tratamiento de gas natural

n BP EXPLORATION COMPANY (COLUMBIA) LTD.

TECHNICAL ASSISTANCE SERVICE

DESIGN, O PERATI ON, FOR THE

AND MAINTENANCE OF GAS PLANTS

SEPTEMBER

2003

Presented by:

O Copyright 1996-2003 by John M. Campbell and Company. These materials are sold or distributed for personal and library use, or for regularly scheduled university classes only; use as a reference or manual for adult training programs is specifically reserved for John M. Campbell and Company. No part of the text or original figures may be reproduced in any form without written permission of John M. Campbell and Company. All rights, including translation rights, are reserved.

'

P'

DISCLAIMER The author, John M. Campbell and Company, takes no position as to whether any method, apparatus

or product mentioned herein is or will be covered by a patent or other intellectual property. Furthermore, the information contained herein does not grant the right, by implication or otherwise, to manufacture, sell, offer for sale or use any method, apparatus or product covered by a patent or other intellectual property right; nor does it insure anyone against liability for infringement of same.

Neither John M. Campbell and Company nor any co-author or other party involved with the writing, prepa- ration, publishing or distribution of these materials shall be responsible or liable in any way for any loss, damage or claim with respect to the use of the information, apparatus, equipment, methods or processes disclosed or described herein. There is no warranty or representation, express or implied, with respect to the accuracy, com- pleteness, or usefulness of the information contained herein. Ail express or implied warranties, including any warranty of fitness for any particular purpose, are expressly disclaimed.

121 5 Crossroads Blvd. Norman, Oklahoma 73072

Website: WWW. JMCAMPBELL.COM E-Mail: [email protected]

Phone: (405) 321-1383 Fax: (405) 321-4533

TABLE OF CONTENTS

4 TABLE OF CONTENTS

I

h

SECTION TITLE

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

OVERVIEW OF OIL AND GAS INDUSTRY

OVERVIEW OF GAS PROCESSING INDUSTRY

PHASE BEHAVIOR

WATER-HYDROCARBON BEHAVIOR

ABSORPTION

GLYCOL DEHYDRATION

GAS SWEETENING / SULFUR RECOVERY

ADSORPTION PRINCIPLES

SOLID BED DEHYDRATION

REFRIGERATION SYSTEMS

VALVE EXPANSION: JT PLANTS

CRYOGENIC GAS PROCESSING

PROBLEM SET

SOLUTION SET

MISCELLANEOUS AND ASSORTED HANDOUTS

@.John M. Campbell & Compnny

~~~~~

Technical Assistance Service for the Design, Operation, and Maintenance of Gas Plants - I ' I I

This page left blank intentionally!

3

3

BP Exploration Company (Columbia) Ltd. /' h

r n ' 1

COURSE: KEY LEARNING POINTS

1.

2.

c

1.

2.

3. 3.

'4

6.

7.

rc'

6.

7.

4. I 4.

9.

5. I 5.

9.

10. 10.

3.

12. 12.

3.

5 . 5.

6 .

7. 7. I 6.

9.

10.

11. 11.

9.

10.

12.

E-mail: registrar (Bjmcampbell. corn Website: www. jmcampbell. com

T

12.

COURSE: KEY LEARNING POINTS

5.

6.

I I

Section: I Section:

5 .

6 .

4. 4. I

7.

8.

9.

10.

11.

12.

7.

8.

9.

10.

11.

12. -.I

11.

12.

Websites

11. -T

12.

ww w. jmcarnpbell. com E-mail: registrar @ jmcampbell. com

Section 1

OVERVIEW OF OIL AND GAS INDUSTRY TABLE OF CONTENTS

PAGE # OIL AND GAS PRODUCTION ...... THE BASIC SYSTEM .....................................................................................

Constraints of the Basic System THE DECISION MODULES .............................................................................................

....................................................................................................... 1.1 ...................... 1.3

................................................................................ 1.4

The Reservoir Module ........................... ................................................................... 1.5

Crude Oil Treating Module .................................................................................................................... 1.6 .............................................. 1.6

Gas Processing Module .............................................. 1.6

The Separation Module ............................................................................... 1.5

Produced Water Treating Module ............................

LIST OF FIGURES FIGURE # PAGE # 1.1 1.2

Schematic View of a Total Production Processing System .......................................................... 1.2 Refrigeration Type of Liquids Recovery Module ......................................................................... 1.3

QJohn M. Cempbeii & Company

Technical Assistance Service for the Design, Operation, and Maintenance of Gas Plants

~

1 .I


NOTES: 3

3

@John M. CampbeU & Company

1 .ii BP Exploration Company (Columbia) Ltd.

! w 1

Section 1

OVERVIEW OF OIL AND GAS INDUSTRY Natural gas is part of the hydrocarbons which are recovered from a petroleum reservoir. Natu-

ral gas processing, which includes sweetening; dehydration; Natural Gas Liquids (NGL) extraction; and fractionation, is critical to both oil and gas production. In most parts of the world gas can not be vented or flared, therefore in order to produce the hydrocarbons (oil and gas), gas processing is required to either make the gaseous hydrocarbons into a product, Natural Gas, or to condition the gas before reinjection into the reservoir. Reinjection is done for reservoir pressure maintenance or enhanced oil recovery.

The gas processing industry is in a precarious place, somewhere between the upstream and downstream segments of the oil and gas industry. The upstream sector is generally where oil and gas exploration and production takes place. While the downstream sector focuses on the refining, marketing, and transportation (RMT) of finished products. The gas processing industry is a microcosm of the entire oil and gas industry. Exploration and production are critical to provide the gaseous feed to a plant while the plant (a miniature refinery) produces finished products which require marketing and transportation. Therefore in an integrated oil and gas company, where does the gas processing division belong? Some have made gas processing part of the exploration and production segment. While others have the gas processing group report through the RMT division. Still others have "spun-down" the gas processing arm into a separate company and others have started companies who focused on the "mid-stream" functions (gathering, processing, and marketing, GPM). Regardless of what you call it, or where you put it, the gas processing industry is a critical link in developing hydrocarbon resources.

4

This manual provides an overview of the gas processing industry and presents the techniques used in recovering natural gas liquids (NGLs). Market, economic, historical, engineering, and operating issues will be presented in the following chapters. This section provides an overview of the oil and gas industry. Section 2, Overview of Gas Processing Industry, presents the market and economic factors for gas processing along with a historical prospective of the industry. Section 3, Qualitative Phase Behavior, reviews engineering principles which are important to understanding how the different NGL extraction methods work. Then Sections 10-8 present the design and operating principles for the five main types of technology (mechanical refrigeration, JT refrigeration, turbo-expansion, absorption oil, and adsorption) used to remove and recover NGLs.

OIL AND GAS PRODUCTION

-. Hydrocarbon fluids produced from a reservoir are seldom (if ever) suitable for direct sale to a buyer. These fluids must be conditioned prior to their disposition. The amount and type of conditioning required vary with several factors such as product specifications, markets, transportation networks, etc.

0John M. Campbell & Company


1.1


le Oil ration

Petroleum reservoirs are a mixture of crude oil, natural gas and water. These constituents are separated and processed to make marketable products.('.')

The gas conditioning and processing equipment is only a part of the entire system. The total system may look very much like that shown in Figure 1 .í For convenience, we divide each system into modules. A dehydration unit, for example, would be a module; as would a fractionation tower with its auxiliary equipment. The choice of modules is governed by convenience, both for calculation and decision purposes.

+ Oil Sales , Oil Crude Oil Field Treating Storage D Reservoir

Water t

Flare

Produced Water Treating Water Disposal

or Reinjection

-+ Gas Sales or Reinjection

t w NGLSales Gas Processing Module

-w COZ Mor Sulfur Sales - Gas

7 1 .

Figure I .I Schematic View of a Total Production Processing System

Unfortunately, one can do a sound job of designing, specifiing and operating each modular unit and yet end up with a poor system. The reason ... each module has varying characteristics under varying loads that may result in a type of internal incompatibility. One modular unit may require a certain incoming analysis to produce the output desired. If a previous unit does not maintain this, then the downstream unit may not prove satisfactory. The fault might not lie so much with that unit but with total system design (even though the unit is usually blamed).

Most of the errors I observe are what I call "errors of omission." Most facets of the problem receiving thoughtful, formal consideration usually are handled satisfactorily. It is the things we fail to consider properly that usually are at the root of most problems. One such omission is to concentrate on the detailed design of each module without proper consideration of the total system within which it resides. Another is failure to properly recognize the degree of uncertainty in the input and output specifications of the system - random variables within practical limits.

The process of simulation is nothing more than performing (in advance) those calculations which characterize system behavior. The most routine form of simulation simply involves solving the equations which (hopefully) describe the operation of concern. Although we currently do much of this on a computer, it adds nothing to the value of the result unless greater true precision is obtained. We may simply obtain more numbers in a given period of time. This in itself is good for more alternatives may be considered. But ... we must remember that better design is not an automatic advantage.

Total simulation must recognize formally the uncertainty (risk) of the numbers used. Using an average or most probable analysis is not an answer. These are only two points on the likely distribu-

@Job M. Campbeli 8s Company

3

1.2 BP Exploratlon Company (Columbia) Ltd.

THE BASIC SYSTEM

Inlet -

4 tion curve (mean and mode respectively). Total simulation must include these concerns so that the system may possess necessary flexibility with minimum use of arbitrary safety factors.

Heat Exchange * Ref rig eration * Separation

r

THE BASIC SYSTEM Figure 1.1 represents a fairly complete processing setup for handling produced fluids. It en-

compasses almost all systems used. Not all elements shown are currently or potentially present in a given system. The purpose is to show most of the common alternatives. The time lag between original reservoir planning and the ultimate disposition of its "goods" (possibly many years hence) requires some initial concern for ultimate potential.

Each of the squares shown represents a calculation module. Within this module there is a body of equations and practice which enables one to design it subject to the imposed constraints. Tradition- ally, adjectives have been used in front of the word "engineer" to loosely define the modular areas chemical, petroleum, mechanical, etc. As the systems have become more complex, calculations within a module can never realistically exclude the other modules.

Not shown in the modular setup are the pumps, compressors, valves and fittings, and lines necessary to move, control and contain the fluids flowing between modules. These are inter-connect- ing modules difficult to show on diagrams.

Some major modules shown have a number of sub-modules representing component parts that involve some unique and/or separate engineering concern. For example, the NGL extraction module could be subdivided as shown in Figure 1.2. This figure is for the very simplest form of refrigeration system consisting of a wellstream exchange, refrigeration source, and separation of liquid from vapor. 4.

* I

1 Gas to Sales

or Other Module

By: 1. Mechanical Refrigeration

2. Expansion Turbine

3. Valve Expansion I Liquid to Sales or Fractionation

Figure 1.2 Refrigeration Type of Liquids Recovery Module

The fact that all of the operations do not occur at, or in the vicinity of, the producing operation does not change the basic system or its needs. The very separation of the functions involved resulting from organizational and geographical considerations dictates the need for an overall planning function. Superimposing this necessary function on top of specific functions ... which have been at least semi- autonomous ... is no easy task. On one hand, the planner does not always possess the technological expertise to impose realistic constraints on each individual element in the system. On the other hand,

4

@John M. CpmpbeU & Company


1.3


7 the people charged with operating each element resist change from those practices which have served them well traditionally. Too often the charge seemingly reduces to, "Reduce cost," when what is meant is, "Increase profit."

Equally often, too much emphasis is placed on "people" and associated costs, within the same system, without adequate concern for the system itself. Profit will result from decreasing cost if all other factors remain the same. Unfortunately, overemphasis on cost usually changes other factors. Handling the system as a system instead of a series of loosely connected individual fiinctions can lead to a more rational basis for greater net profit.

Constraints of the Basic System The system has several basic constraints:

I . The quantity and analysis of fluids entering 2. The market demand (quantity and price) for the effluent products 3. Legal and quasi-legal conditions imposed "no-flare" gas orders, proration, contracts and

agreements, national and political concerns, and the like 4. Environmental factors labor availability and quality, climate, local customs, population

density, availability of utilities and services, and the like

5. The risk tolerance level technological, political and economic 6. The quantity and quality of available data

The problem of predicting the future is self-evident. The techniques for doing this must be reserved for another time. For our immediate purposes let us assume that a forecast of market and associated factors is available and that we possess some realistic measures of the uncertainties involved in that forecast.

9

At this point there are an infinite number of systems that could be devised to market the hydrocarbons available for sale in the reservoir (theoretically). Actually, the choice is limited by a series of practical considerations.

The relative importance of each constraint varies with the individual system. No two systems are exactly alike even though they possess superficial similarity. Many of the errors made are by omission we fail to recognize the sometimes subtle differences in the constraints.

As a practical matter, the first problem is developing the critical constraints (if any). Market- ing is an obvious one but is outside the scope of this discussion. One comment, though, is pertinent technological design must not only serve the present market efficiently but possess sufficient flexibility to accommodate a future market at minimum additional cost. For example, many reserves now exist in areas where there is no significant market for natural gas and natural gas liquids. However, any system design that is incompatible with future gas processing in these areas without unnecessary additional cost is a poor one.

Total reserves might be the paramount constraint. The maximum capital outlay that will yield a fair profit is fixed at some point by this concern. It is then only a matter of finding the best system at this, or hopefully lower, capital outlay to maximize profit.

An uncertain political climate might offer a similar constraint limit the amount of risk capital to that which will afford both a realistic payout time (to reduce time risk) and a satisfactory rate of return.

."5 @John M. Campbell & Company

1.4 BP Exploration Company (Columbia) Ltd.

THE DECISION MODULES

4. These overall economic constraints provide the boundaries for our system "jigsaw puzzle." One then proceeds to the lower order, but equally important, legal and quasi-legal restrictions familiar to all. Compressor capacity rather than the reservoir may limit oil production when a "no flare" order is in effect. Fulfillment of a gas marketing contract may require a production schedule that is "inefficient" from the reservoir viewpoint alone. Many such restrictions are temporary but cause upset in the system.

THE DECISION MODULES Several of the modules shown in Figure 1.1 are discussed to illustrate the considerations in-

volved.

The Reservoir Module A reservoir study generally is undertaken for one of two reasons: to establish value or to

forecast performance under various production strategies, including enhanced recovery. The typical report deals in gross numbers not entirely suitable for productiodprocessing planning and design. Needed is a special report showing greater detail about the character and condition of produced oil andíor gas.

Based on current samples, compositional balances can be made to forecast changes in gas and liquid analysis with time. These are very subjective but order of magnitude changes are detectable with sufficient accuracy to be of value in planning.

4

Geological data are valuable for judgment decisions involving the extrapolation of current data number of wells, likelihood of solids production from core data, gathering system layout, etc.

On almost all new, large reservoirs some form of pressure maintenance is used to permit high initial production rates without excess pressure decline. The injection of water and/or gas usually is involved. At some point in time these will begin to "break through" into the production wells. Well- head pressure will be different; liquid-gas ratios will change. Productioníprocessing system needs will change accordingly. Is the surface system designed to accommodate only current conditions? If so, some major modifications will be necessary eventually. In an offshore or frontier environment the cost of modification plus the hidden cost of inefficient production practices can seriously compromise future profitability and limit reservoir recovery efficiency.

Any forecast of reservoir performance is inexact. But, experience has shown that the formal use of a proper forecast in planning and design considerations by qualified persons leads to more satisfactory surface systems.

The Separation Module With few exceptions, some liquid will be obtained even though the fluid in the reservoir is all

(or primarily) vapor, at reservoir conditions. In this instance, a flash calculation must be made at separation conditions to obtain the quantity and composition of all effluent streams.

If the primary effluent is crude oil or any other liquid stream, containing a reasonable percentage of heavy hydrocarbon molecules (larger than octane), this calculation is difficult. Gas specific gravity alone is inadequate for subsequent liquid recovery computations, less than adequate for even routine dehydration consideration.

4

@John M. Campbell & Company


1.5


Furthermore, even routine changes in temperature and pressure will affect the performance of subsequent modules.

Crude Oil Treating Module This module is required to meet crude oil sales specifications:

1. BS&W (Basic Sediment and Water) 2. Vapor pressure 3. Salt 4. Sulfur content

The BS&W specifications is essentially an entrained water specification. It limits the amount of free water carried with the crude. It often varies from 0.3% to 3.0% by volume with the lower number applied to light crudes and the higher number to very heavy crudes (< 20 N I ) . This specification is typically met by gravity separation. Heat, electricity chemicals, and mechanical coalescers may be used to enhance this process generally referred to as "crude oil dehydration".

The vapor pressure specification limits the volatility of the crude oil. If the crude oil is stored or transported at or near atmospheric pressure this specification will often be equal to or less than 101.3 kPa [14.7 psia] at the system temperature. This specification can be stated in terms of a True Vapor Pressure (TVP) or a Reid Vapor Pressure (RVP).

The specification for salt and sulfur content are frequently met in the refinery rather than at the production facility. Salt is removed by mixing the crude with fresh water and removing the resultant brackish water in a crude oil dehydration module. Sulfur compounds may be removed by gas stripping chemical conversion or a combination of the two.

Produced Water Treating Module Produced water must be treated in order to meet reinjection or disposal specifications:

1. Hydrocarbons 2. Free solids 3. Dissolved solids e.g., CaC03, NaC1, BaS04, etc.

The hydrocarbon specification is particularly important if the produced water is discharged to the sea. For example, in the North Sea the oil content in the discbarged water from an offshore platform is limited to 40 ppm by weight (monthly average). This specification is typically met by gravity separation, flotation units, centrifugal separation (hydrocyclones), or a combination thereof.

Free solids may require removal if the produced water is to be reinjected into the reservoir. Removal methods include gravity separation, filtration and centrifugal separation. Dissolved solids must be analyzed to assess their compatibility with connate water in the reinjection zone or with reinjection water from other sources such as sea water. In this case, specifications can only be established by detailed sampling and testing of the streams involved. These issues are discussed in "Applied Water Technology," available from Campbell Petroleum Series.

Gas Processing Module Natural gas must also be processed to meet basic specifications prior to its sale. These include

water content, hydrocarbon dewpoint, sulfur content and heating value. These will be discussed in more detail in the next section.


'1

r?


THE DECISION MODULES

‘r The gas processing techniques depends on the composition of the gas, product specifications and the markets for natural gas and natural gas liquids (NGLs).

Gas condition and processing are terms used to describe a variety of processes which involve the removal of one or more components from a natural gas or natural gas liquids (NGL) streams. Gas processing is sometimes generically used for any equipment required to make the produced gas marketable. This includes compression, dehydration, sweetening, impurity rejectiodrecovery (nitrogen, helium, etc.) and NGL extraction. In this manual, we will divide gas processing into two categories: conditioning and processing. Gas Conditioning will refer to the dehydration, impurity rejectiodrecovery and sweetening of produced gas. Gas Processing will refer to NGL extraction.

REFERENCES 1.1 1.2

J. M. Campbell, Gus Conditioning and Processing, Vol. 1, Campbell Petroleum Series, Norman, OK (1 98 1). S. T. Pehnec, John M. Campbell and Co., Production and Processing Operations, Sec.2, p.2.1 (May 1995)



1.7


7 NOTES:

@John M. Campbeli LIS Company


rn 1

Section 2

4

OVERVIEW OF GAS PROCESSING I N D USTRY

TABLE OF CONTENTS PAGE #

NATURAL GAS DEFINITION .... .................................. ......................................... 2.2 Paraffin Series Formula: CnH ............................ 2.2

............................ 2.2 Naphthene Series Formula: ............................ 2.3 Aromatic Series Formula: CnH2~-6 ............................................. ......................................... 2.3

............................ 2.3

............................ 2.4

Carbon Dioxide ........................................... ............................ 2.5 .............................. 2.5

.............................. 2.6

.............................. 2.7

.............................. 2.7

.............................. 2.7

.............................. 2.8

.............................. 2.13 .......................................... ........................................ 2.14

.............................. 2.19

.............................. 2.19

.............................. 2.25

............................................................................... 2.32 .............................. 2.33

.............................. 2.40

.............................. 2.41

.............................. 2.42

.............................. 2.42

Olefin Series Formula: CnH2n ....

Helium ........................ ............................................. 2.5

Water ............................................. ..................................... 2.6

Operating Costs ..... ....................................................................................................................... 2.34

A . Absorption ........................................................................................................................................ 2.42 .......................................................................................................................... 2.43

................................................................................................................. 2.45

................................................................................................................. 2.45 A . Absorption ........................................................................................................................................ 2.45 B . Adsorption ........................................................................................................................................ 2.48

@John M . COmpbd 8s CompMy

Technical Assistance Service for the Design. Operation. and Maintenance of Gas Plants


C . Condensation ..................................................................................................................................... 2.50 NGL EXTRACTION OVERVIEW ............................................................................................................... 2.53 NGL FRACTIONATION ............................................................................................................................... 2.54 PRODUCT TREATING ................................................................................................................................. 2.56 How Sweet Gas Can Produce Sour NGL ...................................................................................................... 2.56

LIST OF FIGURES FIGURE # PAGE # 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 2.10 2.1 1 2.12 2.13 2.14 2.15 2.16 2.17 2.18 2.19 2.20 2.21 2.22 2.23 2.24 2.25 2.26 2.27 2.28

Uses of Natural Gas ......................................................................................................................... 2.20 Natural Gas Price ($/MMBtu) - 1985-1997 ................................................................................ 2.21 Uses of Ethane ................................................................................................................................. 2.22 Ethane Price ($/gal) - 1982 to 1997 ............................................................................................. 2.22

...................................................................................................................... 2.23 Propane Price ($/gal) - 1982-1997 ............................................................................................... 2.23 Uses of Butane ................................................................................................................................. 2.24 n-Butane Price ($/gal) - 1982-1 997 ............................................................................................. 2.25 iso-Butane Price ($/gal) - 1982-1997 ........................................................................................... 2.25 Uses of Natural Gasoline ................................................................................................................. 2.26 Natural Gasoline Price ($/gal) - 1982-1997 ................................................................................ 2.26 NGL Pipeline System ...................................................................................................................... 2.28 Waterborne LPG Trade Routes ....................................................................................................... 2.30

Comparison of NGL Pricing with WTI Crude and Spot Natural Gas (Mt . Belvieu) .................. 2.36 Regional Gas Plant Direct Costs ($/Mcf Throughput) .................................................................. 2.36

NGL Plant Capital Costs (1993) ..................................................................................................... 2.39

Process Flow Diagram - Typical Absorption Type Sweetening Process ................................... 2.43 A Glycol Contactor .......................................................................................................................... 2.44 Process Flow Diagram - Typical TEG Dehydration Plant ......................................................... 2.45

Process Flow Diagram - Mechanical Refrigeration System with EG Injection ........................ 2.50 IFPEXOL Process (Brazeau River Plant) ....................................................................................... 2.51 Process Flow Diagram - Typical NGL Fractionator ................................................................... 2.54 Typical Fractionation Plant Contaminants ..................................................................................... 2.57 Amine Liquid Treating Process ...................................................................................................... 2.58

Shrinkage Value of NGL Products Based on Fuel as Ideal Liquid .............................................. 2.35

NGL Plant Operating Costs ............................................................................................................. 2.37

NGL Plant Operating Economics .................................................................................................... 2.40

Process Flow Diagram - Typical Adsorption Dehydration Plant ............................................... 2.48

LIST O F TABLES TABLE # PAGE # 2.1 Elemental Mercury Concentrations in Natural Gas ....................................................................... 2.7 2.2 Liquid Content of Natural Gas ........................................................................................................ 2.8 2.3 Examples of Gas Composition in the U.S ...................................................................................... 2.8 2.4 Example of Sales Gas ...................................................................................... 2.9 2.5 North America Pipeline Quality Specification Survey(’.”) .......................................................... 2.10 2.6 2.7(a) Physical Constants of Paraffin Hydrocarbons and Other Components of Natural Gas .............. 2.14 2.7(b) Physical Constants of Paraffin Hydrocarbons and Other Components of Natural Gas .............. 2.15 2.8 Typical Regional Transportation and Fractionation Fees 1994 (#/gal) ........................................ 2.29 2.9 Fractionation Capacity, bblíd .......................................................................................................... 2.32 2.10 Typical NGL Plant Recoveries, % .................................................................................................. 2.52 2.1 1 NGL Contaminants .......................................................................................................................... 2.55 2.12 Percent Sulfur Compound Distribution in Fractionated Products ................................................. 2.56 2.13 NGL Treating Methods ................................................................................................................... 2.56

Natural Gas and NGL Product Specifications ................................................................................ 2.11

@John M . Campbell & Company

2.ii BP Exploration Company (CdumMa) Ltd .

4

Section 2

r,

OVERVIEW OF GAS PROCESSING I N D USTRY

As we reviewed in Section 1, there are many constraints on an oil and gas production system design. Many of these same constraints apply specifically to the gas processing module.

The gas processing treatment schemes actually used in a particular field depend upon several factors. The most important of these are listed below:

1.

2. 3. 4. 5 .

6 . 7 . 8. 9.

10.

Reservoir conditions - fluid composition, temperature, pressure Field development plan - volumes, temperatures, pressures Legal restrictions - no-flare orders, contracts, nominations, etc. Environmental factors - field location, labor, local customs, etc. Markets - natural gas, NGL, sulfur, and C02 market availability, quantity, price Product specifications (gas) - water and hydrocarbon dewpoint, H2S, heating value Product specifications (NGL) - vapor pressure, water content, H2S and C02 Economic - profitability of treatment process Political - national interests such as resource conservation Processing agreements - for NGL extraction

The reasons for processing are: 1. It is operationally necessary. 2. It is commercially necessary to meet product specifications. 3. It is economically attractive.

While all of the above items are critical to an effective design, four factors which warrant more emphasis are 1) fluid composition, 2) gas contracts and specifications, 3) NGL contracts and specification, and 4) Processing agreements. These are discussed below.

NATURAL GAS DEFINITION Before discussing the industry or the technology of gas processing it is best to define natural

gas. Natural gas is a mixture of hydrocarbons and impurities. There is not a single "type" of natural gas. It will vary from field to field, from well to well, from different formations in the same field, and even over time from an individual well. Characterizing natural gas is imperative to developing the proper gas processing strategy and plant design.

4

@Job M. Campbell & Company


21


What are hydrocarbons? Hydrocarbons are molecules which are composed of only hydrogen and carbon atoms. Atoms are the basic unit of each element that can combine with itself or the atoms of other elements to form a compound. There are different types of hydrocarbons that can occur or are important to gas processing. These include paraffins, olefins, naphthenes, and aromatics.

Paraffin Series Formula: CnH2n+2 (2.1 1

Paraffin hydrocarbons are the most plentiful hydrocarbons in natural gas. Hydrocarbons in this series are saturated compounds - all four carbon bonds are connected either to another carbon atom or a hydrogen atom, with one such atom for each bond.

H H H H H H I l l

H - C - C - C - H I I

H - C - C - H I

H - C - H I

H I I

H H i l l

H H H Methane Ethane Propane

Notice that all names end in -me, the ending used for the paraffin series. In this case, the number of hydrogen atoms is two times the number of carbon atoms plus two more for the ends of the chain. Paraffin hydrocarbons are the most stable of the lot. In referring to a given paraffin hydrocarbon, the abbreviation C1 for methane, C2 for ethane, C3 for propane, C4 for butane, etc. may be used. Statements like "propane plus fraction (C,+)" refer to a mixture composed of propane and larger compounds.

When the paraffin series molecule contains four or more carbon atoms there are different ways these can be connected without affecting the formula. Compounds which have the same chemical formula but a different atomic structure are called isomers. They possess different physical and chemical properties.

There are only two isomers of butane. In the structural diagram shown for i-butane we could draw the carbon atom below instead of above the carbon chain. But, this would be just a "mirror image" of the molecule as drawn. It is the same molecule with the same properties. The adjective "normal" is used to designate a molecule wherein all of the carbon atoms are in a straight line. An "isomer" has the same formula but the carbon chain is "branched" as shown below. In an analysis, these are often abbreviated as "n" and "i" respectively.

H H H I l l

H H H H I I I I

H-C-C-C-H I l l

H - C - C - C - C - H H C H I I I I

H H H H

Normal butane (n-butane)

'1" H H

Icobutane (i-butane)

@John M. Campbell (& Company

2.2 ~ _ _ _ _ ~~~~~

BP Exploration Company (Columbia) Ltd.

NATURAL GAS DEFINITION

Olefin Series Formula: CnH2"

The olefin group of compounds is a simple straight chain in which all the names end in -ene. Ethylene (ethene) C2& is the simplest molecules in the series.

Hydrocarbons in this series combine easily with other atoms like chlorine and bromine, without the replacement of a hydrocarbon atom. Since they are so reactive, they are called unsaturated hydrocarbons.

Unlike the paraffins, the maximum bonding capacity of the carbon atom is not fully satisfied by hydrogen or carbon atoms. Two adjacent carbon atoms form a "temporary" bond (in the absence of other available atoms). It is a necessary but unstable alliance. The structural formula for the olefins uses a double line to indicate the double carbon-carbon linkage, the most reactive point in the molecule.

H H H H H

C = C , H2C = CH2 I l l I I

I t I H-C = C-C-H , HK=CHCH3

H H

Ethylene (Ethene)

H Propylene (Propene)

The amount of olefins in natural gas usually is fairly small. Certain crude oils contain them in measurable amounts. Olefins are important to gas processing because all recovered natural gas liquids (NGL) can be used as feedstock to petrochemical facilities which produce olefins. Olefins are used to make plastics and synthetic fibers.

I

Naphthene Series Formula: CnH2n

The naphthene series has a ring structure. Naphthenes may be found in most crude oils but are seldom shown in routine analyses. Be- ing saturated molecules, they are not very reactive. Cyclohexane is a common member of this series. Its structural formula is C6H12. On chromatographic analysis it occurs between n-hexane and n-heptane. Cyclopentane (CsHio) also occurs. On a chromatographic analysis it is present between n-pentane and n-hexane. If the natural gas which is being processed is rich or produced from a reservoir containing naphthenic crude, then an extended gas analysis should be performed to determine if naphthenes are present in the natural gas. Naphthenes may plug exchangers in cryogenic processes.

Aromatic Series Formula: CnH2n-6

H H \ /

H C - H \ / \ /

H - C C - H I 1

C-H H - C

H / \ / \

H C

H / \

H

Cyclohexane

H

C

H-C C-H

I & \

II

C-H Aromatic is the word used to describe an unsaturated hydrocar-

Benzene, the parent compound of this series, has the structural formula O f C6H6. C

I

I bon molecule where the carbon atoms form a ring, a cyclic compound. H - C * /

H Benzene

4 Since the aromatics are unsaturated, they react readily. They may be oxidized to form organic acids. They also promote foaming and other operational problems in the production and handling of crude oil

@John M. Campbell & Compauy


2.3


and natural gas. Also, aromatics (benzene, toluene, ethyl-benzene, and xylenes) are potential carcino- genic material which are regulated by federal and state agencies.

Most petroleum contains only trace aromatics. Some contain significant amounts. Any analysis of crude oil and natural gas should include aromatics. Even small amounts can influence physical behavior and affect design.

Heavy Hydrocarbons Natural gas is usually composed of methane, ethane, propane, butanes, pentanes, and a mixture

of higher molecular weight hydrocarbons often referred to as "hexane plus" (C,j+). The characterization of the c6+ fraction is important in the proper design of NGL extraction facilities. The primary reason for the gas processing is removing the heavy hydrocarbons to prevent hydrocarbon condensation in the gas transmission pipelines. This characterization is generally determined by performing an extended chromatographic analysis or via a distillation (either ASTM D86 or True Boiling Point Analysis, TBP). It may also be prudent to have a "PNA" (Paraffin, Naphthene, and Aromatic) chromatographic analysis performed to properly characterize the produced gas. Gas sampling is just as important as the analysis performed on the gas, but this subject is beyond the scope of the manual.

Another heavy hydrocarbon which may be present in produced gas is diamondoids. Some reservoirs contain gas which precipitate crystalline solids in surface facilities which can plug flowlines and equipment. These solids are diamondiods which are a saturated, cyclic hydrocarbon compound with a diamond, or cage-like structure with molecular weights of 136 to more than 270. There are three types of diamondiods: adamantane, diamantane, and triamantane. Diamondiods have posed processing problems in the Mobile Bay area in the Gulf of Mexico and in Hanlan Swan Hills gas field in Canada. Diamondiod deposition may be controlled by circulating a hydrocarbon solvent (e.g. diesel) to prevent the diamondiod from precipatating.(2.6)

What impurities are found in natural gas? Impurities can be considered anything that prevents the gas from being marketed. Impurities consist of both hydrocarbons and other compounds. The different types of hydrocarbons were discussed above. Some hydrocarbons are considered an "impurity" because heavy hydrocarbons can form liquids at pipeline operating conditions. These liquid slugs can reduce pipeline capacity, cause compressor failures, increase corrosion, and causes flaring at the burner tip. Heavy hydrocarbons can also cause the heating content of the gas to be too high for the burner configuration. The non-hydrocarbon impurities can be toxic, corrosive, or cause problems in processing, transporting, or burning natural gas. These contaminants are discussed below.

Nitrogen Nitrogen acts as a dilutent in natural gas. Nitrogen is nontoxic, noncombustible and noncor-

rosive. It is generally present in natural gas streams and can vary in composition fiom low percentages (not requiring treating) to concentrations in excess of 50 mol%. High nitrogen reserves tend to be found in the Mid-Continent region such as the Hugoton and Anadarko basins. The Permian Basin, Smackover formation in East Texas, Arkla Basins, Green River, and Sacramento Basins also contain high nitrogen reserves.(2,2) The main problem with nitrogen is that it has no heating content and therefore it can affect burner performance. Another problem with nitrogen is the additional transportation cost incurred by pipelines. Since gas transportation cost is based on heating content ($íMMBtu), pipelines do not receive direct payment for transporting nitrogen present in the gas stream. Nitrogen can be removed through cryogenic separation, absorption facilities or adsorption systems. The rejected nitrogen can be vented to atmosphere, recovered for enhanced oil recovery (EOR), or recovered for liquid nitrogen uses.

@John M. Campbell SE Company


NATURAL GAS DEflNlTION

Helium Helium is the kind of "impurity" that everyone wishes they had in significant quantities. He-

lium is an inert, nontoxic, noncombustible, noncorrosive fluid. Helium is used in the medical field, research, manufacturing, commercial balloons, oil and gas exploration, and defense industry. Helium has recently commanded a price between $40-60 per M s ~ f . ( ~ . ~ ) Not all produced gas streams contain helium. The Hugoton Basin has the greatest concentration of helium recovery units (HRU). Two other regions have HRUs: ArklaíEast Texas and the Wyoming. The commercial processes to recover helium include cryogenic separation and adsorption. Generally helium concentration and gas flows capable of producing 100 MMsc@ of helium are required to economically install and produce a helíum pr~duct . (~.~) When high concentrations of helium are present, there is generally a high concentration of nitrogen which requires a nitrogen rejection unit (NRU) to meet the sales gas specifications.

Carbon Dioxide Carbon dioxide is also a dilutent in natural gas, but unlike nitrogen and helium, carbon dioxide

is corrosive (especially when liquid water is present). When cryogenically processing natural gas, solid COZ can form (dry ice). When the concentration of COZ is above 1%, COZ freezing should be checked in the turboexpander plant design and operations. Carbon dioxide specifications are established by the sales gas pipeline and liquid products containing ethane. Essential all natural gas contains some amount of carbon dioxide. Concentrations may vary from low percentages (not requiring treating) to high concentrations (greater than 80%) found in C02 EOR projects. The Rocky Mountain region contains gas production with carbon dioxide concentrations generally in the 2-20% range. Also the Texas Gulf Coast region and the Austin Chalk formation are known to be high in carbon dioxide.(2.2) Processes to remove carbon dioxide include absorption systems (chemical, physical, and hybrid solvents), extractive distillation (Ryan Homes), Exxon CFZ, and membranes.

-.

Hydrogen Sulfide and Trace Sulfur Hydrogen sulfide and other trace sulfur components (mercaptans, carbonyl sulfide, carbon di-

sulfide, and elemental sulfur) must be removed to low levels due to the toxic and highly corrosive nature (in the presence of liquid water) of these compounds. Hydrogen sulfide has a threshold limit between 10-20 ppm (max) and a lethal concentration of 600 ~ p m f ~ . ~ ) . H2S prone formations include the Green River, East Texas, Permian, and Gulf of Mexico basins.(2.2)

Mercaptans are a strong smelling compound which is added in very small quantities to "stench" natural gas and LPG used in residential and commercial fuel applications. When mercaptans are present in significant quantities, removal is required and generally performed with the removal of the other sulfur compounds, but mercaptan removal is more difficult and affects the process selection.

Sulfur components will distribute into the different products of a natural gas liquids stream (NGLs). Carbonyl sulfide will typically be found in the ethane and propane product. Methyl-mercaptan generally distribute in the ethane, propane, and i-butane and the ethyl-mercaptan distributes in the n-butane and gasoline. Some di-methyl sulfide will be found in the n-butane and all other sulfur species will be present in the natural ga~oline.(~.~)

Even though sulfur compounds are not always present in natural gas, an analysis for H2S should always be performed and if the concentration of H2S is greater than 3-10 ppm then carbonyl sulfide, carbon disulfide and mercaptans analysis should be performed. Sulfur compounds can be removed by absorption systems (chemical, physical, or hybrid), adsorption systems, direct conversion,

14



2.5


and membranes. The sulfur compounds from these treating units (with the exception of direct conversion) are either incinerated to form SO*, reinjected into a reservoir or processed to recover sulfur.

The sulfur compounds which are recovered in these processes may be furthered processed to form elemental sulfur. This is generally performed in a Claus process, but there are other techniques which can be used. It can be difficult to attain very high levels of sulfiir recovery required by some federal or state regulations when carbonyl sulfide and carbon disulfide are present. Carbonyl sulfide can also form nonregenerable compounds with certain chemical sweetening solvents (e.g. MEA).

3

Water must be removed to prevent operational problems resulting from hydrate formation and corrosion. It is also nontoxic and noncombustible, but the hydrate and corrosion issues seem to over- shadow these characteristics. Water is always present in produced gas, but the gas may not be saturated with water by the time it reaches a gas processing facility. The water content of gas can be predicted through a number of different correlation presented in Campbell's Gas Conditioning and Processing- Volume 1. Water removal methods are generally matched to the gas processing technology and/or the sales gas specification. Dehydration processes include absorption systems (TEG and DEG), adsorption systems (molecular sieve, silica gel, etc.), condensation systems (EG and methanol), and membranes.

Trace Elements

Many other elements may be found in produced gas near the lower limit of detection (parts per billion-ppb). The most infamous of these is mercury (Hg). Mercury has been known to be present in natural gas since the ~ O ' S , but its affect on gas processing equipment was not identified until equipment failures in the early 1970's. Elemental mercury is generally found in gases that do not contain hydrogen sulfide. Table 2.1 below lists some mercury concentrations measured in natural gas ~orldwide.(~.~)

Mercury attack is found in brazed aluminum (plate-fin) heat exchangers and aluminum piping. The mercury induced cracks occur in the high magnesium aluminum headers or pipe welds versus the core. The metal corrosion does not proceed at a significant rate if the temperature is maintained below -39°C [-38'F], the freezing temperature of mercury. Above this temperature corrosion may occw at very high rates. Mercury removal entails trapping mercury from the gas on a sulfur impregnated activated carbon or alumina or silver loaded molecular sieve. The material selection is based on cost and presence of heavy hydrocarbons. Hydrocarbon liquids, TEG, and amines may "leach" the sulfur off the activated carbon material.(2.22' Other mercury corrosion control methods include modifications to heat exchanger and piping design such as seal welding of backing rays to prevent liquid mercury from accumulating or operating procedures to desorb mercury before warmup (denme) or shutdowns. Manufacturers of aluminum plate fin exchangers typically recommend that the mercury concentration be reduced to long/Nm3 upstream of the exchangers.(2.8)

Another trace contaminant is arsenic. It was first discovered to be a problem in gas from the Ab0 Field in New Mexico. The potential problem with arsenic is its toxic nature. Arsenic removal equipment is not common in the natural gas industry.f2.2)

Other less toxic metals and naturally occurring radioactive materials (NORM) may be present in natural gas, but their affects on design and operation of gas processing facilities are not significant at this time.


r)

2.6 BP Exploration Company (Columbia) Ud.

NATURAL GAS DEFiNlTlON

Typical Gas Compositions

TABLE 2.1 Elemental Mercury Concentrations in Natural Gas

Location

Algeria (wellhead) Groningen (wellhead)

North Germany (Wellhead) South Germany (wellhead) South America

Far East Far East

Far East

Africa

Middle East Eastern U.S. Pipeline Midwestern U.S. Pipeline

US. - S.W. Wyoming U.S. - San Juan Basin U.S. - Hougoton Basin

U.S. - Permian Basin U.S. - Texas Gulf Coast

CLg/Nm3 50-80

180

15-450, 5000

< 0.1-0.3

69-119

3-20

58-193

0.02-0.16

0.3-130

1-9

0.019-0.44

0.00 1-0.1 o 0.2-24

1.5

< 0.2-2.2

< 0.2-4.9

0.03-579

Even though natural gas compositions can vary significantly, there are trends which are useful for estimating the plant performance. In the continental United States the gas is generally richer in the Rocky Mountains and becomes leaner as you move towards the Gulf of Mexico, with the exception of the Permian Basin which has the riches gas in the country.

The term "GPM" is used to define the liquid content of natural gas. GPM is an abbreviation for Gallons (of liquid) per thousand standard cubic feet (of gas). In the oil and gas industry thousand is usually abbreviated by the Roman numeral "M." Therefore Mscf is the designation for thousand standard cubic feet. To distinguish the liquid content of gas (GPM) from liquid flow (gpm-gallons per minute), the previous will be in all capital letters while the later will be in all lower case letters.

Describing gas as "rich" and "lean" is subjective. Most of the industry designates rich gas as having a liquid content greater than 5 GPM (C,+), a moderate liquid content between 5 and 2 GPM (C,+) and a lean gas with less than 2 GPM (C2+). When defining the liquid content an adjective describing which hydrocarbons are included in the calculation should be used stating the gas GPM. The above numbers are the liquid content for "ethane plus."

There is another adjective which can be used to describe the liquid content. This is the inlet GPM, which is sometimes abbreviated I-GPM and the recovered GPM (R-GPM). The I-GPM describes the maximum liquid recovery if all of the components were condensed into the liquid product stream. Since not all the light components are recovered, the R-GPM will be lower than the inlet GPM. The calculations for these numbers will be presented later in this section.

.4




An example of the liquid content for gas processing units in the continental U.S. is listed in Table 2.2.(2.9)

TABLE 2.2 Liquid Content of Natural Gas

3

Rocky Mountain States

Oklahoma-Kansas

Texas Panhandle

Permian Basin East TexasíTexas Coast

Louisiana, Miss., Ala.

Gulf of Mexico-Offshore

4.8

4.1

4.8

6.2

2.7

2.0

O .9

Examples of gas compositions from these regions is shown in Table 2.3. There may be significant variations from different wells or different fields from the example shown below.

Now that we have seen the many variations in the composition of produced gas, we realize that time and effort must be spent obtaining representative samples and determining realistic characterizations. Understanding and identieing the sales gas quality specifications, the NGL product specifications, and the gas processing agreement is just as important and should not be overlooked. These three legal documents place a "box" around the gas processing systems which constrains the process selection.

TABLE 2.3 Examples of Gas Composition in the U.S.

Component Nitrogen

Hydrogen Sulfide Carbon Dioxide Methane Ethane

Propane i-Butane n-Butane

i-Pentane

n-Pentane

n-Hexane

n-Heptane

n-Octane

Total

c2+ gpm

Rocky Mt 0.12

0.00 1.58

86.75

7.75

2.38 0.45

0.43

0.18

O. 14

0.12

0.08

0.04

100.00

3.23

Okla-KS 2.16

0.00 0.34

81.54

8.48

4.63

0.50

1.42

0.27

0.37

0.32

0.00

0.00

100.00

4.51

Permian Basin 2.89

0.02

0.05

70.45

12.77

7.93

1 .O6

2.66

0.66

0.70

0.51

0.20

0.10

100.00

7.60

E. TX./ TX Gulf 0.83

0.00

0.18

88.88

5.74

2.49

0.73

0.62

O. 14

0.09

0.12

0.1 1 0.11

100.00

2.89

@John DI. CampbeU & Company

Gulf of Mexico 0.22

1.33

4.00

88.15

4.51

1.18

0.26

0.20

0.05

0.02

0.01

0.02

0.05

100.00

1.74

2.8 BP Exploration Company (Columbla) Ltd.

CONTRACT TERMS

CONTRACT TERMS

Gas Contract Quality Gas 1. Minimum, maximum, and nominal delivery pressure 2. Maximum water content (expressed as a dewpoint at a given pressure or as a concentration) 3. Maximum condensable hydrocarbon content expressed as a hydrocarbon dewpoint, analy-

sis, or a cricondentherm temperature (this term will be discussed in more detail in the next section).

4. Maximum delivery temperature 5. Allowable concentration of contaminants such as H2S, carbon disulfide, mercaptans, etc. 6. Minimum and maximum heating value 7. Cleanliness (allowable solids concentration)

These quality specifications are extremely important factors in the selection of the gas treatment process. Quality specifications for gas handled by transmission and distribution companies vary from country to country. In a number of instances specifications have been set for historic reasons rather than technical reasons. Table 2.4 provides a comparison between the U.K., North America, and Middle East specifications. In the case of North America, those given are typical and will vary somewhat from pipeline to pipeline. This is shown in Table 2.5.

usually contain the following basic considerations:

\ TABLE 2.4 Example of Sales Gas Specificationsf2.’2’

4

1020- 1200 Btdscf

Notes: 1) std m3 is abbreviation of standard cubic meters (at 15°C and 101.325 Wa). 2) Wobbe Index is equal to:

Gross Heating Value d Specific Gravity

and is a way of characterizing the heat release at the burner tip for gaseous fuels.

@John M. CampbeU & Compmy


2.9


u> rn c

O E e

s +

N O 9 N

O

rn

2.1 o BP Exploration Company (Columbia) Ltd.

CONTRACT TERMS

Natural Gas NGL 1 Condense Produced Pipeline

Gas Gas or (Inlet Gas, (Residue Mixed Propane Butanes Natural Raw Gas, Gas, Dry NGL or (Mix or Gasoline 1

Liquid Contract Quality Specifications Products from natural gas - Natural gas liquids (NGLs) consist of all hydrocarbons heavier

than methane which can be extracted from a natural gas stream. A list of NGL products is provided below:

1. C2 Ethane 2. C3 Propane

Table 2.6 presents a table of NGL components normally found in natural gas and the resulting NGL products. A more detailed discussion of NGL products is covered below and in the NGL markeí factors of this section.

3. iC4 Isobutane 4. nC4 Normal Butane

5 . Cg+ Pentanes and heavier

Liquid contracts usually contain the following basic considerations:

1. Quality of products expressed as vapor pressure, relative or absolute density, or by standard

2. Specifications such as color, concentration of contaminants, etc., as determined by standard

3. Maximum water content

Liquid products may be classified into two general categories stock tank fluids from separators and fractionated products. The former is normally sold to a pipeline and is subject only to any pipeline limitations such as BS&W, specific gravity, vapor pressure, and presence of "light ends." It is sometimes referred to as a "slop" product to distinguish it from those products falling in the second cate- gory. In essence, the composition of this product will be fixed by the equilibrium relationships at the pressure and temperature of the storage tank.

designation such as Commercial Propane

tests

All other liquid products are a result of a fractionation which separates a raw mixture into its component parts based on vapor pressure and other component physical properties. These are commonly called natural gas liquids and are produced from what are called NGL plants. If the eMuent gas from an NGL plant is totally liquefied, it is called liquefied natural gas (LNG). -

The amount of processing done at the production site depends on the amount of fluids, available transportation to market, and local conditions. Offshore, swamps, jungle or arctic type locations

O J o h DI. Cpmpbell& Company

Technical Assistance Service for the Deslgn, Operation, and Maintenance of Gas Plants

2.1 1


limit the feasibility of more complicated systems. The accent is on merely doing the least necessary at the site to transport the production to more favorable surroundings for future processing.

In a NGL installation one or more of the following products may be specified. The definitions are those of the Gas Processors Association.

Natural gasoline. This is a mixed product whose basic specification is vapor pressure. As a general rule it must meet all of the following specifications.

Reid vapor pressure: 70-235 @a [lo-34 psia] Percentage evaporated at 60°C [ 140"FI: 25-85% Percentage evaporated at 135°C [275"F]: not less than 90% End point: not higher than 190°C [375"F] Corrosion: not corrosive by specified test Doctor test: negative Color: not less than plus 25 (Saybolt) Water content: no free water

Commercial ethane. This primarily is a chemical feed stock for the manufacture of plastics and associated materials. The specifications vary but are usually rigorous for contaminants may affect processing. Limitations on COZ and C1 are particularly important.

Demethanized mixes. An increasing number of pipelines are purchasing a demethanized product containing ethane, propane, butane, and natural gasoline. This is sometimes abbreviated as EPBC (ethane, propane, butane, and condensate). There are no standard specifications on such mixtures, although the methane content is often limited to I-3% of the ethane. Also, there is typically a limit on the COZ (1.0 mol% max or 0.35% C02 in the ethane). The purity tests of natural gasoline usually apply. There may also be a maximum true vapor pressure specification of 600 psia at 100°F.

The contract usually calls for periodic analysis of the stream. This is then used to divide the total purchase into natural gasoline, butanes, propane and ethane. The applicable rate for each component times its quantity is then added to establish the total revenue. There is usually a fee paid for transportation and fractionation (T&F fee) which is subtracted from the liquid price.

Commercial propane. This defines a fluid composed of at least 95% propane andor propylene, whose true gauge vapor pressure must not exceed 1.43 MPa(g) [208 psig] at 38°C [100"F], and which satisfies GPA tests for total sulfur, residue, dryness, and corrosive compounds. Purchasers may offer about 90% of the resale price. This resale price is not tied to any base and fluctuates widely, depending on season and geographical location.

Propane HD-5. This is a special grade of propane for motor fuel and other use which shall not have a true vapor pressure of more than 1.43 MPa(g) [208 psig] at 38°C [10O0F]. It conforms to all other Commercial Propane tests except that it cannot contain over five liquid volume percent (5 LV%) propylene and must contain at least ninety liquid volume percent (90 LV%) propane.

Commercial butane. A product meeting this designation is composed predominantly of butanes andor butylenes and has a true gauge vapor pressure not greater than 483 kPa(g) [70 psig] at 38°C [10O0F]. The temperature at which 95% by volume has evaporated shall not exceed 2.2"C [36OF] when corrected to a barometric pressure of 100 kPa. It must meet the same general tests as propane for contaminating substances.

Butane-propane mixtures. The standard definition of this mixture is that it shall not have a true vapor pressure higher than commercial propane at 38°C [100"F] and shall pass the 95% boiling

@John M. CampbeIi L Company

2.1 2 BP Exploration Company (Columbia) LM.

CONTRACT TERMS

point test for butane. It must further pass all butane purity tests. In most instances it is sold for domestic heating service or used for secondary recovery of oil. The composition, for heating use, is varied to assure volatility in the various seasons, but the vapor pressure of the commercial product seldom exceeds 860 kPa(g) [125 psig] at 38°C [lOO°F].

Vapor Pressure Vapor pressure specifications may be expressed in terms of a TVP (True Vapor Pressure) or

RVP (Reid Vapor Pressure). These concepts will be discussed in a later section.

The vapor pressure or volatility specification is probably the most important factor determining the ultimate amount of NGLs recovered and the type of NGL process selected.

The vapor pressure of an NGL product can be estimated from the following equation:

pvmiX = C xi pvi

Where: PVmi, = vapor pressure of the NGL mixture, bar abs [psia]

xi = mol fraction of each component in the mixture P,, = vapor pressure of each component in the mixture, bar abs [psia]

This equation simply says that the vapor pressure of an NGL mixture is proportional to the vapor pressure and amount of each individual component in the mixture. For specification purposes, vapor pressures are almost always expressed at a temperature of 38°C [lOO°F].

4

For your convenience, Tables 2.7(a) and 2.7(b) present the physical properties of several NGL components, including the vapor pressures, in both SI and FPS units.

Notice in the following example that propane has the largest effect on the vapor pressure of the NGL product. Propane’s contribution to the total vapor pressure of the product is (4.02/5.11), or almost 80%, even though its composition is only 30%.

4

~~~ ~~ ~ ~~ ~ ~ ~ ~ ~ ~ ~

xampie 2.1: Estimate the true vapor pressure (TVP) of an NGL mixture with the following composition.

SI Units FPS Units

Component Xi Ph, bar abs xi pvi hi 7 psis xi pvi c3 0.30 13.41 4.02 188.40 56.5

nC4 0.20 3.77 0.75 51.71 10.3 nC5 0.20 1.16 0.23 15.574 3.1 nc6 0.30 0.373 0.1 1 4.96 1.5

Zxi P, = 5.11 barabs C xi P, = 7 1.4 psia 1 .o0



2.13


I I

- ' I


CONTRACT TERMS

4

- m m -!g

N P - o

2:s N N

* m w o\.?- - - m - m N ' ---t--t


2.1 5


This is an extremely important concept in the design and operation of NGL facilities. We want the extraction and stabilization processes to be selective. This means we want to drive out as many of the highly volatile components as possible from the liquid product while retaining the less volatile components. The highly volatile components raise the vapor pressure of the product unnecessarily and reduce the 'troomtt (in terms of vapor pressure) for the less volatile components. The following example indicates this concept.

In Example 2.2, notice the effect of the selectivity of the stabilization system upon product recovery. If the stabilization system allows a mere 2% ethane in the liquid product, the product rate decreased by almost 10%. The small amount of ethane in the product contributes over 30% of vapor pressure.

Example 2.2: ~ ~~

An NGL product must be stabilized to meet a TVP specification of 3.5 bar abs [50 psia] at 38°C [10O0F]. The unstabilized product has the following composition:

1 nC6 1 35

What percent of the unstabilized feed can be recovered as liquid product if

a) no C2 remains in the liquid product? b) 2% C2 remains in the liquid product?

No ethane in product

P, mix = 50 psia

2% ethane in product

propane

P, mix = 50 psia


'3

3

2.1 6 BP Exploration Company (Columbia) Ltd.

CONTRACT TERMS

Two basic types of stabilization systems are employed in oil production and NGL recovery facilities. These are:

1. Flash stabilization - NGL product is flashed to progressively lower pressures. Flash vapors are recompressed and may be recycled back into the inlet. Heat exchangers may be used to control the product temperature. The oil-gas separation system on the North Slope uses this type of stabilization process to meet the TVP specification of the export crude.

2. Distillation - NGL product is stabilized at constant pressure in a packed or trayed tower. Temperature is varied from the top to bottom. Stabilizer tower may be refluxed or non-refluxed (top tray feed). A non-refluxed stabilizer is often used in gas processing plants to control the vapor pressure of the NGL product mixture.

The flash stabilization method (1) is the traditional method used in field processing facilities, especially offshore. It has the advantages of simplicity, ease of control, and it simplifies topside design. Method (2), distillation, is more common in gas processing facilities. It has the advantages of more selective separation, so product rates are higher. It also minimizes recompression costs.

Processing Agreements In addition to gas and liquid contract quality specifications, there are processing agreements

which may be used to establish the legal framework for gas plant economics. There are four major types of agreements: 1) keep-whole, 2) flat rate, 3) percent of proceeds, and 4) processing fee. There are variations to each of the agreements and there are other types of agreements which can be negoti- ated between parties. Therefore this is a general discussion of gas processing contracts.

4 The creativity of parties is the only limit of how the "pie" is cut. The pie will only be so big and the type of contract which is finally executed is dependent on the risks and potential rewards each party wants (or needs). There is no "right" answer to which type of agreement a company generally works toward. As summarized below, different companies have chosen which risks they wish to be exposed to and there are examples of "successes" for each type of agreement. Also most facilities have a combination of agreements to hedge the risks.

The gas processing industry performs critical functions required to make the produced gas marketable. These functions include gathering the gas, compressing the gas to transmission pipeline pressures, and conditioning the gas into a marketable product (removing impurities to meet pipeline specifications). In addition to these essential functions, gas processors may elect to perform discretionary activities to recover more NGLs than is required to make the gas salable. Gas processors perform this activity because of the "liquid upgrade" realized by selling the components (ethane, propane, etc.) as a liquid versus a gas. These calculations will be explained in the NGL economics part of this section.

4

As the gas is gathered, compressed and processed, there is both a volume "shrink" and a heating content "shrink." The "shrink" is a reduction in the amount of gas available for delivery into a pipeline. This is sometimes called PVR - Plant Volume Reduction. In the gathering operation, the shrink is due to field fuel (from compressors, dehydrators, andor treaters) and line loss (due to leaks in the gathering system). At the gas plant the shrink is from plant fuel (compressors, heaters, etc.), flare (plant upsets), and recovered liquids. In a deep-cut recovery plant (e.g. turboexpander), the liquid shr ink may be the largest component to the overall shrink.

The different type of companies who perform gas processing are listed below: Major Integrated Oil & Gas Companies Independent Oil & Gas Companies



2.17


Regulated Pipelines & Affiliates

Third Party Processors

Below is a discussion of the different agreements, risks, and rewards.

1. Keep-Whole Agreements. Gas is usually sold and transported on a heating value basis (Btu or MJ). Therefore producers who want to be only exposed to the gas market (versus the liquid market) generally want to enter into keep-whole agreements where the gas processor is responsible for keeping the producer "whole" on the Btu's produced. The gas processor can make a cash payment equivalent to the Btu's extracted times a gas price. The gas price is usually indexed on a spot gas price for a given delivery point or an average of spot gas prices for a group of delivery points. Instead of making this payment the gas processor may need to buy gas to replace the shrink Btu's.

One of the variations to the keep-whole agreement is to have a Btu cap or crude cap clause. This clause "caps" the gas price at a maximum price (Btu cap) or a maximum price relative to the price of crude oil (crude cap). To calculate the crude cap the crude price is divided by an approximate crude oil heating content of 5.8 MMBtuhbl.

Under this contract, the risks to the producer are only the gas market price. They are not exposed to the liquid market and therefore will not realize any of the liquid upgrade. The gas processor is exposed to the margin created between the gas market and the liquid market. Historically this margin has been volatile, therefore there are swings between highly profitable operations and negative net margins. Also the processor typically bears all of the cost for field fiel, line loss, flare, and plant fuel.

Enron has performed most of its processing using this type of contract str~cture.(~.'~)

A flat rate contract is similar to a keep-whole agreement in that the producer is not exposed to the liquid market by being paid for all the Btu's delivered to the plant times a certain percentage (40-85%) of the sales gas price. This reduction in gas price is attributed to the processing and conditioning required to make the produced gas marketable. The variation in percentages is also a function of local gas processing competition.

2. Flat Rate Agreements.

The gas processor takes custody of the gas and then bears all of the gas processing risk. This company has all the benefits of liquid upgrade, along with risk that margins may be squeezed. The processor bears all of the cost for field fuel, line loss, flare, and plant fuel. The difference in this costs is that they are lower than the costs under a "keep-whole" agreement.

Some producers/processors, such as Mitchell Energy have offset the risks of keep-whole contracts (or flat rate contracts) through equity gas prod~ction.(~.'~)

3. Percent of Proceeds. The percent of proceeds contract takes the "pie" after the produced gas is transformed into marketable products (natural gas and NGLs) and divides it between the producer and the processor. Both entities are exposed to the gas market and liquid market. Typically the gas revenue is split between the producer and processor with the producer obtaining a larger share. The liquid revenue is divide between the parties with the larger share going to the processor. An example may be a "85/50" percent of proceeds contract where the producer will receive 85% of the gas revenue and 50% of the liquid revenue. Again the splits are a function of contract negotiations and competition. Also a producer may have to pay a marketing fee for their share of the liquids. This marketing fee can be on the order of 5% of the NGL value.

@John M. Campbeii & Company


NGL MARKET FACTORS

One of the main differences in this type of contract is that the producer bears the liquid shrink cost. That means that the producer is paying for the field fuel, line loss, flare, and plant fuel. In exchange for these costs they get the potential benefits of the liquid upgrade.

GPM (formally Phillips) requires a secure NGL supply for its petrochemical operation, therefore it is willing to give up some of the liquid upgrade to maintain liquid production. Midstream players, companies without equity gas production, such as Western Gas Re- sources and the Pan Energy (formerly Associated Natural Gas, Inc.) built their companies on percent of proceeds contracts where they take 40-50% of the liquids and 10-15% of the

4. Processing Fee. As we have gone from keep-whole to percent of proceeds contracts the producers exposure to the liquids market has increased and the processors risk (along with the potential for rewards) have decreased. A processing fee is the most risk free agreement from a processor’s perspective. In a processing fee arrangement the producer pays a fee based on the volume (or Btu’s) handled by the processor. This is generally a flat fee based on the type of conditioning and processing required. If only gathering and compression service are required, then the fee will be lower than the amount required to gather, compress, sweeten, dehydrate, and process (NGL extraction) the gas. Also the fee is a function of competition.

Liquids which are extracted are up for grabs. Depending on the amount of processing fee the liquids may be retained by the producer, but in most arrangements the processor will take title to a portion (if not all) of the liquids.

This type of arrangement is more common in small isolated fields where there is limited competition and which is generally not part of an overall company strategy.

gas. (2.13)

NGL MARKET FACTORS The natural gas liquids market provides feedstock to the petrochemical and refining industries

and supplies fuel for residential, agriculture, commercial, power and transportation sectors. Under- standing the market factors are important in the proper selection of the NGL extraction technology. The disposition of liquids is a primary consideration in determining whether a dew point control facility is installed (recovering the minimum liquids to meet pipeline specifications) or if a high liquid recovery facility is constructed.

The following discussion reviews the gas processing markets, pricing, and transportation systems.

Gas Processing Markets 1.

4

Natural Gas. Gas processing economics, as shown later in this section, are based on the spread or margin between selling the components as gas or extracting and selling the components as a liquid. Therefore understanding the natural gas market is the first step to evaluating NGL recovery economics.

Natural gas transported in transmission pipelines is mainly composed of methane. Figure 2.1 shows the different uses of natural The two main uses for natural gas are fuel (residential, commercial, industrial, power generation, and transportation) and chemical manufacturing feedstock. Worldwide, the primary chemicals manufactured from natural gas are methanol and urea (fertilizer).



2.1 9

OVERVINY OF GAS PROCESSING INDUSTRY

~ ____

Natural Gas

Propane Butane

&heavier I

I * Residential Methane

Hydrogen Sulfide

PARTIAL O X D A TlON

Sulfur

FUEL

~

Monoxide Chloride

etc. Urea JFischer Tropsch] Formaldehyde

Resins

Butyraldehyde

I [OXO ~rocess] I

Figure 2.1 Uses of Natural Gas

2.

Natural gas prices are seasonally cyclic in North America since fuel use is the largest demand. Figure 2.2 shows the U.S. natural gas spot ~rice.(~. '~) This graph indicates that the Rocky Mountains and Oklahoma areas generally have below average gas prices, Texas prices are similar to the average and Louisiana prices are above average.

There have been a number of changes which have impacted the U.S. gas market during the past decade. This is a complicated subject which will not be covered in detail in this manual. Some of the major impacts include deregulation of natural gas prices, open access to gas transmission pipelines (FERC Order 436/500), unbundling of gas transmission pipeline rates (FERC Order 636), and development of a spot market and futures market for natural gas. Ethane. Ethane is the lightest NGL component. It has one end use - petrochemical feedstock for the manufacture of ethylene. The demand for ethane is a function of the demand and price for petrochemicals and the price of competing feedstocks (heavier hydrocarbons). Since there is only one market for ethane, most ethane may be sold under long term contract to a petrochemical buyer. Ethane or "ethane rich" mixed NGL streams (EPBC) can normally be transported only by pipeline. This is because ethane exists as a liquid only under very high pressures 4700 kPa [800 psi] or extremely low temperatures -90°C [-13O0F]. Purity ethane is often transported as a "dense phase." Therefore, the

@John M. Campbeli & Compauy

9

3


NGL MARKET FACTORS

4.5 - 4.0 -

3.5 - h

3 m 3.0 -

o .- a 8 2.0 - 0 E 2 1.5 - m z

-

1.0 - 0.5 -

8 8

Figure 2.2 Natural Gas Price ($/MMBtu) - 1985-1998 \

existence of transportation networks is vital to the decision to recover ethane. This NGL transportation system will be discussed in more detail below. Figure 2.3f2.I4) shows the markets for ethane.

The two major U.S. liquid market centers are Mont Belvieu, Texas and Conway, Kansas. The ethane prices are shown in Figure 2.4.f2.16) There is usually a differential between the liquid prices in these two centers, with the higher price generally found on the Texas Gulf Coast, the location for the majority of the petrochemical facilities.

3. Propane. The market for propane liquid is divided between petrochemical feedstock and fuel (residential, agriculture, commercial, and transportation). Propane is oRen called LPG (Liquefied Petroleum Gas) but the term LPG may also refer to propane-butane mixtures or a predominately butane stream. Figure 2.5 shows the chemical derivatives for pr~pane.(~. '~) Its feedstock use is essentially limited to olefin manufacture (ethylene, propylene).

As a fuel, its most important use is in rural areas or where a natural gas distribution system is not in place. The rural users are both residential and agriculture. The agriculture uses are crop drying (corn, tobacco), irrigation pump drivers, and farm equipment. In some areas propane is processed with air and sold as "town gas." The fuel demand tends to be highly cyclical (high in winter, low in summer). Large storage volumes (normally underground salt domes or mined caverns) are required due to the cyclic demand. When recovered as a separate product, propane is normally transported by pipeline, tractor trailer, or rail tankcar.

Like ethane, there is a differential between Mont Belvieu and Conway propane prices with lower Conway prices. This information is included in Figure 2.6.f2.16)


Technical Assistance Service for the Design, 2.21 Operation, and Maintenance of Gas Plants


Polyethylene Eihyiene Eihylene Eihyiene Ethyl Oxide Chlorohydrin Dibmmide Chloride

ETHANE

ETHYLENE

myi Ethyl

Alcohol BenZC?ne

40

35

30

25

m P -2 20

& IL) c 15 m 6 W

a o .-

10

5

O

Year

Figure 2.4 Ethane Price ($/gal) - 1982 to 1998

3

3

@John M. Cempbeii & Company


4

Ally( nButyr-

Chloride aldehyde

NGL MARKET FACTORS

Curnene Propylene

aldehyd Alwhol Oxide Polymers

iso-Butyr- isoPropyl

I I N E L Rural

Home Heat PROPANE

Cracking

PROPYLENE

Agriculiure

Commercial

Figure 2.5 Uses of Propane

c c m a 7 1

l l 8 8 8 m 5 s 7 7 7

Figure 2.6 Propane Price (#/gal) - 1982-1998




Non9 Alcohol

4. & 5. Butanes. The market for butane is primarily petrochemical, gasoline blending, and fuel. Examples of butanes uses are shown in Figure 2.7.f2.I4) The end use for each isomer differs greatly.

Isobutane is the most volatile isomer and the most valuable. Its primary use is as a refinery blendstock for the manufacture of high octane components for gasoline. Isobutane can converted into methyl tertiary butyl ether (MTBE) or fed to an aklyation unit to produce premium gasoline blending materials.

Normal butane is an important feedstock for the manufacture of monolefms (ethylene, propylene) and diolefins (butadiene, which is used in the manufacture of synthetic rubber). Normal butane may also be isomerized into isobutane. When normal butane is used as a fuel it is normally blended with propane but it can be used as a pure component. The largest use for butane is as a gasoline blending component for octane and vapor pressure control. As a result of their important gasoline and feedstock uses, butanes are normally fed to refineries.

Pricing for i-butane and n-butane are included in Figures 2.8 and Figure 2.9 respectively.(2, “1

ücdecyl Lube Oil Synthetic Methyl-Ethyl Sec-Buts Nylon

Mercaptan Additives Resins Ketone Xanthate

.:igure 2.7 Uses of Butane



3

3

NGL MARKET FACTORS

EO

70

60

= 50 .

0

al o -

40 a> c m 3 m 30 c

c

20

10

6 r

lieu --- conway L Figure 2.8 n-Butane Price (#/gal) - 1982-1998

d m c 7

in co C

-J

W

C

7

m os 7

m C

7

m m m C

-3

O

C

7

m - m m C 7

Year

Figure 2.9 ¡so-Butane Price (#/gal) - 1982-1998




1 Propylene

6. Natural gasoline. Natural gasoline is a mixture of pentanes and heavier hydrocarbons. It may also contain small amounts of butanes. This product is sometimes call condensate. Natural gasoline is almost always shipped to a refinery either as a separate stream or mixed in a crude oil stream. Vapor pressure control is important when it is blended with crude. Recently it has become economical to feed natural gasoline to olefin crackers for petro- chemcial manufacturing. The uses for this product and pricing is shown in Figures 2.10 and 2.1 1(2.’6) respectively.

1 2 Butylenes Butadiene fi Ethylene

r--

_--------I Gasoline I

1 Benzene I I Toluene I 1 Xylene I

- * Refining

Reformer Isomerization

Gasoline -{ Blending I Figure 2.10 Uses of Natural Gasoline

e nr 360-

o Y

2 50 - <I> c O 2 40 a E + 30 -

._ - -

- 3

2 20 -

10 -

0 - m m ;

Figure 2.11 Natural Gasoline Price ($/gal) - 1982-1998



3

r)

NGL MARKET FACTORS

NGL Transportation Systems

A key to recovering and selling NGLs is the transportation, storage, fractionation and distribution network. This systems is one of the key factors in the growth of the industry. Without this network processing economics would be significantly altered due to the seasonal demand for most products. Following is a summary of the transportation methods and storage techniques.

1 . Trucks. Trucks are used to collect mixed NGLs for delivery to a fractionator and for delivery of product to the end user. Tractor trailers with 10 O00 gallon capacity are typically used to transport a propane plus (C,+) NGL stream for Eractionation and to transport finished products. The maximum true vapor pressure for this type of transportation is typically 208 psig @ 100°F. The transportation cost will vary between carriers and regions but the cost is typically prohibitive at distances greater than 250 miles (approximately 6 cents per gallon).

Bob tail trucks are used for delivering LPG to residential users.

2. Railroad. Since the 1860’s when wooden tubs were mounted on wooden flat cars, railroad tankers have been used to move petroleum products. Railroads were used to transport NGL products since the birth of the gas processing industry. One of the hazards of rail transportation resulting from “wild” product produced without stabilization to control the vapor pressure. A number of explosions occurred during this period. Railroads were the main method of transportation until the construction of the pipeline network. Today railroads continue to play an important part in the distribution of NGL products. The capacity of tank cars vary from 4000 gallons to approximately 45 O00 gallon jumbo cars. However, since 1970, the maximum capacity of new DOT specification tank cars containing regulated hazardous materials is limited to 34 500 gallons or not more than 263 O00 pounds gross weight.(2.20) I am not aware of mixed NGL being transported to fractionators.

3. Barge. The Mississippi River, Ohio River, Gulf of Mexico Intercoastal Waterway, and other navigable waterways are another means for NGL transportation. Butane and natural gasoline are the products transported by barge. This method of transportation is for finished product only. The cost is generally less than railroad rates and more than pipeline fees.

4, Pipelines. By far the most important method for transporting and distributing NGLs in this industry is the pipeline system. This network’s growth began in the 1950’s and along with the development of underground storage, provided an outlet for increased NGL production. Figure 2.1 2(2.’7) shows the major pipelines in North America.

The first pipelines put into operation were crude oil lines which batched butane or propane. Mid-America Pipeline Company (MAPCO) completed the first common carrier pipeline system from the Permian Basin through the Panhandle and Hugoton fields to Conway, Kansas. From this point the pipeline reached into the upper Midwest states of Minnesota and Wisconsin. Today the pipelines drain raw NGL production from the Rocky Mountains and Permian Basin into Mont Belvieu (MAPCO, Phillips, Chevron, Chaparral, Mobil and others) and Conway (MAPCO, Phillips, and others). There are also north-south pipelines (Koch’s Sterling Pipeline and Chevron Pipe Line Co.) which connect these marketing ten- ters. From Conway, MAPCO and Phillips transport liquid products into the Midwestern market. There are other pipelines which move product into this region which are not shown in Figure 2.12 (such as Conoco’s Cherokee Pipeline). From Mont Belvieu the two main product pipelines are TEPPCo’s pipeline to Ohio, Pennsylvania and New York and the Dixie Pipeline to Hattiesburg, Mississippi and Apex, North Carolina.




@John M. Campbeil & Company

2.28 BP Exploration Company (Columbia) LM.

3

3

NGL MARKET FACTORS

The combination of storage facilities and pipelines led to the development of two main North American NGL marketing hubs: Mont Belvieu, Texas and ConwayíHutchinson, Kansas. Minor focus areas are Hobbs, New Mexico and Borger, Texas.

Canadian NGL production focuses on the marketing centers of Edmonton, Alberta and Sarnia, Ontario. Amoco's Cochin Pipeline gathers production from the Waterton area and North to the Edmonton fractionators and delivers finished products to the Great Lakes area. There are other systems which gather NGL production to feed the Edmonton fractionators (Cochrane-Edmonton Pipeline-CoEd, Peace Pipeline, and Amoco Pipeline). The Alberta Ethane Gathering System (AGES) is not shown in the figure, but it is a 570 mile pipeline system to gather straddle plant ethane and deliver it to the Alberta Gas Joffre Ethylene plant and Fort Saskatchewan for upgrading to chemical grade ethane and shipment down the Cochin Pipeline and for miscible floods. Raw NGLs are batched from Edmonton down the Interprovincial Pipe LineLakehead Pipeline to Sarnia. Also products are shipped from the Alberta straddle plants to Winnipeg.

Typical transportation and fractionation fees (T&F fees )are shown in Table 2.8(2.'8)

TABLE 2.8 Typical Regional Transportation and Fractionation Fees 1994 ($/gal)

v a r k e t Center I Transportation I Fractionation Permian Basin

Mid-Continent Louisiana Gulf Coast Texas Gulf Coast(') Rocky Mountad2)

ArkLaTex

Mont Belvieu

Conway The River

Mont Belvieu Mont Belvieu

Mont Belvieu

2.25

1.25

1.25

2 .O014 S O

3.9517.55

3 .O0

2.35

1 S O

2.35

2.35

2.35

2.50

('I The Southern Texas Gulf Coast suffers an additional 2.0 to 2.5 $/gal transportation cost (2) The 3.95 $/gal is an average fee for liquid production from Northwestern New Mexico and South Central Colorado (San Juan Basin). The 7.55 $/gal fee applies from Southern Wyoming and Eastern Utah (Overthrust).

5. Ocean going Ships. In the mid 1940's Warren Petroleum began hauling propane from Houston to Newark, NJ. This ocean going ship could carry 35 O00 barrels of product. Today waterborn LPG shipments are critical to the international gas processing industry. A typical long haul LPG carrier has four or five tanks and carries 350 O00 to 550 O00 barrels of product. The product is refrigerated to reduce the vessel design pressure. Figure 2.13 shows the trade routes for waterborne LPG shipments.(2.19)

6. Storage. NGL feed and products are stored in either aboveground or underground. Stor- age may be located at remote gas plants, fractionators, and market centers. The purpose of the storage is different for these locations.

At a remote gas plant, storage serves to provide enough surge to prevent shutdowns due to problems in the delivery system. If product is moved by tractor trailer or rail car, then enough storage must be provided to maintain production between tankers. For tractor trailers this means that the there should be at least four days of storage due to long weekends, holidays, or weather delays. For rail car deliveries the amount of storage (whether in tanks or in tanker spots on the rail siding) must be sufficient to handle plant production



2.29


@John M. CPmpbeU & Comppnv a .


NGL MARKET FACTORS

between switches. Only surge (8-24 hours depending on the volume) capacity is required when delivering NGL into a pipeline.

Fractionator storage may have the same considerations as a remote gas plant and a marketing center. If there is the capability to have underground storage, then storage may be supplied for seasonal swings, otherwise storage capacity will be exclusively dependent on the delivery system.

Market centers have large underground storage which provide capacity for seasonal swings. For example, propane fills caverns during the late spring and summer, then during the fall crop drying and winter heating season these stocks are withdrawn. Underground storage allows the industry to meet demand swings (ratio of fall/winter demand to spring/summer demand) which may be greater than 61.

Aboveground storage can be in bullets, spheres, or refrigerated storage. Standard propane bullets have capacities of 10 000, 20 000, 30 000, 42 000, and 60 O00 gallons. The design pressure is typically 250 psig @ 100°F. These standard propane storage tanks cost approximately $1.00 per gallon new and $0.50 per gallon used. If an EPBC mix is stored, then the propane tanks can not be used due to the higher vapor pressure. Most companies allow a maximum of 300 O00 gallons storage in a given area. If more storage is required then the tank batteries are separated by at least 50 feet.

When storage volumes exceed 100 O00 gallons, then spheres and refrigerated storage should be evaluated. Spheres are more common with lower design pressures (i.e. butanes) or with refrigerated storage. Refrigerated storage has the benefit of lowering the vessel design pressure by cooling the product through direct refrigeration or indirect refrigeration.

If the storage exceeds 4 million gallons, then underground storage caverns are typically economical. The product is stored in either a solution mined cavern or a conventional mined cavern. A solution mined cavern is constructed by drilling a well or wells into a salt dome and circulating low salinity water over the salt to dissolve the salt as brine. Mined caverns can be constructed in any non-porous rock at adequate depths to withstand product pressures. Typical capital costs for solution washed caverns is approximately $2.00/bbl. The total U.S. underground storage capacity is approximately 5 19 million

Fractionators

The bulk of the NGL fractionation is performed in Mont Belvieu and Conway. Table 2.9f2.18) summaries the major fractionators in these locations. There are numerous other fract trains located in the U.S. Small region fractionators are located in Wyoming, Colorado, New Mexico, Ohio, West Virginia.

The bulk of the Canadian NGL fractionators are located in Edmonton and Samia.

4



2.31


TABLE 2.9 Fractionation Capacity, bbl/d 3

Mont Belvieu:

Coastal States(') Enterprise Products Gulf Coast Fractionators

Trident NGL

Warren Petroleum(*) Subtotal

Other Texas:

Mobil - Beaumont

Phillips - Borger

Phillips - Sweeney Vaiero - corpus Christi(') Subtotal

Kansas/Oklahoma:

Enron - Bushton, KS MAPCO - Conway, KS

Trident NGL, Hutchinson, KS Koch - Medford, OK Subtotal

Capacity

33 700 165 O00

120 O00

82 O00

200 O00

604 O00

105 O00

95 O00

95 O00

50 O00

345 O00

91 O00

107 O00

44 O00 150 O00

392 O00

(')De-ethanized mix (*)Includes 30 O00 bpd dedicated to propane, butane, or mixed butane imports

NGL RECOVERY ECONOMICS Liquid hydrocarbons are recovered from natural gas for three reasons: 1. NGL Product Recovery

NGL recovery may be desirable if the market value of the various NGL components in liquid product is greater than the sum of:

a. their equivalent heating value in the natural gas stream, b. the cost to fractionate and transport the NGLs to market, and the cost to build and

operate a NGL extraction plant.

NGL recovery under these circumstances is considered discretionary (i.e., a recovery plant would not be built if it could not generate an adequate rate of return on invested capital).

2. Maximize Liquid Production

In many cases, no immediate market for the natural gas exists. Examples include solution gas from remote crude oil reservoirs such as Prudhoe Bay, or retrograde gas condensate



NGL RECOVERY ECONOMICS

reservoirs where the gas is reinjected for pressure maintenance, i.e. cycling projects. In these cases, NGLs are frequently recovered and spiked into the crude or condensate stream in order to maximize liquid production. NGL recovery levels are set by the liquid product vapor pressure specification. For a TVP vapor pressure specification of 8-15 psia, NGL recovery is typically confined to the Cq+ components. This is strongly dependent upon the relative volume of NGL to crude or condensate. NGL recovery under these circumstances is also considered discretionary.

3. Hydrocarbon Dewpoint Control

Virtually all gas sales contracts include a clause which limits the hydrocarbon dewpoint of the gas. Typical values range from 14 to 41°F [-lo to 5"C]. In order to meet this provision of the contract some type of NGL extraction is generally required. In many cases, the NGLs will be recovered for discretionary reasons, but in some cases the only justification for the NGL plant is to meet the gas sales specifications. In these instances, NGL recovery is mandatory since the gas could not be sold without removal of the heavier NGL components. Recovery levels in these facilities are usually minimal with only a portion of the Cg+ components being extracted.

The construction of a plant to recover NGL products is economically justified if the cash flow generated by NGL recovery is incrementally more attractive than the cash flow resulting from the current or future sale of those recoverable components as part of the natural gas stream.

The economics of gas processing entails the margins, operating costs, and capital costs. These items will be discussed below.

\

Margins

Margins describe the revenue side of the plant economics. Depending on the type of contracts, there may be gas, liquid, and/or processing fee margins (or revenue).

The gas margins and processing fees are straight forward. The calculation of these revenue streams are describe in contracts and will not be discussed in here.

The liquid margins can be more difficult.

If the processing contract is a percent of proceeds type, then the liquid margin is calculated by splitting the liquid revenue received from the product sales (at the plant) as established in the contract. The plant product sales price is sometimes called the netback price. This price is the product price based at a market center (Mont Belvieu, Conway, etc.) less the transportation and fractionation fees to move the NGLs to the market and separate the liquids into final products. This netback price is the products value at the plant tailgate. Transportation and fractionation fees are shown in Table 2.8. In addition, processors may charge producers a marketing fee for the disposition of their NGLs.

In keep-whole or flat rate contracts, the margins are based on the liquid revenue less producer settlements calculated for the plant shr ink. To calculate the liquid shrink, the gallons of each product is converted to a gross heating value. These conversion factors are shown in Tables 2.7(a) and 2.7(b). This calculation is demonstrated in Example 2.3.

4 From these values we can also determine the breakeven price relationship between recovery and rejection of individual components. This is shown in Figure 2.14. This breakeven analysis does not include transportation and fractionation fees and operating costs.



2.33


xample 2.3: What is the liquid s h r i n k from a 100 MMscfd turboexpander with the followinj inlet gas composition and recoveries.

A

Inlet Gas (moiX)

0.83 0.18

88.88 5.74

2.49 0.73

0.62 0.14

0.09

0.12 0.11

0.11 ~

100.00

B

Recover ("/.I

25.0

0.2 84.9

92.3 99.5

99.7 99.9

100.0 100.0

100.0 100.0

Volume (scf/gal)

58.807

59.135 37.476

36.375 30.639

31.791 27.380

27.673 24.379 2 1.725

19.575

Liquid Rec. (GPD)

765

3006

130 285 67 290 23 707

19 444

5108 3252

4922 5063

5619

268 461

O

59 728

65 869

90 830 98 917

102 913 108 754

110 084 115 061

118 623 121 393

O 1792

8582 61 12

2345 200 1

556 358 566 608

682

23 595

'herefore, the liquid shr ink from recovering 268 461 GPD of NGL is 23 595 MMBtdday. Noticl iat the volume (scf/gal) is referenced to standard conditions of 60°F and 14.696 psia. If the bas1 onditions are different from standard conditions, then column C can be corrected by the followinj quation.

Where: CB = volume corrected to base conditions Cs Ps = standard pressure = 14.696 psia Ps = base pressure from contract, psia Ts = standard temperatureOR = 460+60"F TB = base temperature from contract, "R

= volume at standard conditions (60°F, 14.696 psia) &om Table 2.7(b)

Jso notice that the gross heating value can be calculated based on fuel as liquid or as gas. Th ifference between these values is the latent heat of vaporization. Using the gross heating value fo iel as a gas, benefits producers in a keep-whole contract while processors are benefited by fuel a quid GHV values. In this example the effect is $87 M/yr at $1 .5O/MMBtu shrink value

lotes: ('1 "C" = from Table 2.7(b)

1OOx 106scf (B) gal (*) "D" = [%I( day )(loo)[ (C)scf)

(3) "E" = from Table 2.7(b)


3

3

")



70

60

50

$ 3 -

40 \ u> c a, o

c

.- 30 a 6 Z

20

10

O O 1 2 3 4 5 6 7

Gross Heating Value Price, $/MMBtu ~~

Figure 2.14 Shrinkage Value of NGL Products Based on Fuel as Ideal Liquid

Gas processing economics for discretionary processing is based on the spread between natural gas price and liquid prices. Figure 2.15 provides a historical perspective on this relationship and the relationship between NGL prices (Mont Belvieu) and crude prices (West Texas Intermediate - WTI). The composite NGL prices are based on the data from Gas Processor Report. This shows that the ratio of NGL market prices (on a heating value basis) relatively to natural gas has typically ranged between 1.50:l to 2.00:l. This relationship also indicates that the two markets are independent based on the wide swings in this ratio. NGL and WTI pricing tends to track together. NGLs generally run between 60% and 75% of crude. This is very close to the differences in gross heating value. Crude has approximately 5.8 MMBtu per barrel while NGL have approximately 4.0 MMBtu per barrel.

Operating Costs

Operating costs have been the focus of many bench marking studies during the early 1990's. The McKinsey study(2.21) provided a comprehensive study looking at the cost for each U.S gas production basin. This information is shown in Figure 2.16.

Another operating cost estimation is shown in Figure 2.17. This tool provides a correlation for cost based on throughput. This graph shows the McKinsey study and Purvin & Gertz (2.2) along with actual plant data. As can be seen from this graph, costs can be all over the place. Factors such as company philosophies, plant configurations, plant utilization, related services (gathering, compression, treating, etc.), contract terms, electrical costs, etc. impact the operating costs. Therefore the best method for determining the operating cost is from actual operations or from a budget "build-up."

-4

O John M. Campbell & Company


2.35


I

450

400

250

0 200

150

1 O0

0-

h

c Q LT

0.10

IC m

11%

Year

7% 2% 12% 1 Yo 1 Yo 1 Yo 42% 4% 19%

Figure 2.15 Comparison of NGL Pricing with WTI Crude and Spot Natural Gas (Mt. Belvieu)

1 .o

0.8

0.6

0.4

0.2

0.0

Vol u me, YO

Location

1 .O6

0.79

+-- 0.04-C

L

0.17 16 L L

( -

0.26

i

Figure 2.16 Regional Gas Plant Direct Costs ($/Mcf Throughput)

@John M. CampbeU & Compeny

2.36 BP Exploration Company (Cdumbla) Ud.

3

"1


1 .o0

c o . 5 c u)

0 0.10

o"

m

9 c .- c a,

0.01

Figure 2.17 NGL Plant Operating Costs

The major costs for operating a gas plant are listed below:

1. Fuel and Electricity.

The fuel and power costs are usually the most significant cost to operate recovery processes. This cost is typically associated with compression (inlet, refrigeration, residue, recycle, etc.). Compressor power calculations can be performed through a number different methods as discussed in Gas Conditioning and Processing - Vol. 2. An approximation of power requirements is the "Rule of 22."

hp = 2 2 x F x n x C R x Q ,

Where: hp = the compressor power F = correction factor

= 1.00 single stage = 1.08 two stages = 1.10 three stages

n = number of stages CR = compression ratio per stage Q,, = std volumetric flow (MMscfd)

The compression ratio per stage is calculated as follows

Where: Pd = Discharge pressure (psia) P, = Suction pressure (psia)

@Job M. Campbell & C 0 m p . n ~


2.37


The fuel usage is generally 8-9000 Btdbhp-hr for engine driven reciprocating compressors and 10-12 O00 Btuhhp-hr for turbine driven centrifugal compressors.

2. Labor.

The major consideration for labor cost is the operator coverage. This is primarily a company operating philosophy, but guidelines can be provided. There are four main types of shifts: 1) three 8 hour shifts, 2) two 12 hour shifts, 3) 8 hour attend (16 hour unattend), and 4) unattend. The first two shifts require four personnel per shift-operator. The fourth person is required to rotate for weekends, holidays, and vacation. The unattended operations, personnel usually spend less than 4 hr/day at the plant.

Unattended or 8 hour attend facilities are generally used for small gas plants that are not critical to a companies operation. These plants typically do not handle associated gas (gas produced from oil wells) and sour gases. These types of plants are not usually located near population centers. The product disposition can be by truck, rail, or pipeline provided the truck loading is key-stop (automated) or the liquid production is low. The increased use of SCADA for gathering makes this type of operation more attractive.

For large plants, most facilities are attended 24 hours per day. Multiple operators may be required if a significant amount of time is spent loading or unloading trucks and tankers. Also multiple units (gas treating, compression, utilities, etc.) may require multiple operators.

The cost of operations staff is dependent on too many variables to provide guidance in this manual.

Other staff which may be required for a gas plant is plant manager, engineer, account- adclerk, maintenance foreman, mechanical technicians, instrument technicians, and roust- abouts. The cost of staff is generally 1.30-1.45 times the annual salary. This covers legali- ties (FICA, federal unemployment insurance, state unemployment insurance) and benefits.

3. Maintenance.

The maintenance budget is another major cost area. It is generally divided into major maintenance and repair maintenance. Costs can be determined from actual plant operations or vendors recommendations. Compressors are a significant factor in this budget. For electric motor driven compressors, some people use $40-50/bhp-yr for budgeting purposes and $60-70/bhp-yr for other compressors.

4. Chemicals and Operating Supplies.

Chemicals may include lube oils, coolants, methanol, plant water, heat transfer fluids, desiccant, glycols, filter elements, etc. These costs may be estimated from typical usage and current pricing.

5. Other.

There are other cost associated with gas plant operations. These include royalties, taxes, insurance, general overhead and administration (G&A), etc.



Capital Costs

Figure 2.18 provides an order of magnitude estimate for plant costs(2.2’. The assumptions for this cost data are listed below.

Included: 10% Contingency Start-up Costs at 2 months of operating expenses Initial start-up supplies and minimal spare parts Sales taxes Direct costs Single product storage

Land Home office costs Interest during construction Construction insurance and/or bonds Gathering, inlet compression, treating, fractionation

Excluded:

It must be emphasized that these costs are order of magnitude and not +/- a certain percentage. The capital costs are dependent on factors beyond the purpose of this manual, but this information provides a feel of plant investments. The key to developing realistic cost estimates is defining the scope of the project.

\

100.0

10.0 5- 5 Y?

o c v) O

m a

- c ._

o 1 .o

1 O 0 0 0.1

1 10 100 Inlet Gas Capacity (MMccfd)

Figure 2.18 NGL Plant Capital Costs (1993)

@John M. CampbeU US Compnny

Technlcal Assistance Service for the Design, Operation, and Maintenance of Gas Plants

2.39


Economic Summary

Figure 2.19 provides a generic analysis of gas processing economics for 1985-1995. These economics are based on a keep-whole contract. From the gross margin, the transportation and fractionation fees, fuel and shrink costs, and operating cost are subtracted to determine the cash margin. The “tough” years for gas processing were in 1986 and 1993. The major cause for these poor economics were higher shr ink costs resulting from high gas prices. The good times occurred in 1995 and during the Gulf war (1990-1991). This analysis is similar to the Oil and Gas Journal’s Wright-Killen Gas Processing Margin with the following exceptions: 1) Gas Processor Report composite NGL pricing was used, 2) Oil and Gas Journal’s Natural Gas Spot Price (average of all points) was used for gas prices, 3) T&F fees were held constant at 4.35 cents per gallon, 4) shrink costs were based on Gas Processor Report composite NGL analysis, 5 ) fuel cost were calculated based on plant simulations assuming a 100 MMscfd plant and 6) operating costs were based on Figure 2.17 assuming a 100 MMscfd plant.

50

40

- ell 3 30 i. 9 g 20 O

oc

10

O

Trans & Frac Fee Fuel & Shrink

d - V i n i n < o < o h b co ? ? ? ? ? ? 9 ? ? 7 7 7 - 7 2 = 5 = m z $ < < c - C

Qgure 2.19 NGL Plant Operating Economics



TYPES OF PROCESSES

TYPES OF PROCESSES There are basically three’ general types of processes(2.’2) employed for NGL extraction, dehy-

dration, and sweetening. These are: 1. Absorption 2. Adsorption 3. Condensation

This manual covers only NGL extraction in detail in the later sections. Other courses and materials are available for gas sweetening, dehydration, fractionation, and liquid sweetening. In order to understand the entire system, the following discussion is presented for each process in the context of its application.

GAS SWEETENING Sweetening is a term generally applied to a group of processes which are used to remove H2S,

C02 and/or other sulfur components from a natural gas or NGL stream.

Sales gas specifications for H2S content are very low. This is due to the fact that H2S is extremely toxic. Even at very low concentrations (less than 100 ppmm) H2S can cause sickness and, in some cases, death. In addition, H2S is corrosive to carbon steel in the presence of water and can induce stress corrosion in carbon steels at high pressures.

C02 removal is not as critical as H2S removal. Examples of COZ removal include:

1. Upstream of turboexpander plants to prevent COZ freezing 2. Recovery of CO2 for commercial use (miscible flood EOR projects) 3. To meet sales gas specifications

A. Absorption Absorption is the most common type of process employed for the removal of H2S and COZ

from natural gas. (It is also widely used for the removal of H2S and COZ from NGL streams as well.) The absorption process may be chemical or physical. Several common solvents used in absorption type sweetening processes are listed below:

Chemical Absorption

1 . Amines (MEA, DEA, MDEA, DGA, DIPA, etc.) 2. Alkali Salts (Hot Potassium, Benfield, Catacarb, Flexsorb HP, etc.)

Physical Absorption

1. Selexol 2. Methanol-Rectisol 3. Water 4. Glycol 5. Propylene Carbonate (Fluor)

1 A fourth type of process is based on membrane technology. The use of membranes is currently confined to special applications such as the removal of COZ from natural gas and the separation of nitrogen from air. Membranes will not be discussed in this course.



2.41


Hybrid 1. Sulfinol 2. Flexsorb PS 3. Optisol

The difference between chemical absorption and physical absorption is that in chemical absorption a chemical reaction takes place between the component being absorbed (solute) and the absorption liquid (solvent). In physical absorption the solute simply dissolves into the solvent.

An example of a chemical solvent is amine. When monoethanolamine (MEA) is used to remove H2S from a gas stream, the following chemical reaction takes place.

Amine + H2S += Amine Sulfide + H20

3

This is a common reaction ... that of a base reacting with an acid:

Acid + Base -+ Salt + Water

Actually a number of complex reactions take place simultaneously, but the net effect is shown above. The salt that forms is not heat stable, so this reaction can be reversed by lowering the pressure and heating the solution. An example of a typical amine treating process is shown in Figure 2.20. This is a typical process flow scheme and differences will exist depending upon the solvent.

As a general rule of thumb the chemical solvents tend to be used when the partial pressure of the H2S is less than approximately 4.0 bar abs [60 psia] and when selective absorption of COZ is not required.

If the partial pressure of H2S is greater than about 4.0 bar abs [60 psia] then physical solvents are typically employed. Selective solvents such as Selexol are particularly attractive when coabsorption of C02 is not desired. Selexol is a proprietary solvent, that is a mixture of polyglycols. The acid gas components, H2S and COZ, are absorbed into the Selexol exactly like water into TEG, by physical attraction of the molecules.

Regeneration of physical solvents can be accomplished by heating the solution in a still column at atmospheric pressure, as shown in Figure 2.20, or by flashing the solvent to a low pressure and thereby vaporizing the acid gas from the solution.

B. Adsorption Adsorption systems may also be used for gas or NGL sweetening. Specially manufactured

molecular sieves will adsorb hydrogen sulphide as well as water vapor. There are a number of units which use this process (essentially the same as adsorption dehydration) to remove relatively small amounts of H2S from a gas stream. This process has the disadvantage that upon regeneration the H2S ends up in the regeneration gas and unless this gas can be used for fuel, it must be treated by another process (e.g., amine, Selexol) to get the total gas stream free of H2S. The secondary treating facilities are smaller, however, since all the H2S is concentrated in a smaller (10%) volume of gas. The adsorption process is a common method of sweetening NGL.

The iron oxide (iron sponge) process is also used to remove small amounts of H2S from the gas stream. The gas and oxide are contacted in a tower in which the material is held and through which the gas is routed. In this process the H2S in the gas reacts with the iron oxide (Fe2O3) to form iron sulphide (Fe2S3) and water. When exposed to air, the iron sulphide is oxidized back to iron oxide plus free sulfur. Theoretically, this reaction-regeneration cycle can be repeated several times until enough

@John M. Campbeü & Company


3

'7

GAS SWEETENING

@John M. Campbeli & Company


2.43


sulfur accumulates in the bed to cover up the iron oxide particles. At this time the vessel is opened and the bed replaced. This process is simple and relatively inexpensive but has the disadvantage of being "dirty" in that the residual iron sulphide and sulfur must be disposed of when the bed is replaced.

The zinc oxide process is similar to, but more efficient than the iron oxide process. This process is effective for removal of small amounts of hydrogen sulphide and will remove H2S down to well below a 5 ppm level. If large amounts of H2S must be removed the beds become prohibitively large and material cost prohibitively expensive. The beds are ordinarily designed to last 3 to 6 months before replacement is required. In some instances two beds are operated in series to assure that all H2S will be removed even if the first bed has some gas bypass due to channeling. Also, the second bed can be utilized while the first bed is being replaced.

")

C. Condensation Condensation techniques for gas sweetening are generally used for COZ removal. Two meth-

ods include RydHolmes and CFZ. RyaníHolmes is the only method with wide spread commercial use. Exxon developed the CFZ (Controlled Freeze Zone) process which performs a solid-vapor separation.

DEHYDRATION Dehydration is the removal of water from natural gas, NGL, condensate, or crude oil. The

water may be free water (as is often the case in crude oils or condensate) or it may be in equilibrium.

A review of natural gas dehydration processes follows.

I -e=

Figure 2.21 A Glycol Contactor

A. Absorption The absorption process, when applied to

natural gas dehydration, is typically referred to as glycol dehydration. Triethylene glycol (TEG) is contacted with the gas stream in a trayed or packed contactor tower at elevated pressure and ambient temperature. The TEG is hygroscopic; Le., it has a strong affinity for water. The water molecules in the natural gas are attracted into the liquid TEG. This process is called absorption.

Once the water has been absorbed by the TEG it must be removed so that the reconcentrated TEG may be recirculated to the contactor. This process of water removal is called regeneration and in the case of TEG takes place at high temperature, about 204°C [400°F], and atmospheric pressure.

Figure 2.21 is an example of a glycol contactor and Figure 2.22 presents a process schematic for an entire dehydration unit.

1 .



DEHYDRATION

B J o h M. Campbell & Company


I

- 2.45


Figures 2.21 and 2.22 are typical of most glycol systems. These will be used for an overview of glycol dehydration.

In Figure 2.22, wet feed gas enters the inlet separator for free liquids removal. From here it flows to the bottom of the glycol contactor as the lean glycol (TEG) solution enters at the top of the tower. The countercurrent flow of the gas bubbling up through the downward-flowing liquid on the trays or packing causes water vapor to be absorbed by the glycol. The now dry gas exits the top of the contactor and after passing through an outlet scrubber, flows on for further processing, as necessary. The diluted (rich) glycol solution collects in the base of the contactor.

From the contactor the relatively cool rich glycol flows to the tube bundle of a reflux condenser mounted in the top of the glycol stripper column. Here, the glycol acts as a condensing medium for some of the vapor flowing up the stripper column, thereby providing reflux for the top portion of the column.

From here the rich glycol flows through the contactor’s level control valve, taking a pressure drop as is enters the rich glycol flash tank, where dissolved hydrocarbons can flash to vapor and be removed. A liquid hydrocarbon phase may also form and be skimmed off. The rich glycol then passes through a sock filter, charcoal “filter” for removal of dissolved heavy hydrocarbons and then the downstream sock filter. Next, the rich glycol is heated in the leadrich glycol heat exchanger before entering the atmospheric stripper column of the regenerator.

Inside the stripper column, the rich glycol feed flows downward over the packing (or trays) and contacts the rising hot stripping vapor from the reboiler below. The water in the rich glycol vaporizes and joins the upward flow of stripping vapor. Some of the rising vapor condenses on the coils of the reflux condenser and rains down as reflux onto the top section of the stripper, cooling the rising vapor to prevent excess TEG vaporization losses.

In the reboiler, the glycol is heated to the temperature required to drive off the water. The hot glycol level is maintained by a weir or other device in the reboiler shell; excess solution overflows the weir and enters the top of a secondary stripper column. The secondary stripper provides more complete water removal as the hot solution contacts the rising dry stripping gas supplied from an external source. The hot glycol is now considered to be lean glycol; it collects in the lean TEG surge tank.

From the surge tank, the lean TEG is pumped through the opposite side of the leadrich glycol exchanger and then back to the contactor by the circulation pump (and booster pump if required by contactor pressure). Prior to entering the contactor, the lean glycol is cooled for more efficient absorption, but only to within 5 to 10°C [9 to 18”FI of the inlet gas temperature, so that no liquid hydrocarbons condense inside the contactor.

The factors that determine the extent to which TEG gas dehydration is effective are: 1. The glycol circulation rate

2. The inlet gas rate 3. The lean glycol concentration 4. The number of contacting stages in the contactor 5. The temperatures of the inlet gas and lean TEG 6. The contactor’s operating pressure

Glycol dehydration is generally the most economical method of removing water vapor from the gas stream for dewpoint depression of around 60” to 120°F. It is also the most commonly employed method, worldwide. Glycol dehydration using TEG is generally adequate for meeting gas pipeline

9 @John M. Campbell & Compmy


DEHYDRATION

\ water dewpoint specifications and the requirements of most downstream plant processes. For downstream plant processes operating at temperatures lower than about O"F, enhanced EG dehydration methods (e.g. stripping gas, vacuum, Cold Finger, or Drizo) an adsorption process is generally used. Drizo achieves the maximum dewpoint depression of 180 to 220°F.

B. Adsorption

Adsorption is the process by which water is removed from the gas by contacting the gas with a solid desiccant. The water molecules are held onto the surface of the desiccant by physical attraction. The most common dry desiccants are alumina, silica gel, and molecular sieves. Molecular sieves are the most widely used adsorption agents in recent years. They are manufactured as pellets or spheres with a diameter of 2-3 mm [1/16-1/8 in.].

A process flow diagram for a typical adsorption system appears in Figure 2.23. This shows a two-bed system with one bed in operation and the other in regeneration. The wet gas enters the system by first passing through an inlet scrubber where free liquids are removed. Some systems then employ the solids or coalescing filter, as shown. In the diagram, the wet gas then enters the top of adsorber (dehydrator) number 1. The gas flows downward through the bed of adsorbent material and is stripped of its water contaminant. The dry gas then flows to the outlet filter where desiccant dust is removed prior to transportation or further processing steps. This bed will remain in operation until saturated with water.

Adsorber number 2 is in the heating portion of its regeneration cycle. A supply of dry regen-

creased to that required by the adsorbent type. (Sometimes as high as 350°C [660"F]). This hot gas is routed to the bottom of adsorber number 2. The high temperature gas drives the adsorbed water off the surface of the desiccant. The liberated water vapor joins the flow of regeneration gas and leaves the top of the dehydrator tower. The wet regeneration gas enters a cooler where the bulk of the water vapor is condensed. The regeneration gas is then scrubbed and returned to the plant as a recycle stream or is used as fuel.

\ eration gas (often the filtered dry outlet gas) is passed through a heater where its temperature is in-

This process can achieve water dewpoints of the following:

0 Silica gel -70 to 80°F 0 Alumina sieve -100°F 0 Molecular sieve -150°F

Water removal using molecular sieve is essential if the gas is to be processed in an LNG plant or a "cryogenic" gas plant.

An adsorber containing molecular sieves will have a desiccant volume equivalent to around 10 kg of sieves per kg of water to be removed (or 10 lbs of desiccant per lb of water to be removed). Two, three and four-bed systems are utilized, depending on which results in the lowest costs. The operation of the beds is set on a time cycle with automatic switching of valves. Eight to twelve hour cycles are common.

The desiccant material will deteriorate over time and a three-year replacement period is normal. It is essential that any foreign material (well treating chemicals, heavy hydrocarbons, dirt, etc.) is removed upstream of the beds. Heavy liquids will be adsorbed by the sieves and can form coke upon regeneration. Coke and dirt can drastically reduce the efficiency of water adsorption.

\

@John M. Campbcil& Company


2.47


@John M. CPmpbeU & Company

2.48 BP Exploration Company (ColumMa) Ltd.

3

DEHYDRATiON

C. Condensation If the natural gas is to be refrigerated for NGL recovery, the water will condense out of the gas

due to the cooling process just as the hydrocarbons do. However, if the gas temperature is dropped below the hydrate point, the water will solidiSr into hydrates which are similar to ice. Hydrates will plug off the gas flow through the exchangers, valves and piping, so their formation must be inhibited.

1. Glycol injection

Hydrates can be prevented by removing the water from the gas stream using one of the two dehydration processes previously described (absorption or adsorption), and this often is done. How- ever, if the gas is to be refrigerated it often makes sense to allow the refrigeration system to condense the water (thereby removing it from the gas) and to prevent freezing by injecting an inhibitor into the gas stream. The most common inhibitors used in gas processing are methanol and ethylene glycol (EG). Ethylene glycol is preferred in most systems because it can be recovered, regenerated and reused. Methanol can be recovered by it is generally uneconomical. Also methanol vapor losses are significant.

A glycol injection system is typically the most economical method of dehydrating natural gas if moderate levels of refrigeration are applied to the gas stream for NGL extraction. This is because no contactor towers are required. The EG is contacted with the gas by spray nozzles or static mixer elements. This results in a considerable savings over other methods.

Figure 2.24 presents a typical EG injection system for a natural gas process employing refrig-

of the gas-to-gas exchanger and the chiller. The injection point is usually immediately upstream of the exchangers, and spray nozzles or static mixers are used to ensure good mixing.

4 eration for NGL removal. The EG solution (typically 70-80 wt% EG) is injected into the gas upstream

The gas, condensed hydrocarbons and rich glycol solution enter the 3-phase cold separator. The gas passes through a mist extractor and heads back to the shell side of the gadgas exchanger. The liquid hydrocarbons float on top of the glycol solution, overflow a weir and dump out on level control. The glycol solution collects in the "boot" section of the separator where an liquid interface level controller maintains the proper level. The rich glycol dumps from the separator, is regenerated to the proper concentration, is cooled and then reinjected into the process gas stream. The regeneration unit in an EG injection system is quite similar to that used in a TEG dehydration unit except that the reboiler temperature is much lower (125°C or 257°F) and no stripping gas is required.

The primary operating problems in glycol injection systems are high glycol losses due to poor separation in the glycol/condensate separator and poor regenerator operation due to hydrocarbon canyover with the rich EG. Also, the control of glycol concentration is critical in low temperature systems (less than -30°C [-22"FI). Minimum operating temperatures are -40°C [-40°F] due to increased EG viscosity which causes difficulties in EG-NGL separation.

2. Methanol/lFPEXOL

IFP has recently touted a "new" NGL extraction technology called IFPEXOL. This process was demonstrated in Petro-Canada's East Gilby Plant in 1992. The IFPEXOL technology is really a hydrate inhibition system versus NGL extraction technique. The unique feature of IFPEXOL is the regeneration of rich (water) laden methanol. Figure 2.25 provides an example of this process. This plant used a refrigerant to cool the produced gas stream to -35°C [-35"F] and condense both liquid hydrocarbons and water, Hydrate formation is inhibited by the regenerated methanol in the IFPEXOL

4



2.49


w a E L o 8 $ - e I5 2

I- o)

z

gJ t W a:

I

I I I I I 1 I I I 1 I I

J ' f

íci

I



3

3

DEHYDRATION

* I I

I

L

d c

:I O J o h M. Campbell & Compauy


2.51


applications for each of these tech- nologies: Ethane

Propane Butanes Pentane'

contactor and the make-up methanol. The methanol absorbs the condensed water. This rich methanol stream is regenerated by using the feed gas to "strip-out" the methanol leaving the water to condense in the bottom of the IFPEXOL contactor. Any method to cool the inlet gas can be utilized including mechanical refrigeration, turboexpansion or valve expansion. Six operating plants have used this technology with capacities varying from 8-320 MMscfd and operating temperatures of -30 to -150°F. The benefits of this technology include to BTEX emissions, no heater emissions, potential for combining the dehydration and sweetening units and potentially reduced operating and capital costs. The main disadvantages include methanol contamination in waste water, methanol contamination in the NGL product and methanol losses in the residue gas. The first two items can be addressed in facility design as shown in Figure 2.25. To reduce methanol in the waste water more stages in the contactor and a higher stripping gas rate are used to get the methanol content down to 40-100 ppm. To remove methanol from the NGL product a water wash may be employed. Overall methanol losses are approximately 2 galmscf. This value is dependent on the operating temperature. IFPEXOL is generally not competitive above -20°F.

SCU ROA Refrig JT Expander 20-40 50-60 60-95 70-90 30-60 80-90 90- 1 O0

10-20 100 60-90 90-100 1 O0 70-80 100 1 O0 1 O0 1 O0

NGL EXTRACTION OVERVIEW



NGL FRACTIONATION

NGL FRACTIONATION Once recovered from a produced gas stream, the raw NGL mixture must sooner or later be

distilled or fructionuteú into useful products. Fractionation may take place at the production facility or the recovered NGL may be sent to a central fractionator located some distance from the field. NGL fractionation may consist of from two to four separate steps, depending upon whether or not ethane is part of the mix and if butanes will be split into isobutane and normal butane fractions.

The process unit shown in Figure 2.26 is a four-tower fractionation train. Each tower is named for the top or overhead product made by that tower. The unit produces 5 separate products:

1. Ethane 2. Propane 3. Isobutane 4. Normal butane 5. Natural Gasoline

Each tower in the fractionation train includes a bottoms reboiler to provide the heat that drives the process; an overhead condenser; reflux accumulator; and reflux / product pump. Tower pressure generally decreases and tower bottoms temperature generally increases from left to right through the train pictured, with the exception of the last tower (Deisobutanizer) which will be operated at conditions as if it were third in line.

The entire raw NGL mixture is fed to the first tower in the train, the Deethanizer. The overhead product of this tower is ethane. Because ethane is so volatile, the overhead condenser is the chiller in a refrigeration system. The chiller acts as a partial condenser since the overhead product is in the vapor phase, as shown in this example; only enough liquid is condensed to provide the necessary reflux, and no more. The liquid which dumps from the bottom of this first tower contains the mixture's propane and all heavier NGLs. It becomes the feed to the second tower in the series.

\

The Depropanizer produces a propane overhead product. The overhead condenser in this case uses cooling water to completely condense the propane vapor to liquid. A portion is pumped back in as reflux while the remainder flows as a liquid product to storage. The single refluxíproduct pump does double duty in this regard. The liquid that dumps from the bottom of the depropanizer consists of isobutane and heavier NGLs and it feeds the third tower.

The Debutanizer is operated at a temperature and pressure that will cause the total butane content of its feed to leave the tower top as a vapor. This tower also has reflux facilities but in this case, the mixed-butane overhead is not a final product, since it becomes the feed to the fourth tower. The liquid that dumps from the bottom of the debutanizer consists of isopentane and all the remaining heavier NGLs. This is the product known as natural gasoline.

The fourth and last tower is the Deisobutanizer. It produces a liquid-phase overhead product, isobutane, and a liquid-phase bottom product, normal butane. Since this tower is feed from the overhead of the previous tower, we say that it is "out of sequence" or that the sequence of fractionation is "indirect sequence." By contrast, the first three towers are in "direct sequence." Some plants will use deisobutanizers that are in direct sequence between the depropanizer and debutanizer, but there are design and operational considerations that make the indirect sequence deisobutanizer attractive.

If the NGL contains but a scant quantity of ethane, and if butanes will not be split, then only the two towers in the middle will be used, and three products will be made. The depropanizer will make a propane product that contains all of the ethane in the raw NGL feed. The debutanizer will

4



2.53


o

t

I .


PRODUCT TREATING

Water coz H2S

4

Ethane Propane Butane Gasoline X X X X X X X X

make mixed-butane overhead and natural gasoline bottom products. The decision to split butanes is essentially driven by markets and economics.

cos cs;! DMS RSH Other

PRODUCT TREATING

X X X X X X

X X X X X .

Gas contaminants such as COZ and H2S can concentrate in NGL under certain circumstances. This is especially the case with very high ethane extraction processes at 75% and higher recoveries. Trace compounds such as carbonyl sulfide (COS) and mercaptans can also concentrate in the NGL products.

How Sweet Gas Can Produce Sour NGL Assume contactor off-gas: 1000 psia [6.9 MPa], with H2S content of 1/4 gain /lo0 scf [4

ppmv or 0.21 mmHg].

Contactor NGL Unit (80% Cz Recovery) Off-Gas Residue Gas NGL

Cl 93 92.95 0.05 co2 1 0.67 0.33 c2 3.1 0.62 2.48 H2S 4 PPm" see below c3 2.0 0.02 1.98 C4'S 0.5 0.5 C5'S 0.4 O .4

100.0 mols 5.74 mols

Initially assume all H2S goes to NGL, therefore:

loo PPmv ( 5.74 mols ) = 70 ppmv (molar basis) H2S concentrations in NGL

and if NGL is fractionated the H2S will split roughly in the ratio

c2 95 c3 5 (Ref: Petty & Naeger, GPSA Convention 1987)

But after the contactor some H2S (and trace sulfur compounds) may be removed in the dehydration step to reduce the H2S level in the NGL a little below the 70 ppmv predicted above. Table 2.11 shows how water, COZ and sulfur compounds are distributed as contaminants in the NGL.

TABLE 2.11 NGL Contaminants

OJohn M. Campbell & Company


2.55


Hydrogen Sulfide

H2S 95

5

Table 2.12 illustrates how sulfur compound contaminants are fractionated as raw NGL passes

Carbonyl Carbon Methyl Ethyl Propyl Sulfide Disulfide Mercaptan Mercaptan Mercaptan cos cs2 CH3SH C2HsSH CzH7SH 38 2 10 3 1 62 5

5 85 54 2 93 43 97

through a fractionation train.

TABLE 2.12 Percent Sulfur Compound Distribution in Fractionated Products

Ethane Propane Butane Gasoline

This is shown graphically in Figure 2.27.

A number of methods are available for treating NGL products. Some of the more common processes are listed in Table 2.13. Perhaps the most common treating process for the removal of sulfur compounds from raw NGL is the familiar amine absorption process, specially modified for liquidliquid contact. (Figure 2.28)

TABLE 2.13 NGL Treating Methods

Glycol Solid Bed Desiccant Mole Sieve MEA DEA ADIP DGA (Malaprop) Caustic

Non-regenerative Regenerative

KOH Iron Sponge

~

H20 X X X

co2

X X X X

X

X X X X

X X X X

Mercaptan

X

X X

cos

X

X X

X

The flows of amine and hydrocarbon in the liquidliquid contactor are the similar to that in the contactor used for gas sweetening. The major difference being that the hydrocarbon in this case is in the liquid phase. The lighter NGL liquid floats upward through the heavier amine solution, which sinks to the bottom.

The liquidliquid contactor may be of either a packed design or a specially-modified tray design. In either case care is taken in the basic design to keep the velocities of both streams low, to prevent excessive agitation and foaming. Operationally, care must be taken to prevent pressure drops or heating that could cause the hydrocarbon liquid to vaporize.

Some of the amine solution will continuously be carried over in the sweet NGL as it leaves the top of the contactor. Separators called settlers perform the job of removing the carried over amine through the use of specialized internal devices. Two or three settlers in series may be needed to perform an adequate overall separation.



PRODUCT TREATING

o w n a o n a z - N W U

o

t



2.57

OVERVIEW OF GAS PROCESSNG INDUSTRY

a

I I I I I I I I 1

I I I I I I I

I I

3

3



PRODUCT TREATING

The sweet NGL must then be dehydrated, usually in a mole sieve unit, since it leaves the settler(s) saturated with water from the amine solution. After dehydration, the NGL is ready for transport or further processing.

The rich amine that dumps from the liquidliquid contactor is regenerated in the typical way. One amine of choice for NGL treating is Diisopropanolamine or DIPA. A 30-40% solution of DIPA is used in the Shell patented process called ADIP. Another amine with good properties in this service is Diethanolamine or DEA. Raw NGL liquid entering a DEA sweetening process is commonly contacted by the amine solution in a pressurized, agitated mixer vessel rather than in a liquidliquid contactor.

REFERENCES

2.1

2.2

2.3

2.4

2.5

2.6

2.8

2.9

2.10

2.1 1

2.12

2.13

2.14

2.15

2.16

2.17

2.18

2.19

2.20

4 2.21 2.22

Campbell, J. M., "Gas Conditioning and Processing - Vol. I ," The Basic Principles, 7'h Edition (1994), p. 5-8.

Tannehill, C. C. and C. Calvin, "Business Characteristics of the Natural Gas Industry, "GRI Report No. GRI-9310342, Ptirvin & Gertz Inc, (May 1993).

"Private communications with Jim West", CCI Cryogenics, (Jan. 1996).

"Statewide Rule 36 Hydrogen Sulfide Safe@," Railroad Commission of Texas - Oil & Gas Division, (June 1993), p. 5.

Harryman, J. M. and Bill Smith, "Sulfur Compound Distribution in NGL: Plant Test Data GPA Section A Committee, Plant Design,'' 73" Annual GPA Convention Proceedings, New Orleans, (Mar. 1994), p. 118-122.

Cullick, A. S., J. L. Magouirk and H. J. NG, "An Analysis of Soild-Forming Characteristics From Produced Gas Stream," 73rd Annual GPA Convention Proceedings, New Orleans, (Mar. 1994),

Lewis, L., "Measurement of Mercury in Natural Gas Streams," 74th Annual GPA convention Proceedings, San Anto- nio, (Mar, 1995) p. 104-108.

Cameron, C. J., Y. Barthel and P. Sarrazin, "Mercury Removal From Wet Natural Gas," 73" Annual GPA Convention Proceedings, New Orleans, (Mar. 1994), p. 256-261.

"Gas Plant Operating Cost," Gas Processor Report, (Feb. 1991), p. 3-5

Campbell, J. M., "Gas Conditioning and Processing - Vol. 1," The Bmic Principles, 7th Edition, p, 34-39. (There have been modifications to the text to provide additional details to Gas Processing Agreements}.

Tannehill, C. C. and C. Calvin, "Business Characteristics of the Natural Gas Industry," Ptirvin & Gertz, Inc., Gri Report No. GRI-93íO342, (May 1993), p. 19. Pehnec, S. T., "Production and Processing Operations - CASCO Abtr Dhabi UAE," John M. Campbell & Co., (May 1995), Section 2, p. 2.12.

Gill, D., "Liquid Profits," Oil & Gas Investor, (May 1992), p. 34.

Cannon, R. E., "The Gas Processing Industry: Origins and Evolution," Gas Processors Association, Is' Edition, (1993),

"Spot Gas Prices," Oil and Gas Jozrrnal Statistics, Based on Natural Gas Clearinghouse Information, (Mar. 1985 - Feb. 1996).

Gas Processor Report, Houston, TX

Product Pipelines of the United States and Canada, Pennwell Maps, (1988 & 1995)

Tannehill, C., L. Echterhoff and K. Trimble, "Assessing the Value of NGL in Natural Gas," GPA 74'h Annual Con- vention Proceedings, San Antonio, (Mar. 1995) p. 206-221.

Lindsey, L., "The Impact of Waterborne LPG Imports on U.S. Markets," 73" Annual GPA Convention Proceedings, New Orleans, (Mar. 1994), p. 171-173.

"A General Guide to Tank Cars," Union Pacific Railroad, (Sept. 1990).

Langford, T. R., "Capturing Value From Gas Processing,'' 70* Annual GPA Convention Proceedings, p. 248-252.

Lund, L. D., "Causes and Remedies for Mercury Exposure to Aluminum Cold Boxes," 75'h Annual GPA Convention Pre-Print, Denver, Colorado (Mar. 1996).

p. 21-28.

p. 3-19.

@Job M. CampbeU & Company

Technical Assistance Service for the Design, Operation, and Maintenance of Gas Plants .' I '


NOTES: 3

3

3 @John M. Campbell & Company


i i

1 I a i

Section 3

PHASE BEHAVIOR TABLE OF CONTENTS

PAGE # SINGLE COMPONENT SYSTEMS ............ ................................................

......................................... 3.4 .............................................................................. 3.7

........................................... 3.9 ........................... 3.9

Effect of C7+ Characterization ............... ...........................................

PREDICTION OF PHASE ENVELOPE .........................

Critical Pressure and Temperature .....................

............................................................. 3.10 ......................................... 3.12 Cricondentherm and Cricondenbar T and P .................................

..................................................

LIST OF FIGURES FIGURE # PAGE #

3.1 3.2

3.4 3.5 3.6 3.7 3.8 Refrigeration Processes ....................................... ..................................................................... 3.9

P-V-T Diagram for a Single Component System .......................................................................... 3.2 P-T Diagram for a Single Component System ..

3.3 Vapor Pressure Chart for Paraffin Hydrocarbons ............... ............................................. 3.5 Vapor Pressure Chart for Normal Parafin Hydrocarbons Based on Normal Boiling Point ...... 3.6

................................................................ 3.3

Typical Phase Diagram for a Multi-Component Mixture .............................................. Effect of Cg+ Characterization on Phase Envelope for Non-Associated Gas .............. Effect of Pressure on Dewpoint Control Processing Temperatures ..............

4



PHASE BEHAVIOR

NOTES:

3 @John M. Campbell & Compauy

3.ii BP Exploration Company (Columbia) Ltd.

Section 3

PHASE BEHAVIOR n this manual we are concerned primarily about the application of energy to achieve desired goals. I The energy possessed by any substance depends on its phase. There are three different familiar

phases ... solid, liquid and gas (vapor). You have learned to identify them by sight.

A solid possesses a definite shape and is hard to the touch. It is composed of molecules with very low energy that stay in one place even though they vibrate. There is space between these molecules so a solid is not impenetrable.

A liquid has a definite volume but no definite shape. It will assume the shape of the container in which it is placed but will not necessarily fill that container. The molecules of which the liquid is composed possess more energy than in a solid; enough energy to move from place to place. By virtue of this energy there is more space between molecules. So, a substance is less dense in the liquid form than in the solid form. \

A vapor has no definite volume or shape and will fill a container in which it is placed. The molecules have more energy than in the liquid form. In fact, they are very active. Vapor density usually is low enough that one can see through a vapor. As might be expected, the density of a given substance is lower in the vapor phase than in the liquid phase.

Our primary concern is the difference in energy level between phases, If we wish to melt a solid to form liquid we must add energy. If enough additional energy is added, the liquid can be vaporized.

We must know the phase or phases that exist at given conditions of pressure, volume and temperature in order to ascertain the corresponding energy level. To do this, we separate substances into two classifications - pure substances (single component systems) and mixtures of substances (multi-component systems).

SINGLE COMPONENT SYSTEMS

The word component refers to the number of molecular or atomic species present in the substance. A single component system is composed entirely of one kind of atoms or molecules. We often use the word "pure" to describe a single component system.

4 Figure 3.1 is a typical phase diagram for a pure substance. It has three axes - P, V and T. It is composed of a series of plane surfaces, each of which represents a given phase or a mixture of phases. We are particularly interested in the two-phase planes:



3.1

PHASE BEHAVIOR

L I L a

Pi

3

Specific Volume

Figure 3.1 P-V-T Diagram for a Single Component System

BDHG - solid plus liquid, FGIJ - solid plus vapor, and an irregular-shaped plane HCI for the liquid plus vapor. All of these planes are perpendicular to the temperature axis.

The liquid-only plane is the "cliff' to the left of plane HCI and adjacent to plane BDHG. The vapor-only plane is the "slope" to the right of plane HCI.

Although all planes are of interest, we are primarily concerned with plane HCI, the vapor-liquid region of the phase diagram.

A three dimensional phase diagram like Figure 3.1 is awkward to use. So, we normally draw a projection of this diagram. Both P-T and P-V projections are important.

P-T Plot for a Pure Substance Since all of the two-phase planes in Figure 3.1 are perpendicular to the T axis, they appear as

single lines in a P-T projection like Figure 3.2.

Lines HD, HC, and FH are the equilibrium lines - combinations of pressure and temperature at which the adjoining phases are in equilibrium. At equilibrium one can change phase, at constant pressure and temperature, by simply adding or removing energy from the system. Point H, the triple point, is the only combination of pressure and temperature at which all three phases can exist together.

Along line FH no liquid phase is ever present and solid sublimes to vapor. The use of "dry ice" for cooling is an example of this. Line HD is the equilibrium line between solid and liquid. Ice


3.2 ________ ~~


SINGLE COMPONENT SYSTEMS

Vapor

water at 0°C [32"F] and atmospheric pressure occurs on this line. Line HD can have a positive or negative slope depending on whether the liquid expands or contracts on freezing. The energy change occurring along line HD is called the heat offusion. At any P and T along this line the system can be

This line could be called the solid-liquid saturation or solid-liquid equilibrium line.

\ all solid, all liquid or a mixture of the two depending on the energy level.

Line HC is the saturation or equilibrium curve between vapor and liquid. It starts at the triple point and terminates at the critical point "C." The pressure and temperature conditions at this latter point are known as critical temperature (T,) and critical pressure (P,). At this point the properties of the liquid and vapor phases become identical. For a pure substance the critical point can be defined as that point above which liquid cannot exist as a unique separate phase. Above P, and T, the system is often times referred to as a dense fluid to distinguish it from normal vapor and liquid.

Line HC is often referred to as the vapor pressure curve. Such vapor pressure curves are available from many sources. Line HC is also the bubblepoint and dewpoint curve for the pure substance.

In Figure 3.2, consider a process starting at pressure PI, and proceeding at constant pressure. From "m" to "n" the system is entirely solid. The system is all liquid for the segment o-b. At "b" the system is a saturated liquid - any further addition of energy will cause vaporization at constant pressure and temperature. At "d," the system is in the saturated vapor state. At temperatures above "d," it is a superheated vapor.

Line HC is thus known by many names - equilibrium, saturated, bubblepoint, dewpoint and vapor pressure. For a pure substance these words all mean the same thing.

\ At the pressure and temperature represented by HC the system may be all saturated liquid, all saturated vapor or a mixture of vapor and liquid. The exact phase condition of the system depends on the energy level at the P and T involved.



3.3

PHASE BEHAVIOR

The rectangle "bfghd" illustrates another important phase property that is confirmed experimentally. Suppose we place a liquid in a windowed cell at condition "b" and light it so it is easily visible. We then increase pressure at constant temperature (isothermally). As we proceed toward point 9'' the color will begin to fade. At some point (as we blink our eyes) the color disappears completely. The cell now contains what looks like a vapor, but no bubble of vapor was ever seen to form.

At 'If' (above the critical) the system is in a fourth phase that cannot be described by the senses. It is usually called dense phase fluid, or simply fluid. The word "fluid" refers to anything that will flow and applies equally well to gas and liquid.

This fluid at "f' looks like a gas but possesses different properties from regular gas found to the right of line HC and below the critical pressure. It is denser than regular gas but is more compress- ible than a regular liquid. Gas type correlations are used but must be modified to reflect the different behavior patterns of this dense phase fluid.

From "f' one can proceed at constant pressure (isobaric) to "g," reduce pressure isothermally to "h," and then proceed isobarically to "d." One has gone from saturated liquid to saturated vapor without encountering any discernible change of phase.

One could go from "b" and "d" directly by just adding energy to the liquid at constant pressure. In the cell you would observe bubbles of vapor forming and an interface would develop between gas and liquid. As energy input continued the liquid level would fail until the liquid phase disappeared. No temperature change would occur in going from "b" (saturated liquid) to "d" (saturated vapor).

Refer again to Figure 3.1. On the temperature axis you will note t,. If you follow this line it will go through the critical point "c" and is tangent to the phase envelope HCI. The temperature lines between line HC and CI inside the phase envelope occur at constant pressure. This is a basic characteristic of all such diagrams for pure substances.

Figure 3.3 is a vapor pressure plot for light paraffin hydrocarbons. The lines shown are curve HC in Figure 3.2. They have been straightened artificially by using an odd scale on the abscissa. Figure 3.4 is a corresponding vapor pressure chart for paraffins based on their normal boiling point.(3.''

Although the true vapor pressure curve of a pure component must terminate at its critical point, the line often is extrapolated above that point in the calculation of mixture properties. This pseudo value may be used to estimate the contribution of that molecular species to total mixture properties.

MULTI-COMPONENT SYSTEMS For a multi-component system, another variable must be added to the phase diagram - com-

position. The location of the lines on a phase diagram depends on composition.

For a pure substance phase envelope HCI is a plane surface parallel to the temperature axis. For a multi-component mixture of substances this phase envelope is not a plane; it has thickness, somewhat like your tongue. Composition is the variable that reflects this thickness. If you replace specific volume in Figure 3.1 by composition and then make a pressure-temperature projection of the solid diagram, you obtain a figure like 3.5. The location of the lines in Figure 3.5 thus varies with composition. This figure is a projection showing only the liquid-vapor portion of the total phase diagram.

@John M. Cunpbeii & Company


MULTI-COMPONENT SYSTEMS

Temperature, OF

Temperature, OC

Figure 3.3 Vapor Pressure Chart for Paraffin Hydrocarbons

There are several terms used to define the location of various points on the phase envelope.

Cricondenbar - maximum pressure at which liquid and vapor may exist (Point N).

Cricondentherm - maximum temperature at which liquid and vapor may coexist in equilibrium (Point M).

Retrograde Region - that area inside phase envelope where condensation of liquid occurs by lowering pressure or increasing temperature (opposite of normal behavior).

Quality Lines - those lines showing constant percentages which intersect at the critical point (C) and are essentially parallel to the bubblepoint and dewpoint curves. The bubblepoint curve represents 0% vapor and the dewpoint curve 100% vapor.

Line ABDE represents a typical isothermal retrograde condensation process occurring in a condensate reservoir. Point A represents the single phase fluid outside the phase envelope. As pressure is lowered, Point B is reached where condensation begins. As pressure is lowered further, more liquid

inflection points of such lines. As the process continues outside the retrograde area, less and less liquid forms until the dewpoint is reached (Point E). Below E no liquid forms.

4 forms because of the change in the slope of the quality lines. The retrograde area is governed by the

@John M. Camphii & Company


PHASE BEHAVIOR

Temperature, K

Figure 3.4 Vapor Pressure Chart for Normal Paraffin Hydrocarbons Based on Normal Boiling Point

Retrograde

2

9? a

3 v> 0-J

O 20 40

Temperature

Figure 3.5 Typical Phase Diagram for a Multi-Component Mixture

@John M. Cpmpbeii & Company


1 1’ n ’


In my experience the critical point has always occurred to the left of the cricondenbar for naturally occurring hydrocarbon gas mixtures. It is not, however, necessarily in the position shown. It may be further down on the phase curve or closer to the cricondenbar. Location of 'IC" is most important, for it fixes the shape of the quality lines which in turn governs the vapor-liquid ratio at a given temperature and pressure within the phase envelope.

Effect of C 7 t Characterization The analysis and/or characterization of the cg+ or C7+ fraction in a natural gas mixture is not

routine, yet has a significant effect on the shape of the phase envelope. This is illustrated for one gas in Figure 3.6.

Figure 3.6 Effect of C6+ Characterization on Phase Envelope for Non-Associated Gas

Figure 3.6 presents phase envelopes for 4 different characterizations of the cg+ fraction - C7, Cg, C11 and full characterization based on a distillation analysis.

Special gas chromatographic techniques can identify individual components through about cg- C ~ O . This is called an extended analysis. If the phase behavior of the gas has a significant effect on the system design it is highly recommended that an extended analysis be performed. If an extended analysis is not available, predictive characterization techniques may be used.

For lean natural gases the C7+ characterization has a dramatic effect on the dewpoint line. The effect on the location of quality lines is much less significant.

@Job M. CPmpbeU & Company


3.7

PHASE BEHAVIOR

Refrigeration Processes Refrigeration is frequently used to cool a gas in order to meet a hydrocarbon dewpoint specifi-

cation. When the refrigeration takes place at high pressure, it is very important to have an accurate description of the phase behavior. Figure 3.7 shows the phase envelope for a lean gas typical of that found in the southern North Sea or Gulf of Mexico. The cricondentherm of this gas is 30°C [86"F] and occurs at a pressure of 3000 kPa [435 psia].

10 O00

9000

8000

7000 m Y - 6000 p! 3 "> v> p! a 4000

3000

2000

1 O00

" -1 50 1 O0 -50

Temperature, "C O 50

Figure 3.7 Effect of Pressure on Dewpoint Control Processing Temperatures

Assume that this gas is to be processed in a mechanical refrigeration plant to meet a hydrocarbon dewpoint specification of -3°C [27"F]. If the processing pressure is 3000 kPa [435 psia], the gas must be cooled to at least -3°C [27"F] to meet the dewpoint specification. However, if the processing pressure is 7500 kPa [lo88 psia], the gas must now be cooled to -25°C [-13"FI. The following shows the processing temperature required to meet the dewpoint specification at various process temperatures.

P, kPa

7000 -19.4 -25.1 -34.0

Above about 8000 kPa [1160 psia] this gas could not be processed in a mechanical refrigeration plant to meet the hydrocarbon dewpoint specification. In order to meet the specification an expansion process (valve or expander) or an adsorption process must be used.

@Job0 M. Campbeii & Company

3.8 BP Exploratlon Company (Columbla) Ltd.


O 20 80 100

Temperature

Figure 3.8 Refrigeration Processes

Operation Near the Critical Region In some cases, having an accurate phase envelope is not good enough; you also need a reliable,

true critical point. This may be illustrated by the quality lines discussed previously. If, in Figure 3.8, the critical is at C1 the 20% and 80% quality lines might be as shown. Suppose, instead, that the true critical is at C2. Notice the possible change in the location of the same quality lines. If the operating pressure is much less than C2, the error in the amount of liquid predicted will be relatively small.

The pseudocritical found fiom PVT models are seldom reliable.

\

However, if the operating pressure falls between C2 and Cl, the error can be significant - having a profound effect on system design andor operation. In general, operation near the critical point should be avoided. Common sense dictates that if the system is difficult to model it will be difficult to operate and control. If near-critical operation is unavoidable, the process design should reflect the uncertainties involved. the most common examples of potential near-critical operation are low temperature expander plants and de-ethanizers.

Practical Suggestions The above are merely examples of typical problems that have arisen. One could cite many

others. Failure to handle phase behavior in a workmanlike, professional manner has proven, and will continue to prove, very expensive for the petroleum industry. The cost of obtaining good data is usually trivial compared to the economic benefits obtained.

There are some good guidelines that should be followed:

1. Obtain good samples using experienced samplers. 2. Handle the samples carefully.

3. Analyze the samples in a proven laboratory. 4. Develop all, or that portion needed, of the phase curve. Above about 14 MPa [2000 psi]

calculated points are suspect no matter how big the computer or how complex the program. The results should be regarded as estimates only. If an estimate is good enough, fine. If



PHASE BEHAVIOR

not, go to a reliable laboratory. In many cases a few laboratory points supplemented by calculations will suffice.

PREDICTION OF PHASE ENVELOPE The location of the bubblepoint and dewpoint lines may be calculated using vapor-liquid equi-

librium (VLE) methods. For most naturally occurring systems above about 14 MPa [2000 psia], the validity of the standard calculation becomes questionable. If the location of the curve at high pressure is very important - and if a reliable fluid sample is obtainable - a laboratory determination of at least selected parts of this curve is recommended.

However, there are circumstances where the calculation of all or part of the upper phase curve may be satisfactory. References 3.2 on, at the end of this chapter, provide further details.

Estimation of the critical, cricondentherm and cricondenbar points is particularly important.

The application of K-values to calculate phase quantities and compositions proceeds as follows.

This K, may be incorporated into a material balance around a separator, where:

zi = mol fraction of any component in total feed stream to separation vessel

yi = mol fraction of any component in the vapor phase xi = mol fraction of any component in the liquid phase Ki = equilibrium vaporization ratio (equilibrium constant)

v9 Y¡ - = yilxi

F = total mols of feed V = total mols of vapor L = total mols of liquid

Component balance: Fzi = Vyi + Lxi

If we set F = 1.0 so that L and V are now liquid and vapor-to-feed ratios and if we remember that yi = Kixi

rearranging, Zi x. = ’ L + V K j

Since the summation of liquid fractions must equal one, we can write the following equation.

The equation serves as the objective function in an interactive calculation to determine the

1. Determine K values of each component at the temperature and pressure of the system. 2. Estimate a value of L (remember, V = 1 - L)

3. Solve the above equation. If Xi f 1.0 assume a new value of L and repeat step 2. 4. When xi = 1 .O, the phase quantities L and V are known as well as the liquid phase compo-

quantity of L or V. The calculation procedure is as follows:

sition. Vapor phase compositions may be calculated by remembering that yi = Kixi, @John M. CampbeU & Company

9


PREDICTION OF PHASE ENVELOPE

Component

c2 c3

iC4 nc4

\ The foregoing calculations is known as a flash calculation and is used to predict the equilibrium quantities and compositions of two phase systems. This calculation may also be made using the following objective functions.

mol%

10

50

15

25

= 1.0 Zi

= V+(L/Ki)

zi (Ki - 1) = o C y i - C x i = V(Ki-1)+1

or

The second of the above equations is widely used in computer programs.

Special cases of a flash calculation include bubble point (V = O, L = 1) and dew point (V = 1, L = O) calculations. In these cases the iteration is made around T or P and the objective functions are as follows:

Bubblepoint

Dewpoint

Examples of bubblepoint, dewpoint and flash calculations for a simple hydrocarbon mixture r, follow



3.1 1

PHASE BEHAVIOR

~ ~ ~~

Example 3.1 (Cont'd.):

Dewpoint -

Component

c2 c3 iC4

Total

I Assume = 65OC 1 Assume T = 93°C I Assume T = 77OC

Yi Ki Y i/Ki Ki Y i/Ki Ki Y W i 0.10 3.24 0.031 4.26 0.023 3.64 0.028 0.50 1.23 0.407 1.83 0.273 1.46 0.343 0.15 0.610 0.246 0.993 0.151 0.750 0.200 0.25 0.535 0.535 0.584 0.428 0.583 0.429

1 .o00 1.219 0.875 1 .o00

Calculate the amount of vapor and liquid and the composition of the two phases if this feed entered a vessel @ 1724 kPa [250 psia] and 65°C [ 150"FI.

Calculate the bubble point and dewpoint temperature at 1724 kPa (250 psia) of the following hydrocarbon mixture.

Flash Calculation -

Assume Assume Assume L = 0.5 L = 0.75 L = 0.649

Component zi Ki xi xi xi yi = K;x~

c2 0.10 3.24 0.047 0.064 0.056 0.181

c3 0.50 1.23 0.448 0.473 0.463 0.569 iC4 0.15 0.610 O. 186 0.166 0.174 O. 106 nC4 0.25 0.467 0.341 0.288 0.308 0.144

Total 1 .o0 1 .o22 0.99 1 1 .o00 1 .o00

For manual calculations, K-values can be found in Chapter 25, "Equilibrium Ratio (K) Data" in the GPSA Engineering Data Book, or Appendix 5A Volume 1 "Gas Conditioning and Processing - The Basic Principles," Campbell Petroleum Series.

Cricondentherm and Cricondenbar T and P Grieves and Thodos have prepared correlations for these points based on 123 binary and 15

multi-component m i ~ t u r e s . ( ~ . ~ > ~ . ~ ' For the systems tested, the agreement between predicted and measured values was very satisfactory. The prediction of temperature is better than pressure. The maximum is less than 5%, whereas the maximum pressure deviation was around 13%.

These correlations are useful in conjunction with vapor-liquid equilibrium (VLE) calculations and critical point predictions. Usually in all of these it may be possible to construct a phase curve (or portion thereof) that is useful for at least planning purposes. This may be the case in the early stages of development when good fluid samples are not yet available. Sometimes the oniy samples are fiom drill-stem tests or nonproduced exploratory wells, which taints the validity of the laboratory results. Some calculated cricondentherm and cricondenbar numbers may prove helpful, for they denote the maximum T and P of any phase curve.


1

3


PREDICTION OF PHASE ENVELOPE

\ Critical Pressure and Temperature The prediction of the location of the critical point is difficult. The most reliable way remains a

laboratory study of a reliable sample. But, there are many circumstances where a calculated number may be suitable.

Based on an empirical fit of 25 natural gas systems possessing a molecular weight less than 30, the following equation has been developed.

T,/T,‘ = 1 .O + (0.03)(M gas - 16)

Where: T, = actual critical temperature T,’ = pseudocritical temperature from Kay’s Rule

For the samples tested the error was less than plus or minus 5%.

A more accurate m e t h ~ d ( ~ . ~ ) uses the equation

Where: Tcmix = critical temperature of mixture

TCi = critical temperature of each component

cpi = critical volume fraction as defined in Equation 3.7

Yi Vci

C Yi vci cpi =

Where: yi = mol fraction vci = critical volume of each component

A similar has been proposed which employs third parameters mixing rules, but the increased accuracy is marginal.

No simple, satisfactory correlation has been found for critical pressure. outline some methods available.

(3.5)

(3.7)

and more complex

References 3.6-3.8

One problem is prediction of the critical values for the hydrocarbon fractions heavier than hexane shown in analyses. Correlations based on the mean boiling point of the fraction, as found from distillation test, have proven ~se fu l . (~ .~ , 3.9)

Simon and Yarb~rough(~.”) present a correlation for predicting critical pressure based on 14 gas-solvent reservoir systems using composition and the heptanes plus fraction molecular weight and characterization factor as parameters. A total of 37 other systems shown in the literature were tested. The average deviation was about 5%.

Whit~on(~.’’) has proposed a series of generalized equations for property prediction, including critical values for heavier fractions such as heptanes plus. The following equations were proposed that used earlier

The value of Tb in Equations 3.8 and 3.9 may be found from a distillation analysis or from a mixing rule.

pc = (“I)” @John M. Campbell & Company


3.13

PHASE BEHAVIOR

SI Units K

Where: T b = average boiling point y = relative density

P, = critical pressure a = empirical constant b = empirical constant c = empirical constant

FPS Units "R

SI Units FPS Units

W a psia 5.53 x io9 3.12 x io9

-2.3 1 2.32

Equation 3.8 applies for any liquid having a boiling point less than 472 K [850°R].

Where: T, = critical temperature a = empirical constant b = empirical constant c = empirical constant y = relative density

19.1 I 24.3 0.59 0.36

It is recommended that references like those cited immediately above be used with caution and only by those very familiar with phase behavior. They are shown to illustrate the type of simple, manual correlations available.

Equations of state provide the most accurate and consistent methods of predicting critical properties. However, even these methods provide an approximation only.

3.1 3.2

3.3 3.4 3.5 3.6 3.7 3.8 3.9 3.10 3.11 3.12

REFERENCES Lee, B. 1. and Kesler, M. G., Hydr. Proc. (July 1980), p. 163. Maddox, R. N. and Erbar, J. H., Gas Conditioning and Processing, Vol. 3 (1982), Campbell Petroleum Series, Nor- man, Oklahoma, USA. Grieves, R. B. and Thodos, G., SPE Jour. (Dec. 1963), p. 287. Ibid. (Sept. 1964), p. 240. Li, C. C., Con. J. Chem. Eng., Vol. 19 (1971), p. 709. Church, P. L. and Prausnitz, J. M., AIChEJ., Vol. 13 (1 967), p. 1099. API Technical Data Book, Ch. 4, "Critical Properties." Spencer, et aL, AIChEJ., Vol. 19 (1973), p. 522. Maddox, R. N. and Erbar, J. H., Hydr. Proc. (Jan. 1984), p. 119. Simon, R. and Yarborough, L., Jour. Petr. Tech. (May 1963), p. 556. Whitson, C. H., SPE Jour. (Aug. 1983), p. 683. Riazzi, M. R. and Daubert, T. E., Hydr. Proc. (Mar. 1980), p. 115.

3


3.1 4 BP Exploratlon Company (Columbia) LW.

Section 4

WATER-HYDROCARBON BEHAVIOR TABLE OF CONTENTS

PAGE #

WATER CONTENT OF GAS ....................................................................................................................... 4.1 P-T Correlations ................................... ............. 4.3 P-T, Composition Correlations ... ......... ................................................................. 4.8 PVT Correlation ..................................................... 4.11

...................................................... 4.13 Summary of Methods ................................... ............................... 4.14

........................................................................ 4.14 Katz et al., Approaches ................................................................................................ 4.17 Campbell et al., Approaches ............. ......................................................... 4.24

..................................................... 4.32 HYDRATE INHIBITION .............................................................................................................................. 4.32

HYDRATES . . . . . . .. ..

Inhibitor Losses to the Hydrocarbon Phase ......... ..... .......................... 4.35 Calculation Summary ............................................................. 4.37 Crystallization (Freez ..................................... ... Glycol Losses .......................................................................................................................................... 4.39

LOW TEMPERATURE IN GLYCOL INJECTION GAS PROCESSING FACILITIES ........................... 4.40 Basic System Operation ......................................................................................................................... 4.40 Regeneration .. ... . . . . . . . . . . . . . . . . . , . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . __. . . . . . . . . . .. . . . . . . . . . . .. . . . . ............................ 4.42

.................................................................. 4.43 Filtration.. . . .. .. .. . . . . . . .. . . . . . . . . . ... .. . . .. . . .... .... . . .. . . .... ..... ................................................................... 4.44 Glycol Losses ..................................................................................................

................................................ 4.44 ........_......... .......... .. ............ ...................... ... 4.46

Glycol-Water-Oil Separation ................ ............

REFERENCES

LIST OF FIGURES FIGURE # PAGE # 4.1 4.2(a) 4.2(b) 4.3(a) 4.3(b) 4.4(a) 4.4(b) 4.5(a) 4 4 b ) 4.6(a)

Solubility of Water in Liquid Hydrocarbons ................................................................................. 4.2 Dewpoint of Natural Gas (SI Units) .............................................................................................. 4.4 Dewpoint of Natural Gas (FPS Units) ........................................................................................... 4.5 Water Content of Lean, Sweet Natural Gas (SI Units) ................................................................. 4.6 Water Content of Lean, Sweet Natural Gas (FPS Units) .............................................................. 4.7 Effective Water Content of Saturated C02 in Natural Gas Mixtures (SI Units) ......................... 4.9 Effective Water Content of Saturated C02 in Natural Gas Mixtures (FPS Units) ..................... 4.9 Effective Water Content of Saturated H2S in Natural Gas Mixtures (SI Units) ......................... 4.10 Effective Water Content of Saturated H2S in Natural Gas Mixtures (FPS Units) ...................... 4.10 Water Content Ratios - Sour Gas (SI Units) ............................................................................... 4.12

@John M. Campbeil& Company


4.1


LIST OF FIGURES (CONT'D.) FIGURE # PAGE # 4.6(b) Water Content Ratios - Sour Gas (FPS Units) ............................................................................ 4.12 4.7 Water Content of 5.31% C3 - 94.69% Ci Gas in Equilibrium with Hydrate ...__ 4.8 Schematic of Natural Gas Hydrate Lattice ........ .......... ............ .......... ............ ........ 4.9 Hydrate Forming Conditions for Natural Gas Components .......................................................... 4.16 4.10 Conditions of Hydrate Formation for a Hydrocarbon Mixture ..................................................... 4.16 4.1 1 Pressure-Temperature Curves for Predicting Hydrate Formation (SI & FPS Units) ................... 4.18 4.12 Permissible Expansion of a 0.6 Gravity Natural Gas Without Hydrate

Formation (SI & FPS Units) ........................................................................................................... 4.19 4.13 Vapor-Solid Equilibrium Curves (Methane, Hydrogen Sulfide, and Carbon Dioxide) ............... 4.20 4.14 Vapor-Solid Equilibrium Curves ( ....................................................... 4.21 4.15 Vapor-Liquid Equilibrium Curves ....................................................... 4.22 4.16 Vapor-Solid Equilibrium Curves (iso-butane and n-butane) .......................................................... 4.23 4.17 Hydrate Prediction Chart for 69 bar [ 1 O00 psia] .... ....__.... ................................ 4.1 8 Hydrate Prediction Chart for 138 bar [2000 psia] ......................................................................... 4.26 4.19 4.20 Hydrate Prediction Chart for 276 bar [4000 psia] ........ 4.21 Hydrate Prediction Chart for 345 bar [5000 psia] ....__. .............................................. 4.29 4.22 Hydrate Prediction Chart for 414 bar [6000 psia] ....... .............................................. 4.30 4.23 Hydrate Depression Correction for Non-Hydrating M O00 psia ............................... 4.31 4.24 Hydrate Depression Correction for Non-Hydrating Molecules at 2000-6000 psia .............. ....... 4.3 1 4.25 Incipient Hydrate Formation for a Typical Gas Condensate Fluid in the

Presence of Methanol Solutions ........................................................................................... 4.25 Hydrate Depression Correction for Non-Hydrating cules at 2000-6000 psia ..................... 4.33 4.26 Vapor-Liquid Equilibrium of Methanol over Water .................................. ~ .......... 4.36 4.27 Methanol Solubility in NGL Liquid ......................... ............................................................. 4.37 4.28 Freezing Points of Glycol-Water Solutions .................................................................................... 4.40 4.29 Typical MEG Injection System ............................. ..................................... 4.41 4.3 1 Glycol Injection Practices ................................................................................................................ 4.42 4.31 Temperature-Composition, T-x, Diagram for MEG-Water System at 1 atm ................................ 4.43 4.32

Hydrate Prediction Chart for 207 bar [3000 psia] ....... . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Example IFPEXOL@ Dehydration Process Flow Diagram ... ................... , ........ .. .........

1

LIST OF TABLES TABLE # PAGE # 4.1 Comparison of Methanol and Monoethylene Glycol ..................................................................... 4.32


4.H BP Exploration Company (Columbia) Ltd.

Section 4

WATER-HYDROCARBON BEHAVIOR he amount of water present at saturation in a natural gas stream is relatively low, as shown

T b e l o w . at 38.7"C [lOO°F] and 69 bar [io00 psia] - mol. fr. water = 0.0013 at 38.7"C [100"F] and 6.9 bar [lo0 psia] -mol. fr. water = 0.0090

The water content of natural gas increases with increasing temperature and decreases with increasing pressure. The contents shown above are low on an absolute basis and not particularly significant when normal errors in gas analyses are considered.

The solubility of water in sweet liquid hydrocarbons also is very low, as shown in Figure 4.1. In sour hydrocarbons the solubility may be 2-3 times higher but is still very small. For this reason, in phase behavior calculations it is often assumed that water is insoluble in liquid hydrocarbons.

\

Water, even in amounts this small, may need to be removed from the gas and/or liquid to meet sales specifications or processing andíor transportation requirements. Condensed water is also the potential source of two operating problems:

1. The formation of hydrates or ice which can plug the piping and equipment. 2. Corrosioníerosion in the system when sulfur compounds or carbon dioxide are present.

Removal or inhibition of water to a desirable level is thus an integral part of system planning. To achieve this, there are three basic calculations involved:

1. Water content of gas 2. Prediction of hydrate formation conditions 3. Inhibition of hydrates with chemical inhibitors

This section of the manual is devoted to these topics.

WATER CONTENT OF GAS

4

Up to a pressure of about 3.45 bar [50 psia] the behavior of the water-natural gas system is ideal. Thus, Raoult's Law can be used reliably.

(Yi) (P) = (Pv) (xi) (4.1)

Where: y; = mol. fr. of any component in the gas phase P = total pressure

P, = vapor pressure of component xi = mol. fr. of same component in the liquid phase



4.1

WATER-HY DROCARBON BEHAVIOR

Temperature, "C 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80

1 .o00 0.900 0.800 0.700 0.600

0.500

0.400

0.300

0.200

8 0.100 2 0.090

0.080 2 0.070 1. 0.060 I 0 0.050 E g 0.040

0.030

-0

O O

E 0 s z 0.020 .- a I3 O m

la

-

ai I 0.010 4-

0.009 0.008 0.007 0.006 0.005

0.004

0.003

0.002

0.001 ' ' ' I I 1 I 'I I ' ' ' 1 I I I ' I I I I l l I I ' 1 ' I'I I I I ' I ' I I'I I I I ' I I I I I I ' I I I ' I " I 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180

Temperature, OF

7

Figure 4.1 Solubility of Water in Liquid Hydrocarbons


4.2 ~~



Where: W = kg/106 m3 (std) A = 761 400 y = mol fraction H20

Because water is basically insoluble in hydrocarbons it will exist as a separate liquid phase (even though it may be emulsified in the hydrocarbon liquid). In this water phase, xi = 1 .O, so the mol. fr. water vapor in the gas (yi) is equal to P,/P can be found from a vapor pressure chart or a steam table at the system temperature.

It is common to express water content in kilograms per million standard cubic meters, kg/106 m3 (std) [pounds per million standard cubic feet, IbmíMMscfl. The conversion from mass of H20 per standard volume of gas is shown below:

W = A y (4.2)

SI units II FPS Units r IbmiMMscf

47 430 mol fraction H20

SI Units 101.325 Wa Psid

FPS Units 14.696 psia

When the system pressure is above 3.45 bar [50 psia] other approaches are needed to predict water content. There are three basic methods:

1. Empirical correlations based only on gas pressure and temperature. 2. Empirical correlations based on P, T and gas composition. 3. The use of PVT equations of state.

P-T Correlations Many P-T correlations have been developed - McCarthy, Boyd, and Reid; McKetta; AGA;

etc. All are based on the same basic body of data for sweet, relatively lean natural gas - usually containing over 70% methane. The data, for the most part, were taken on pipeline quality gases where most of the "heavy ends" had been removed.

Most such correlations give about the same result. Figures 4.2 (a&b)@'p4.') is a chart originally published by McKetta and Wehe published in 1958. The correction shown for water salinity is usually ignored because it is seldom known. This correction is not significant, since the water phase is essentially "fresh" - being condensed from the gas phase. The correction for gas specific gravity is not statistically significant. Thus, only the basic chart is normally employed.

Figure 4.3 (a&b) is a composite correlation based on a previously published correlation of this type.

All such correlations are used in the same manner. The left hand ordinate is gas-water content in kg/103 m3 (std) [lbm/MMscfl at saturation (i.e., the maximum amount of water the gas can hold). The abscissa is the water dewpoint temperature of the gas - the temperature at which the gas is saturated with water vapor at a given pressure. The diagonal lines represent system pressure.

4

Enter at 40°C [104"F] in Figure 4.3 (a&b), proceed vertically to 7.0 MPa [lo15 psia] and thence horizontally to the ordinate. Read a water content of 1100 kg/106 m3 (std) [68 lbm/MMscfJ.




I 500

1 O00 800

600 500

400

300

200

150

000

O00 O00

O00 O00 O00

' O00

1 O00

1 O00 20 25 30 35 40 45 50

1 O0 O00

80 O00

-50 -40 -30 -20 -10 O 10 20 30 40 50 60 70 80 100 120 140

Temperature, O C

kPa I (abs)

Figure 4.2(a) Water Content of Lean, Sweet Natural Gas



3

4

4


60000

JOOOO

?3000

1

O

o a

a

20 25 30 35 40 45 50 Molecular Weight

4000 4000

Water contents of natural gases with corrections lor salinity and relativa density After McKetta and Wehe, Hydrocarbon Processina. Auaust. 1958

2 I 2

Y I 1 I I I I I " -

'L I .?'o ' 40 I :O ' A I I 1 1 1 1 1

S'O ' i0'0' 120' l4O'i6O'I~OZOO 240 280' '

Temperature, "F

I

Figure 4.2(b) Water Content of Lean, Sweet Natural Gas

@John M. Cpmphii & Company


.E I


100000 80000

60000 50000 40000

30000

c (a 10000

8000

6000 5000 5 4000

3000 c

m

E 2000 c v)

O <D

Y \ y 1000

800

600 500 400

300

200

1 O0 80

60 50 40

30

20

10 -40 -20 O 20 40

Water Dewpoint, "C 60 80 1 O0

Figure 4.3(a) Water Content of Lean, Sweet Natural Gas

@John M. cimpbcii & Company

4.6 BP Exploration Company (CoiumMa) Ltd.

3

3


O000 8000

6000 5000 4000

3000

2000

1 O00 800

600 500 400

300

200

1 O0 80

60 50 40

30

20

10 8

6 5 4

3

2

1 -40 O 40 80 120

Water Dewpoint, O F 160 200 240

Figure 4.3(b) Water Content of Lean, Sweet Natural Gas



4.7


SI Units Where: W = water content of gas kg/106 m3 (std)

kg/106 m3 (std) kg/106 m3 (std) kg/106 m3 (std)

Whc = water content from Figure 4.2 or 4.3 W, = effective water content of COZ from Figure 4.4 W2 = effective water content of H2S from Figure 4.5

y = mol. fr. of all components except H2S and COZ yi = mol. fr. of COZ y2 = mol. fr. of H2S

~

This is the maximum water content of the gas at 40°C [104"F] and 70 bar [lo15 psia]. It is not necessarily the actual water content. The gas may not be fully saturated with water vapor.

FPS Units 1bMMscf IbíMMscf lb/MMscf lb/MMscf

Example 4.1: Water saturated natural gas at 3OoC [86"F] and 34.5 bar [500 psia] is compressed to 103.4 bar [1500 psia]. The gas leaves an aftercooler at 50°C [122"F] before entering a scrubber ahead of a dehydration unit. Should we expect to drain water from the scrubber?

Estimate water content from Figures 4.2 or 4.3 @I 50°C [12Z°F] & 103.4 bar [1500 psia] = 1.32 kg/103 m3 (std) [82 Ibm/MMscf @I 30°C [86"F] & 34.5 bar [500 psia] = 1.10 kg/103 m3 (std) [69 lbm/MMscfl

Answer: The gas entering the scrubber can hold more water than the gas entering the compressor. Therefore, no water can condense. The gas is not saturated at the scrubber.

It may be assumed that the gas is saturated with water vapor at the last set of conditions where liquid water is detected. It is also saturated at reservoir pressure and temperature.

P-T, Composition Correlations

Composition of natural gas has an effect on water content, particularly when the gas contains hydrogen sulfide and carbon dioxide. Several methods are used to correct correlations like Figures 4.2 and 4.3 for these acid gas components.

Example 4.2: A natural gas is 71% hydrocarbons, 16% H2S and 13% COZ. Determine the water content of this gas at 14 O00 kPa [2000 psia] and 38°C [100"F] using Equation 4.3 and Figures 4.3-4.5

From Figure 4.4, From Figure 4.5, From Figure 4.3, From Equation 4.3,

W1 = 1000 kg/106 m3 (std) [óO IbmíMMscfJ W2 = 1250 kg/106 m3 (Sta) [75 Ibm/MMscf)

wh, = 580 kg/106 m3 (std) [38 lbm/MMscfl W = (0.71)(580) = (0.13)(1000) + (0.16)(1250)

W = (0.71)(38) + (0.13)(60) + (0.16)(75) = 742 kg/106 m3 (std)

= 36.8 1bdMMscf

4.8 ~

BP Exploration Company (Columbla) Ltd.

WATERCONTENTOFGAS

1 O0 O00

m E 0 VI u>

10 O00 3 In Y

i c o> C c

s 3 1000

8

L

c <u m

a> > .- c = w

1 O0 20 40 60 80 1 O0 120 140 160

Temperature, "C

Figure 4.4(a) Effective Water Content of Saturated COZ in Natural Gas - SI Units

Temperature. "F

Figure 4.4(b) Effective Water Content of Saturated C02 in Natural Gas - FPS Units



4.9


I 1

1 O0 140 160

Figure 4.5(a) Effective Water Content of Saturated H2S in Natural Gas - SI Units

Figure


4.1 O BP Exploration Company (Columbia) Ud.

3

3

WATERCONTENTOFGAS

\ Figures 4.4 and 4.5 are based on the effective water contents of carbon dioxide and hydrogen sulfide in natural gas mixtures. These should not be used for gases that are predominantly COZ or H2S.

Correlations for estimating the water content of sour gases have a higher level of uncertainty than those for sweet gases. The above approach has been shown to yield "reasonable" values of water content for gases containing up to 30-40% H2S and/or COZ. For gases containing higher quantities of H2S and C02, experimental data is limited. In this region, all correlations (including EOS methods) are suspect. If accurate water contents are required in this region, it is suggested that they be measured experimentally.

have developed a correlation to estimate the water content of C&-H2S binary mixtures. This correlation was originally developed using an SRK Equation of State (EOS). The correlation was later expanded by Wichert & Wi~hert(~.~' to pressures up to 1000 bar [14 500 psia] .

Robinson et

When using the Robinson method, the carbon dioxide content is converted to equivalent hydrogen sulfide content by multiplying the carbon dioxide percentage by 0.75. This is then added to the actual H2S concentration to arrive at the "% H2S equivalent" for use in the correlation which is depicted graphically in Figure 4.6.

Example 4.3: Determine the water content of the gas in Example 4.2 using Figure 4.6.

% H2S Equivalent = 16% + (0.75)(13%) = 25.75%

Enter Figure 4.6 at T = 38°C and proceed horizontally to % H2S Equiv. = 25.75%. Proceed vertically to P = 14.0 MPa and read the water content ratio from the ordinat e.

Water Content Ratio = 1.28

Water Content of Sour Gas = (1.28)(640) = 820 kg/106 m3 (std) = (1.28)(40) = 51 I b W M s c f

Consider the 820 [51]; from W = 640 [40].

three water contents for the gas in the previous example. From Figure 4.6, Equation 4.3 and Figures 4.4 and 4.5, W = 840 [52.4]; from Figure 4.3 (sweet A sour gas should be expected to have a higher water content than a sweet gas.

The correlation used for design will depend on individual circumstances and preferences. From a practical viewpoint the uncertainty in the water content due to temperature and pressure effects is far greater than the uncertainty in the water content correlation.

PVT Correlations

The same PVT equations of state used for predicting physical and thermodynamic properties, equilibrium constants, etc., also may be used to predict water content. The four most common forms of the equations used are:

1 . BWR,

2. Soave-Redlich-Kwong, 3. Peng-Robinson, and 4. Eykman

Many modifications of these basic equations are employed. O J o h M. Campbell & Company

Technical Assistance Servlce for the Design, Operation, and Maintenance of Gas Plants

4.1 1


I

I

Figure 4.6(a) Water Content Ratios - Sour Gas (SI Units)

Figure 4.6(b) Water Content Ratios - Sour Gas (FPS Units)



3

3


An approximate solution of the Eykman approach is a~ailable(~.j). The others are strictly computer solutions. Different companies tend to adopt a basic program and modify it as needed to fit available data. All are empirical so the value of the output is affected by the quantity and quality of data used in developing the working program. None of the basic models are inherently superior to the others.

Many companies currently use simulation programs such as HY SIM@, PROVISION@, PROSIM@, and Chemcad@, etc. to predict water contents as well as the solubility of water in liquid hydrocarbons and hydrate formation ~onditiond~.~). The process simulators use selected Equations of State with empirically derived binary interaction parameters to model the H20-hydrocarbon, H2O-CO2 and H20-HzS behaviors.

Saturated Water Content in Equilibrium with Hydrates Figures 4.2 and 4.3 are based on the assumption that the condensed water phase is a liquid.

However, at temperatures below the hydrate temperature of the gas, the "condensed" phase will be a solid (hydrate). The water content of a gas in equilibrium with a hydrate will be lower than equilibrium with a metastable liquid.

Hydrate formation is a time dependent process. The rate at which hydrate crystals form depends upon several factors including gas composition, presence of crystal nucleation sites in the liquid phase, degree of agitation, etc. During this transient "hydrate formation period" the liquid water present is termed "metastable liquid." Metastable water is liquid water which, at equilibrium, will exist as a hydrate.

References 4.6-4.8 present experimental data showing equilibrium water contents of gases above hydrates. Data from Reference 4.6 is presented in Figure 4.7. For comparative purposes, the "metastable" water content of a sweet gas from Figure 4.2 is also shown. The water content of gases in the hydrate region is a strong function of composition. Figure 4.7 should not be extrapolated to other compositions.

1 O0

10

c

2

c 5 1

;

z 0 c

O

- o 1

o o1 -40 -30 -20 -10 O 10 20 30 40

Temperature, O F

~

Figure 4.7 Water Content of 5.31% C3 - 94.69% C, Gas in Equilibrium with Hydrate



4.13

WATER=HYDROCARBON BEHAVIOR

When designing dehydration systems, particularly TEG systems to meet extremely low water dewpoint specifications, it is necessary to determine the water content of the gas in equilibrium with a hydrate using a correlation like that presented in Figure 4.7. If a metastable correlation is used, one will overestimate the saturated water content of the gas at the dewpoint specification. This, in turn, may result in a dehydration design which is unable to meet the required water removal. Where experimental data is unavailable, utilization of a sound thermodynamic-based correlation can provide an estimate of water content in equilibrium with hydrates.

9

Summary of Methods The EOS-based methods for estimating water content should be inherently more reliable be-

cause they employ some consideration of the fundamental variables affecting water content. In real world applications, however, the limitation in the models often suggests that the empirical correlations are more accurate and convenient to use.

HYDRATES Gas hydrates are an ice-like solid that form above the freezing point of water, and many times

form in preference to ice below the freezing point of water. Free water need not be present for hydrates to form, but the hydrocarbon stream must be saturated with water. Favorable hydrate formation conditions are typically low temperature and high pressure. Hydrates are a crystalline solid which is stabilized by a structure which includes small gas molecules and water molecules. Normal butane forms hydrates with difficulty, while molecules larger than butane are generally thought to be too large to enter the hydrate structure and may actually inhibit hydrate formation.

A hydrate is a solid formed by the physical combination of water molecules and certain of the gas molecules. It is solid, like ice, but possesses different characteristics.

Ice is a simple crystalline solid. Hydrates, on the other hand, crystallize in one cubic structure, or in a combination of two cubic structures, in which gas molecules are "trapped" in cavities. These cavities occur in a framework composed of water molecules linked together by hydrogen bonds. The water molecules are the structural members similar to the steel lattice of a building. However, the structure is very weak and will collapse unless supported by any mass occupying the cavities. This is similar to the subsidence of the earth's surface that may occur when fluid is withdrawn from porous rock.

The basic lattice of Structure I has a cell constant of 12A and contains forty-six water molecules arranged to form eight voids, two of which are pentagonal dodecahedra. The remaining six voids are somewhat larger cavities having fourteen faces. Molecules of methane are able to enter either size void, while a larger hydrocarbon such as ethane can be entrapped only in the larger voids. Other hydrocarbon molecules such as propane and larger are not entrapped in this structure due to their unfavorable size.

A second crystal form, Structure 11, designated as a pentagonal dodecahedral-diamond lattice, is formed in the presence of hydrocarbons larger than ethane. For this structure, a unit cell having 136 water molecules and a cell constant of 17A is present. There are sixteen small pentagonal dodecahedral voids and eight larger voids. Entrapment of propane, n-butane, and isobutane occurs in these structures as well as possible entrapment of ethane and methane. It can be surmised that the absence of propane and heavier molecules would cause only Type I hydrates to be formed, while the presence of both ethane and propane would result in both.


1


HYDRATES

\ It is apparent from the above that molecular size is an important factor. A second important factor is gas solubility. Solubility primarily affects the rate of clathration since it governs the statistical probability that a given molecule will be present when the lattice is closing. Hence, the greater the solubility of a given component, the faster a hydrate will form.

It is believed that the guest molecule dictates which type of structure will be formed, and its solubility governs the rate of formation. Size also affects the formation rate. This may be seen from the hydrate behavior of methane, hydrogen sulfide, and propane. At a given temperature the pure methane hydrate requires a higher pressure to form than the propane hydrate, although its solubility in water is considerably greater than that of propane, It is much easier for a small molecule to avoid entrapment as the cage is closing than a larger one. There are more methane molecules present, but their more active, random movement, and smaller size make clathration more difficult. The similarly sized hydrogen sulfide molecule is far more soluble than methane, forms in the same structure, and requires about one-twentieth the time to hydrate.

Hydrates form in two separate crystal lattices designated as Structures I or I1 shown in Figure 4.8.

Hydrates tend to form at the gas-water boundary, with most of the molecules coming from those in solution in the water phase. Consequently, system components such as valves, fittings, etc. which induce mixing (increasing interfacial area) promote hydrate for-

\ mation.

Hydrates grow like crystals. They build up and plug a line at orifice plates and valves where the full force of the flow stream cannot prevent their build-up. The process is time dependent with the initial precipitation of microscopic crystals occurring first followed by the agglomeration of crystals into a snow-like solid which can plug flow.

Hydrate Fomerc

N2 coz H2S CH4 C2H6

C3H8 IC4H10 nC4H10

l6-Hedron

Figure 4.8 Schematic of Natural Gas Hydrate Lattice

Hydrate forming boundary conditions for hydrocarbon components are shown in Figure 4.9. The hydrate conditions are above and to the left of the curve for the gas component interest. These curves can be modified by the addition of an antifreeze (glycol or methanol). Additional hydrate prediction methods are discussed in Section 20 of the GPSA Engineering Book, including permissible expansion charts and vapor-solid K-values. A hydrate prediction method based on the Peng-Robinson equation of state is available through GPA. All commercial process simulators have EOS based hydrate equilibrium calculation routines.

The diagonal short dash - long dash lines represent the saturation curve for the pure component in question. To the right of the intersection between the hydrate and saturation lines, the hydrate curve becomes essentially vertical. Above this point, liquid hydrocarbon is present along with liquid water (two liquid phases present).

\ For multicomponent systems the effect of the liquid hydrocarbon phase is less significant on the slope of the hydrate cure. This is illustrated in Figure 4.10. Note however, that the hydrate curve is essentially vertical in the dense fluid region.

@Job M. Campbell & C o m p y


4.15


m co P ._

2-

c a 6Y

10 000

7000 5000

3000

2000

1 O00 700 500

300

200

1 O0

70 50

30

20

Temperature. "C O 5 10 15 20 25

I I I I I I 30 40 50 60 70 80 90

Temperature, "F

50.0

20.0

10.0

7.0

5.0

3.0

2.0

1 .o 0.7 0.5

0.3

0.2

0.1

Figure 4.9 Hydrate Forming Conditions for Natural Gas Components

1 1 I 1 I 1

I I t I , I I , I I I 1

I 1 1 I I

Temperature

Figure 4.10 Conditions of Hydrate Formation for a Hydrocarbon Mixture

@John M. Cunpbeii & Company


3

1 1 I' ' il

HYDRATES

, Katz ef u/., Approaches

4

Figure 4.1 1 is an approximate hydrate curve, first published in the early 1940's for sweet, lean pipeline gad4,'). It is useful as a general guideline only. Figure 4.12 is a corresponding chart for predicting the amount of pressure expansion across a valve without hydrate formationf4.'). It assumes that Figure 4.1 1 is correct and that no liquid forms on expansion.

One correlation commonly used recognizes the contribution of gas analysis to hydrate formation condition^.(^.^^-^.'^) It uses a series of vapor-solid (hydrate) equilibrium constants in a manner comparable to a dewpoint calculation. The basic equation is:

c(L] KV-, = 1.0 (4.4)

The term "yi" is the mole fraction of each component in the gas. The value of is shown in Figures 4.13-4.16. The K-values of nitrogen and all non-hydrate forming hydrocarbons are taken as infinity.

In the original work it was assumed that the K-value for n-butane was the same as for ethane. Later work has shown this to be incorrect. Figure 4.16 presents K-values for n-butane. Note that the n-butane K-value goes to infinity at temperatures of 13-15°C [55-6O0F].

Also, the original i-butane curve was based on only two points of experimental data. Figure 4.16 is a new chart for i-butane suggested for use in the Katz

These two changes from the original method are incorporated in the example below.

~~ ~ ~ ~ ~ ~ ~

Example 4.4: Find the hydrate formation temperature of the following gas composition at 20 bar [290 psia].

I I At 10°C 150°F1 Component

Nitrogen Methane Ethane Propane Iso-Butane n-Butane coz

Y¡ 0.094 0.784 0.060 0.036 0.005 0.019 0.002 1 .o00

0.12 0.047

Yif Kv-s

0.384 0.073

0.106 0.090

0.954

The Katz approach has proven reliable for sweet gases up to pressures of about 103 bar [1500 psia]. For gases containing significant amounts of H*S, it should be used, with caution.

Reference 4.13 provides a hand calculation method for gases containing significant quantities of H2S. Although this correlation was developed using HYSIM@' it has proven accurate when compared to the limited amount of experimental data available.

@John M. Cempbeli & Company


4.1 7


8 O 0 0 o O 0 0 o O 0 0 o < D l n P m O N


3


HYDRATES

O O O 0 o 8 g38m 8 s o N 0 0 0 0 0 g 8 o O 0 0 0 o O 0 0 0 o

N g m r n r - <o

O’

8

v

h

Y m



4.19


I Temperature, O F

4.0

3.0

y>

x5 2.0

1 .o O 5 10 15 20 25 30

Methane Temperature, OC

Temperature, O F

Carbon

z Y

Hydrogen

Dioxide

Sulfide

3.0

2.0

1 .o 0.9 0.8 0.7 0.6 0.5

0.4

0.3

0.2 -2.5 0.0 2.5 5.0 7.5 10.0 12.5 15.0

Temperature, "C

Temperahire, O F

I Temperature, "C

Figure 4.13 Vapor-Solid Equilibrium Curves



HYDRATES

~~ ~~~

Temperature, "F 35 40 45 50 55 60 65 70 75 80 a5

2.0

1.5

1 .o 0.9 0.8

0.7

0.6

0.5 ?

Y 0.4

0.3

0.2

n i v. I O 5 10 15 20 25 30

Temperature, "C

Figure 4.14 Vapor-Solid Equilibrium Curves (Ethane)



4.21


Temperature, O F 35 40 45 50 55 60 65 70 75 80 85

Q> c ce n

a L

Temperature, "C

Figure 4.15 Vapor-Liquid Equilibrium Curves (Propane)



3

1 I .' 1 1

HYDRATES

O 0 o O O O > W b ( o v) -s m (u

u ? ? r 9 9 9 9 9 9 00000 o o 8 8 O 0 2 2

'-9



4.23

Y

1000 psia

WATER=HYDROCARBON BEHAVIOR

2000 psia

3 Campbell et u/., Approaches The principle of "additive effects" has been used to develop two manual approaches for par-

ticular application above 69 bar [lo00 p~ia](~.~,"'~). Figures 4.17-4.24 are for hydrate forming molecules. Each figure is for a given pressure. At pressures between those shown, arithmetic interpolation is satisfactory.

The hydrate formation temperature for methane is shown on each figure. The temperature displacement, AT on the abscissa is the amount the hydrate point of methane is changed by a given amount of a component. Carbon dioxide is not shown on these figures and should be ignored as it has no AT effect.

The pentanes and heavier molecules cannot form hydrates because they are too large to fit into the cavities of the hydrate lattice. In fact, experimental data indicates they inhibit hydrate formation to some degree. Figures 4.23-4.24 provide corrections for this inhibiting effect. Thus, the effect of these larger molecules is negative rather than zero, as assumed in the earlier Katz correlations.

Example 4.5 illustrates the use of Figures 4.17-4.24.

Both the Katz and Campbell methods calculate only that portion of the curve between the freezing point of water and the point where the hydrocarbon dewpoint curve intersects the hydrate curve, i.e. no liquid hydrocarbon phase is present. However, if the amount of condefised hydrocarbon liquid is small, say less than 5%, the effect of this liquid phase on the hydrate formation temperature is negligible.

Calculate the hydrate temperature at 69 bar [lo00 psia] and 138 bar [2000 psia] a natural gas with the composition shown below using Figures 4.17-4.24.

Temperature Displacement; 1 A T Mol % 69 bar 138 bar

88.36 6.82 2.54 0.38 0.89 1.01

9.5 2.9 4.6 1.3 0.2

-1.2

Example 4.5:

Component

C1 c2

c3

iC4 nC4 c5+

Total

The temperature displacement for the iCSf fraction can be found from Figures 4.23-4.24.

This procedure shows that pentane and heavier have a calculatable effect on the hydrate point, and the K-value is not infinity.

15.4 2.2 3.2 0.8 O

-0.1 100.00 I 17.3 I 21.5

Temperature Displacement; ATOF

49.10 5.20 8.20 2.30 0.35

-2.20

59.65 3.90 5.80 1.47 O

-0.14 62.95 I 70.68

= 9.5 (Yc,+) (100)

1 - Yc, - Yc,+ Cg+ effect:

0Joh M. Campbell & Company


HYDRATES

Temperature Displacement AT, "C

2

1

Temperature Displacement AT, O F

Figure 4.17 Hydrate Prediction Chart for 69 bar [IO00 psia]



4.25


Temperature Displacement AT, "C

Figure 4.18 Hydrate Prediction Chart for 138 bar [2000 psia]

@John M. CUnpbeU & Compiiny


HYDRATES

Temperature Displacement AT, "C O 1 2 3 4 5 6 7



@John M. Campbd & Company


4.27

WATER-HYDROCARBON BEHAViOR

Temperature Displacement AT, "C O 1 2 3 4 5 6

20.0

10.0 9.0 8.0 7.0

6.0

5.0

4.0

3.0 C O e 2 2 2.0

I" U

c C

1.0

I 0.8 0 0.9

0.7

0.6

0.5

0.4

0.3

0.2

0.1 O 1 2 3 4 5 6 7 8 9 10 11



OJohn M. CampbeU & Company


3

HYDRATES

Temperature Displacement AT, OC O 1 2 3 4 5 6

2

1





4.29


3 20.0

10.0 9.0 8.0 7.0 6.0

5.0

4.0

3.0

2.0

1 .o 0.9 0.8 0.7 0.6

0.5

0.4

0.3

0.2

o. 1


Figure 4.22 Hydrate Prediction Chart for 414 bar [SO00 pcia]

@John M. Cpmpbell& Company

4.30 BP Exploration Company (Columbia) ud.

”,

3

HYDRATES

O 7

v> O

O O0

@Job M. Campbeii & Company


4.31

Computer Predictions

cost Toxicity


less expensive more expensive high low

3

Flash point, "C [OF] Vapor pressure, 37.8"C [100"F] Molecular weight

Advances in computer technology make it possible to incorporate hydrate predictions into EOS calc~lations~~.~~~~.~~~. EOS methods are more general than the empirical methods discussed previously. They also address the presence of a second liquid phase (hydrocarbon). The empirical correlations are based on the assumption that the only liquid phase present in the system is water. Most EOS methods will also calculate the hydrate temperature in the presence of an inhibitor.

12°C [54"] 119°C [247"F] 3 1.9 kPa E4.63 psia] < 1 m m H g

32.1 62.1

HYDRATE INHIBITION

Freezing point, "C ["F] Relative density Relative volatility to H 2 0 @ 1 aún Viscosity, cp @ 25°C [77"F]

a, 600c r 1400~1

The positive manner to prevent hydrates (and corrosion) is to keep the pipelines, process piping and equipment "dry" of liquid water. There are occasions (rightly or wrongly) when the decision is made to operate the system with liquid water present. As long as the system temperature is maintained above the hydrate temperature, hydrate formation is not a problem. This is frequently done by using line heaters or wellhead heaters. However, if the minimum system temperature is below the hydrate point, inhibition of this water is necessary.

The most common inhibitors are equilibrium inhibitors. These are frequently added to the water to depress both hydrate and freezing temperatures. For many practical reasons an alcohol or one of the glycols is used, typically methanol, diethylene glycol (DEG) or monoethylene glycol (MEG). All may be recovered and recirculated, but the economics of methanol recovery may not be favorable in some cases.

-97.6"C [-143.8"F] -13.4"C [7.9'F] 0.796 @ 15°C

3.5 @ 73°C [163"F] 1 .11 @ 25°C

*27.5 @ 138°C [280°F] 0.55 16.9

5.2 -

Methanol may be used effectively at any temperature. I do not recommend DEG generally below about 10°C [14'F] because of its viscosity and the difficulty of separation if liquid hydrocarbons are present. Above 10°C [14"F] it is sometimes preferred because there is less vaporization loss. Since DEG can also be used in an absorption system, some people use DEG as both an inhibitor and absorbent to minimize and simplisr chemical inventory.

Monoethylene glycol (MEG) is by far and away the most popular equilibrium inhibitor in gas processing systems. It has a lower viscosity, molecular weight and cost when compared to DEG. When using an equilibrium inhibitor the choice is generally between MEG and methanol. Table 4.1 below presents a comparison of methanol and MEG properties.

TABLE 4.1 Comparison of Methanol and Monoethylene Glycol

I I Methanol I MEG I

I Flammability I highly flammable at ambient temp. I low flammability at ambient temp. I

1 *HiO to MEG I


3


HYDRATE INHIBITION

Both inhibitors are widely used. MEG tends to be more common for continuous injection due to the high relative volatility of H20 to MEG which makes regeneration easy. Methanol is more difficult to regenerate, but has a much lower viscosity and lower molecular weight which makes it a more effective inhibitor. Methanol's relatively high vapor pressure can result in high methanol losses unless the system temperature is cold, say less than -20°C [-4"F]. Methanol also has a higher solubility in liquid hydrocarbons which also contributes to methanol losses.

Aromatic hydrocarbons tend to be more soluble in MEG than in methanol so the regeneration of MEG can result in releases of benzene, toluene, ethylbenzene (BTEX) and xylenes to atmosphere. This is not an issue in methanol systems.

Finally, MEG's high viscosity makes the physical separation of the rich inhibitor solution from hydrocarbons difficult, particularly at low temperatures. This necessitates long retention times and can result in significant hydrocarbon carryover to the glycol-water phase. Separation of the methanol-water mixture from hydrocarbons is much easier and straight forward. Methanol, however, has a higher solubility in the hydrocarbons than MEG. This sometimes requires a water wash of the liquid hydrocarbon stream to remove soluble methanol.

The degree of suppression (or depression) of the hydrate temperature is directly related to the concentration of inhibitor in the aqueous phase. This is illustrated in Figure 4.25 below which shows the inhibition of a hydrate with methanol in a gas condensate

The relationship between the hydrate suppression and inhibitor concentration can be calculated using thermodynamic methods. For calculations at low inhibitor concentrations (less than 25 wt% for

Temperature, "F

4000

3000

2000

1 o00 800

g 600 500 400

m .-

f?"

!! 300 a v)

200

100 80

60 50 40

30

Temperature, "C

Figure 4.25 Incipient Hydrate Formation for a Typical Gas Condensate Fluid in the Presence of Methanol Solutions




methanol and less than 50 wt% for MEG) the Hammer~chmidt(~."~ equation, Equation 4.5, has proven sufficiently accurate for field calculations.

Equation 4.5(4.'8) has been found reliable for prediction of the necessary inhibitor concentration in the water phase to lower the hydrate point a given amount.

(4.5)

SI Units FPS Units Where: d = suppression of hydrate point

X = weight percent inhibitor in the liquid water phase M = mol wt of inhibitor Ki = constant 1297 2335

As stated earlier, Equation 4.5 is a simplification of the actual thermodynamic relationship between hydrate depression (d) and inhibitor concentration (X). However, when using the constants shown above, Equation 4.5 does an excellent job of matching laboratory equilibrium data for hydrate suppression with methanol solutions up to about 25 wt% and ethylene glycol to about 50 wt%.

Correlating Equation 4.5 with data collected on actual flowing systems is more difficult. On some glycol injection tests the back calculated values for Ki were nearly twice the numbers quoted above.

In truth, no one number applies to all systems, because it is affected by system dynamics, configuration, location and method of injection, etc. Most experienced operators will adjust the injection rate by trial-and-error following initial start-up. This is one of many process calculations that provides little more than a "ballpark" estimate to guide operations.

Equation 4.5 can also be written in terms of the hydrate suppression, d.

S i ) (XI M (1 -X)

d =

Correlations other than Equation 4.5 and 4.6 have been proposed. For methanol concentrations up to about 50 wt% the following equatiod4.I9) may be more accurate.

d = A h ( 1 -xm) (4.7)

SI Units FPS Units

Where: x, d = = mol depression fraction of of hydrate methanol point in the liquid water phase p A = constant 129.6

Note that the inhibitor concentration in Equation 4.7 is expressed as a mol fraction rather than a weight percent.

Maddox, et has also proposed a correlation to estimate hydrate depression versus inhibitor concentration. Based on the limited empirical data available, it appears to be more reliable at high inhibitor concentrations (> 50 wt%).

The quantity "d" is found by first predicting the hydrate forming temperature at the maximum pressure in the line segment being protected. The minimum flowing temperature is then estimated. The quantity "d" is the difference between these two numbers. In the absence of a more definitive


/1

3


HYDRATE INHIBITION

Molecular Weight Density: g/cm3

kg/m3 lb/ft3 ib/gal

number, 4°C [40"F] may be used as a reasonable minimum flowing temperature for buried lines and lines at the bottom of a body of water over 30-40 meters [loo-130 ft] deep.

Total inhibitor injection rate to satis@ the inhibitor concentration needed is found from the equation

32 62 106 0.80 1 . 1 1 1.12

800 1110 1120 49.7 69.4 69.6

6.64 9.28 9.30

Where: mI = mass of inhibitor solution mw = mass of liquid water XR = rich inhibitor concentration XL = lean inhibitor concentration

Vapor Liquid

SI Units FPS Units

ibm wt Yo wt Yo

3.5 1iters/iO6 m3 (std) ' 0.23 1bmMMscf nil nil

The lean inhibitor concentration, XL, is typically 90- 100% for methanol, depending on whether the methanol is new or regenerated, and 6040% for glycols. The rich inhibitor concentration, XR, is determined from Equation 4.5 or 4.7 for field and pipeline inhibition. For inhibition in gas processing plants, the ability to adequately mix the inhibitor and gas in the piping or exchanger is the primary concern. In these cases, the rich inhibitor concentration is set based on two criteria: 1) hydrate suppression and 2) mixing. The mixing requirement usually dominates and often results in a specified dilution (X, - X,) of 5% or less. The concentrations of both the lean and rich glycol are kept in the non-freezing region (see Figure 4.28).

The following properties are useful for the above calculation.

II I MeOH I EG I DEG

Inhibitor Losses to the Hydrocarbon Phase Inhibitor may be lost due to its solubility in the hydrocarbon liquid and vapor phases. For

glycol systems these losses are small. The following provides useful guidelines.

II I Glycol Losses I

Methanol losses are more significant, particularly vapor phase losses.

Figure 4.26 provides reliable estimates of vaporization loss for pressures less than about 69 bar [lo00 psia] and water phase methanol concentrations less than about 40 wt%. Enter the ordinate at the minimum pressure of the line segment. Proceed horizontally to the minimum temperature. From this point read vertically to the abscissa, loss per unit volume of gas flowing divided by X, the weight percent methanol in the liquid phase. Knowing X from Equation 4.5 you can solve for methanol vaporization loss per unit volume of gas flowing.



4.35


40 I I I I I I I

I I O 5 10 15 20 25 30 35 40

kg MeOH per lo6 std m3 Gas (101.325 kPa, 15°C) wt% MeOH in Water Phase

Lbc MeOH per 106ft3 Gas (14.7 psia, 6 0 O F ) wt% MeOH in Water Phase

Figure 4.26 Vapor-Liquid Equilibrium of Methanol over Water

@John M. Camphit & Company


HYDRATE INHIBITION

\ At pressures greater than about 6900 kPa [lo00 psia] vapor losses may be several times higher than those indicated in Figure 4.26, particularly at high methanol concentrations.(4.2')

Methanol solubility in the hydrocarbon liquid phase is low, however, in systems containing substantial amounts of hydrocarbon liquid the total liquid phase losses can be significant. At typical pipeline inhibition conditions, a solubility of about 0.4 kg/m3 E0.15 lbmíbbl] is generally adequate for planning purpose^.(^.^^) This assumes a paraffinic hydrocarbon liquid. Methanol solubility in aromatic hydrocarbons can be 4-5 times higher than this. Figure 4.27 shows the solubility of methanol in NGL streams. It is based on empirical data and was developed by IFP, Institut Francais du Pétrole for their IFPEXOL process. IFPEXOL is a patented process of IFP. To use it either for design, fabrication, or exploitation requires the execution of a licensing contract with IFP.

Temperature, "F -112 -94 -76 -58 -40 -22 -4 14 32

10000

1 O00

100 .. -80 -70 -60 -50 -40 -30 -20 -10 O

Temperature, "C

Figure 4.27 Methanol Solubility in NGL Liquid

Calculation Summary The inhibition calculation procedure may be summarized as follows:

1. Determine the hydrate formation temperature of the gas. 2. Establish the lowest temperature expected in the system. 3. Compute the amount of liquid water present at the temperature in Step 2, using the water

dewpoint at that temperature and a suitable water content correlation. 4. Use Equation 4.5 to solve for "X." In the equation "d" is the temperature in Step 1 minus

that in Step 2. Calculate the mass of inhibitor from Equation 4.8. XR is equal to X from Equation 4.8. (Note, if Equation 4.7 is used to calculate x,, it must be converted to a weight percent, X, before being used in Equation 4.8.)

The volume rate of injection of solution will be the weight of inhibitor per unit time divided by its density, after correcting for concentration.

@John M. CempbeU & Company


4.37


5. If methanol is used, one must correct for that amount lost to the hydrocarbon vapor and liquid phases. Figure 4.26 and 4.27 is used for this purpose. For Figure 4.26, enter the right-hand ordinate at the pressure at which the temperature in Step (2) occurs. Read horizontally to the lowest temperature and then vertically to the abscissa. The denominator of the abscissa value is the value of "X" previously determined from Equation 4.5 (step 4). Losses to the liquid hydrocarbon phase for pipeline conditions may be estimated using a methanol solubility of 0.2 mol% or Figure 4.27 can be used for temperatures below 0°C [32"F].

6. Total injection rate equals that found from Step 5 plus that from Step 4. This is the rate of inhibitor solution.

ixample 4.5: 3.5 x lo6 m3 (std)/d [124 MMscfl of natural gas leaves an offshore platform at 40°C [ 104"FI and 80 bar [ 1 160 psia]. The hydrate temperature of the gas is 17°C [63"F]. The gas aníves onshore at 5°C [41"F] and 65 bar [943 psia]. Associated condensate production is 60 m3/106 m3 (std) [10.7 bbl/MMscfJ. Calculate the amount of 100 wt% methanol and 80 wt% EG inhibitor required to prevent hydrate formation in the pipeline.

Methanol: Step 1 - Hydrate temperature = 17°C Step 2 - Lowest temperature in system = 5°C "dtf = 17 - 5 = 12°C Step 3 - Water content at 40°C and 8000 kPa = 1000 kg/106 m3 (std) (Figure 4.3)

Water content out at 5OC and 6500 kPa = 160 kg/106 m3 (std) (Fig. 4.3)

Water condensed = 3.5 x lo6 std m3 looo - 160

= 2940 kg H20/d E6485 lbm H20/day]

Step 4 - Calculate inhibitor concentration from Equation 4.5

Calculate massofinhibitor required in water phasefromEquation4.8

ml = (2940)(22*8) = 868 kg/d [1915 lbdday] (100 - 22.8)

Step 5 - Calculate losses to the hydrocarbon phases

Vapor - from Figure 4.26 at 5°C [41"F] and 65 bar [943 psia]

3*5 'O6 m3 (stdl ( 22.8 wt% MeOH ) 1 wt% MeOH Vapor Losses =

= 1357 kg/d [29941bm'day] Liquid - use 0.4 kg MeOH/m3 condensate

kg 60 m3 condensate )[ lo6 m3 (std) m3 condensate Liquid Losses =

= 84kg/d



1

3

3

HYDRATE INHIBITION


Total Injection Rate = 868 + 1357+ 84

= 2309 kgíd = 0.12 m3/h [5094 lbmíday, 0.53USgpm]

For 80% EG solution - calculate X from Equation 4.5

12)(62)(100) = 36.5% 1297 + (1 2)(62)

Calculate mass of inhibitor required in the water from Equation 4.8

(2940)(36*5) = 2467 kgíd mI = 80 - 36.5

= 0.095 m3/h [5442 lbmíday, 0.42 USgpm]

Crystallization (Freezing) of Glycols Glycols do not freeze solid but form a "mushy" solution that nevertheless does not flow very

well. The concentration must be such that this is avoided.

Figure 4.28 shows the "freezing point" for the three most common glycols. The curves go through a minimum. At a given temperature the concentration should be kept between the lines. A value of about 75-80 wt % glycol is safe at any likely temperature.

If the minimum system temperature is below 10°C [14"F], the injection rate must satisfjr both Figure 4.28 and Equations 4.5 and 4.8. Typically the lean glycol solution injected will contain 15-25 wt% water. Thus, the total mount of water to be protected is that in the system plus that injected with the glycol. Equation 4.8 accounts for this water.

With EG and DEG the reboiler should not be operated above 150-160°C [300-32O0F] to minimize thermal decomposition.

Glycol Losses The major loss of glycol does not occur from vaporization. It is from losses in the regenera-

tion system, spillage, salt contamination and losses in the separation of oil from the glycol water phase.

Regeneration losses should be small in a well designed unit unless salt contamination tends to plug the still column. Salt is a problem in its own right for the water is distilled off and leaves the salt behind. Salt can be economically removed from glycol only by vacuum distillation reclaiming.



4.39


O

-1 o

-20

-30

-40

- Ethylene -50 - - - - . Diethylene

- - - Triethylene

-60

-40

-58

-76 O 20 40 60 a0 100

wt% Glycol in Aqueous Solution

Figure 4.28 Freezing Points of Glycol-Water Solutions

LOW TEMPERATURE IN GLYCOL INJECTION GAS PROCESSING FACILITIES

Basic System Operation

Figure 4.29 shows a basic glycol injection system used to inhibit hydrates in a low temperature gas processing plant. An inlet separator removes all free water so it is necessary only to inhibit water condensing from the gas. The gas-gas heat exchanger pre-cools the entering gas with cold gas from the low temperature separator. Since the gas temperature will most likely drop below the hydrate point in this exchanger, glycol is injected ahead of it.

A chiller is used to produce cooling. The flow sheet, however, would be very much the same if this chiller was replaced by an expander or expansion valve.

In this configuration, the cold separator is 3-phase with the cold gas going to the gas-to-gas exchanger, the hydrocarbon liquid to stabilization, and the rich glycol solution to the flash tank. The rich glycol is separated from any entrained hydrocarbons and then regenerated for reinjection.

@Job M. Campbeii & Cornpamy

4.40 BP Exploration Company (Cdumbia) Ltd.


f

B E :j m a

t

J



4.41


There are several possible variations of this scheme: 1. A two-phase separator can be used, with the glycol and condensate being separated at the

flash tank. This must be carefully considered as it is difficult to separate these immiscible phases efficiently after they have mixed across the cold separator dump valve. Since glycol is relatively expensive, it is desirable to avoid losses that would be inherent with such an arrangement.

2. Stabilization by distillation may be used on the liquids from the low temperature separator, or the stabilizer may be replaced by stage separation. In this case, supply gas for fuel and general usage is frequently taken from the second stage separator. The glycol-hydrocarbon separator would be the third stage with some gas vented from it also.

If the chiller is a horizontal shell and tube (or plate-fin) exchanger, supplementary injection will be required at the inlet end. Glycol tends to separate from the gas by gravity, hence the upper portions of the exchanger may plug with hydrates. The usual approach is to place one or more nozzles to spray glycol against the tube sheet. This is illustrated in Figure 4.30. An alternative approach is to use a vertical exchanger. The tube sheet provides distribution. Slightly more efficient heat transfer is also obtained.

Glycol G W cuueoi Spray Pattern at Proper Spray Pattern at Hlgh Spray Paitem ai Low

Giycd Flow Rate Gtycd Aow Rate Glycol Flow Rate

1 1 1 t=F

Spray Pattern wiih Proper Node Location

Spray Pattern with Nonle too Clase to Tube Sheet

Spray Pattern vvith Nonle too Far from Tube Sheet

Figure 4.30 Glycol injection Practices

Regeneration

Figure 4.31 is a T-x diagram for L e IEG-I ,O system at 1 atm. This figure may be used to determine the reboiler temperature necessary to produce the required lean MEG concentration. In most glycol injection systems the lean MEG will leave the regenerator at 80 wt% (this is equivalent to about 53.7 mol% MEG). From Figure 4.31 the reboiler temperature required at 1 atmosphere will be about 127°C [260"F]. Lean MEG concentration is controlled by reboiler temperature.

@John M. Campbeii & Compauy

1



220

200

-

- I.--' - ~ - - 110 l- T--- -- - ~- - -

-- ~

- -- ~

I - ~-~

eC ~ ~ ---- . - - ~

- - 100 .. -~ ~ - -

Figure 4.31 Temperature-Composition, T-x, Diagram for MEG-Water System at 1 atm

Carryover of hydrocarbons with the MEG entering the regenerator can contribute significantly to MEG losses. Most of the hydrocarbons easily vaporize in the regenerator effectively lowering the partial pressure of H20 and MEG. This has the same effect as stripping gas in TEG systems. While this does assist in the regeneration of the MEG it can cause MEG losses by literally "stripping MEG" along with the water. This is one reason why glycol-hydrocarbon separation is so important in a glycol injection system.

Very little absorption type gas dehydration occurs in an injection system. The glycol concentration is too low and the gas-glycol contact is concurrent. The water is removed from the gas by condensation. The glycol is injected to prevent solid formation (hydrates, ice). For all practical purposes the water dewpoint of the gas exiting the cold separator will be equal to the cold separator temperature.

Glycol-Water-Oil Separation

All vessels for separating MEG from hydrocarbons (below 0°C [32"F]) should be designed for a minimum residence time of 30 minutes in the glycol phase. This residence time is found by dividing the volume of the vessel by the volumetric flow rate. Above O°C [32"F], a residence time of 20 minutes may be adequate, if foaming or emulsions are not a problem. If possible, it is recommended that this separation be carried out at 30-40°C [86-104"F], which lowers the viscosity and promotes settling. At colder temperatures, separation becomes increasingly difficult.



4.43


Distillate Product Bottoms Product

A full-flow filter comparable to that in a glycol dehydrator is needed to remove solid particles that promote corrosion, degradation, emulsions, and foaming.

98-99 wt% methanol 99.5 wt% water

Glycol Losses In general, the solubility of the glycol in hydrocarbon liquid increases with:

1. An increase in the molecular weight of the glycol (Le., TEG is more soluble than DEG.

2. An increase in temperature. 3. An increase in weight percent of glycol in the water-glycol mixture.

Glycol solubility losses typically are small and are dependent upon the type of hydrocarbon liquids present as well as the type of glycol used. The glycol losses due to true solubility in paraffin hydrocarbon liquid mixtures normally occurring at a separation temperature of 15°C [59'F] and for glycol concentrations of 50-70 wt%, range from 10 to 50 ppm for MEG and 20 to 100 ppm for DEG. This is approximately equal to 7-70 liters of MEG per 1000 m3 of hydrocarbon liquid [0.3-3 US gall1000 bbl] of hydrocarbon liquid. Glycol solubility losses will be lower at lower separation temperatures. Aromatic components increase glycol solubility in the liquid hydrocarbon phase.

If good separation is not obtained, losses can be substantially higher. Total losses of 200-250 liters per 1000 m3 of hydrocarbon liquid [lo US gal/l000 bbl] are not uncommon. Very little vaporization loss occurs at temperatures below 38°C [ 100"FI.

MEG is the least soluble of all.)

Methanol Injection Systems The primary difference between glycol and methanol injection systems is the regeneration of

the inhibitor. Methanol is more difficult to separate from the water phase due to its much lower relative volatility. The two most popular methods of methanol regeneration are distillation and gas stripping (IFPexol).

The regeneration of methanol by distillation is accomplished in an atmospheric distillation tower (actual operating pressure is 0.4-0.7 bar(g) [20-25 psig]. The bottom product is water and the distillate product is methanol. Purity requirements and fractionation parameters vary but typical values are shown below.

II Fractionation Parameters I Theoretical Stages

Reflux Ratio (L/D)

An alternative to methanol fractionation is regeneration by gas stripping. This process is patented under the trade name IFPEXOL@. The IFPEXOL@ process uses a portion of the feed gas stream


3

3

r)



. to "strip" the methanol from the methanol water mixture. A simple flow diagram for IFPEXOL@ is shown in Figure 4.32.

The "COLD PROCESS" depicted in Figure 4.32 can be a refrigeration process similar to the one shown in Figure 4.29 or it can be a valve expansion (LTX, LTS, J-T) or an expander process. In the IFPEXOL@ process the decanted methanol-water solution is returned to the IPEXOL@ contactor (stripper) for regeneration.

The purity of the water leaving the bottom of the stripper is typically less than 500 ppm methanol and depends on the quantity of feed gas routed to the column. No heat is required for the process and no atmospheric venting takes place.

Dry Gas Make-up Methanol

Wet Raw Gas Feed Ij

Stripping Gas

1 - ......

......

- 1

Water 1

IFPEX-1 Contactor

Decanted Methanol-Water Solution Pump

NGL

Figure 4.32 Example lFPEXOL@ Dehydration Process Flow Diagram



4.45


4.1 4.2 4.3

4.4

4.5

4.6

4.7

4.8

4.9 4.10 4.1 1 4.12 4.13 4.14 4.15 4.16 4.17

4.18

4.19

4.20

4.21

REFERENCES GPSA Engineering Data Book, Gas Processors Suppliers Assoc., Tulsa, OK. (1972), p. 15-9. Wichert, G. C. and E. Wichert, "Chart Estimates Water Content of Sour Natural Gas," OGJ (Mar. 29, 1993), p. 61. Campbell, J. M., "Gas Conditioning and Processing," Vol. 1, Ch.6, Campbell Petroleum Series (CPS), Norman, OK, 8th Ed., 2nd Prtg (2001). Maddox, et ai.; "Estimating Water Content of Sour Gas Mixtures," Gas Conditioning Conference, Univ. of Oklahoma, Norman OK (Mar. 1988). Robinson , J. M., et ai., "Estimation of the Water Content of Sour Natural Gases," Trans AIME, Vol. 263 (Aug. 1977), p. 281. Song, K. Y. and R. Kobayashi; RR-50, "Measurement and Interpretation of the Water Content of a Methane-5.31 mol% Propane Mixture in the Gaseous State in Equilibrium with Hydrate," Gas Processors Association (Jan. 1982). Aoyagi, et al.; RR-45, "I. The Water Content and Correlation of the Water Content of Methane in Equilibrium with Hydrates. 11. The Water Content of a High Carbon Dioxide Simulated Prudhoe Bay Gas in Equilibrium with Hy- drates," Gas Processors Association (Dec. 1980). Song, K. Y. and R. Kobayashi; RR-80, "The Water Content of CO;! - Rich Fluids in Equilibrium with Liquid, Water, or Hydrate," Gas Processors Association (May 1984). Maddox, R. N. and J. H. Erbar, "Gas Conditioning and Processing, " Vol 3., CPS (1980). Noaker, L. J. and D. L. Katz, Trans. A M E , Vol. 201 (1954), p. 237. Unruk, C. H. and D. L. Katz, /bid., Vol. 186 (1 949), p. 83. Robinson, D. B. and H. J. Ng, Hydr. Proc., (Dec., 1975), p. 95. Baillie, C. and E. Wichert; Chart Gives Hydrate Formation Temperature for Natural Gas; OGJ (April 1987), p. 37. McLeod, H. O. and J. M. Campbell, Jour. Petr. Tech. (June, 1961), p. 590. Parrish, W. R. and J. M. Prausnitz, iEC Chem. Pro. Design and Dev., Vol. 11, No. 1 (1972), p. 26. Ng, Heng-J. and D. B. Robinson, IEC Fund., Vol. 15, No. 4 (1976), p. 293. Ng, H., Chen, C. J., and D. B. Robinson; RR-92, "The Effect of Ethylene Glycol or Methanol or Hydrate Formation in Systems Containing Ethane, Propane, Carbon Dioxide, Hydrogen Sulfide or a Typical Gas Condensate," Gas Proces- sors Association (Sept. 1985). Hammerschmidt, E. G., "Formation of Gas Hydrates in Natural Gas Transmission Lines," Ind. Eng. Chem., Vol. 26 (1 934), p. 85 I . Nielsen, R. B. and R. W. Bucklin, "Why not use Methanol for Hydrate Control?," Hyd. Proc., Vol. 62, NO. 4 (April 1983), p. 71. Maddox, R. N., et al., "Predicting Hydrate Temperature at High Inhibitor Concentration," Proceedings 1991 Gas Con- ditioning and Processing Conference, Univ. of Oklahoma, Norman, OK. Falk, C. and R. A. Hubbard, private communication.

@John M. CampbeU & Company ~~


Section 5

ABS O R PTI O N TABLE OF CONTENTS

PAGE # ............................................................................................. 5.3 ............................................................................................. 5.5

.............................. 5.5 ............................................................................................. 5.9

Rigorous Calculations ...

Tray Efficiencies ..................................................................................................................................... 5.1 1 Factors Affecting Absorption.. ...... Calculations Based on Operating Data .................................................................................................. 5.14 Distillation and Stripping Equipment ..................................................................................................... 5.15

. . . . . . . . . .

LIST OF FIGURES FIGURE # PAGE # 5.1 5.2 5.3 Kremser-Brown Relationship 5.4 5.5

5.7 5.8

Process Flow Diagram - Typical Refrigerated Lean Oil Plant for NGL Extraction .........__..... 5.2 Typical Process Flow Diagram - Ambient Lean Oil Plant ......... ................................... 5.3

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..................................... 5.6 Typical Process Flow Diagram - Absorber Overhead Presaturator ........................................... 5.9 O’Connell Correlation for Determination of Actual Trays from Theoretical Trays ................... 5.11

Determination of Optimum Lean Oil Quality ................................................................................ 5.13 Checking Efficiencies of Still Operations ....................................................................... ............ ... 5.16

5.6 Effect of Lean Oil Quality on Absorber Performance ....._.._........._ ..... ........... 5.13

@John M. Cunpbcii & Company


5.1

ABSORPTION

? NOTES:


5.ii BP Exploration Company (Columbia) Ltd.

I I- 1 1

.

Section 5

A BSO R PTI O M GL extraction by absorption is called lean oil absorption. The lean oil absorption process was the N dominant method for recovering NGLs until the early 1970's. The process has largely been su-

perseded by the turboexpander method following the development of reliable turboexpander machinery.

A schematic flow diagram of a lean oil absorption plant is shown in Figure 5.1. The gas stream is contacted with a naptha-like liquid hydrocarbon stream (lean oil) in an absorber, and upon contact the lean oil absorbs hydrocarbons from the gas. The heavier components are absorbed most readily and the light component least readily. The actual percent absorption of any one component depends on its volatility at the contacting pressure and temperature, the amount of lean oil (molecules) contacting the gas and the number of contact stages (trays). For a given volume of gas, the lower the temperature and the greater the amount of lean oil circulated, or the more stages the higher the percentage of each stream component absorbed.

*.

The most modern of existing lean oil plants typically operate at contacting temperatures of -30°C to -40°C [-22"F to 40°F] (both lean oil and gas chilled to that level) and with a lean oil of a molecular weight of around 100 to 140. Lean oil circulation rates of around 1 m3 of liquid per 3000 m3 of gas [1 U.S. gal per 400 scfl are common. Recoveries of up to 20-40% ethane, 90 to 95% propane and 100% butanes can be achieved.

The heart of the process is the absorber which operates in the same manner as the glycol absorber. Since considerable methane is coabsorbed along with the ethane and heavier, one or more flashes are applied and a rich oil demethanizer (or deethanizer) is utilized. The demethanized rich oil (lean oil containing absorbed hydrocarbons) is separated into lean oil and recovered hydrocarbons in a still.

Such plants require a large investment in heat exchange equipment, towers and pumps. These are, by nature, large consumers of the fuel required for process heat. The turboexpander option is simpler, less capital intensive, and more thermally efficient.

Two basic types of absorption or lean oil plants are employed in the gas processing industry:

1. Ambient Temperature

Ambient temperature plants were historically the first lean oil plants utilized. They relied on ambient temperature cooling (usually water) for process cooling. Steam is utilized as a stripping medium in the still to help remove the NGL from the oil stream. The molecular weight of the lean oil typically ranges from 140 to 190. As the demand for the lighter NGLs has increased, the ambient temperature plant became uneconomic because of the large lean oil circulation rates required. Figure 5.2 shows a typical ambient temperature lean oil plant.


Technlcal Assistance Service for the üeslgn, Operation, and Maintenance of Gas Plants

5.1

ABSORPTION

”I

@Job M. Campbeii & Company


I

I

RESIDUE GAS

T

- ABSORBER RICH 011 REABSORBER

FLASH RICH OIL T A N K TO LEAN OIL -

HEAT EXCHANGER -

\ Figure 5.2 Typical Process Flow Diagram - Ambient Lean Oil Plant

2. Refrigerated

In the 1950’s and early 60’s’ refrigerated lean oil plants replaced ambient temperature plants as the most common gas processing plants. These plants employ refrigeration systems to cool the lean oil and natural gas streams which increases the recovery of light NGL components and decreases lean oil circulation rates. The molecular weight of the lean oil normally ranges from 100 to 140, depending on the cooling level. Figure 5.1 is a simplified process flow diagram for a typical refrigerated lean oil plant.

Absorber The performance of a lean oil absorber may be modeled using either rigorous or short cut

1. Rigorous (tray-to-tray) methods.

Tray-to-tray calculations are trial and error involving equilibrium, heat, and material balances around each theoretical tray. With the development of process simulators, it has become very easy to evaluate absorption processes using the rigorous methods. A complete absorber calculation can be made in a few seconds. Careful attention must be paid to the characterization of the lean oil.

@John M. CunpbeU & Company


2. Shortcut

ABSORPTION

Short cut methods are extremely useful for desk calculations and for illustrating the rela- I tionships between the various operating variables. All short cut calculation methods are based on the assumption of constant L N and the assumption that the "K"-value of the component being absorbed (solute) remains constant through the absorber.

The Kremser-Brown method is the simplest short cut model. It is generally adequate for parametic evaluations and scoping studies. Other methods such as Kirbride-Bertetti and Edminster have been developed to provide a better estimate of the average L N to be used in the Kremser-Brown equation. These "enhanced" methods are tedious to use and offer little advantage today given the reliability and utility of the rigorous methods.

A listing of the key independent and dependent variables in a lean oil absorber is shown below.

Independent Variables Column characteristics - Number of trays - Tray design

- Rate - Composition - Temperature

- Rate - Composition - TemDerature

Rich gas

Lean oil

Dependent Variables Tray

Vapor from Tray - Efficiencies

- Rate - Composition - Temperature

Liquid from Trays - Rate - Composition - Temperature

The independent variables are those which actually determine the degree of absorption. Each is of prime importance in the performance of an absorber and most may be varied independently of the others. A change in any one of the independent variables will cause a corresponding change in absorber extraction, and will also cause changes in the dependent variables.

The dependent variables are those which may be changed only by changes in the independent variables. They are very important in analyzing the operation of absorbers. Much as the doctor "diag- noses" or "analyzes" a disease by its symptoms or effects, so may the process engineer analyze independent variables by their effects on dependent variables.

Only changes in the independent variables will affect absorber operation, but these changes will always be reflected by changes in the dependent variables. This allows calculation of independent variables under conditions which will permit direct measurement of the variable involved, or calculation of an independent variable needed to produce a desired effect or dependent variable.

None of the generally used absorption calculation methods consider the variations in overall tray efficiency for different components. Although it is not correct, calculation methods make it necessary to use one overall tray efficiency for estimating the absorption of all components. A tray efficiency based on the lightest component being extracted and recovered (usually propane or ethane) often gives the best results. Because of this, absorber models will frequently over predict the absorption of methane and COZ.

@John M. Campbeii & Company I I

5.4 BP Exploration Company (Columbia) Ltd. I

Short-Cut Calculations

The Kremser-Brown equation is rewritten here in the context of a lean oil absorber.

Where: E, = efficiency of absorption YN+, = mols of solute entering absorber per mol of rich gas entering absorber

Yl = mols of solute leaving absorber per mol of rich gas entering absorber Yo = mols of solute in equilibrium with incoming lean oil per mol of rich gas entering

absorber A = absorption factor = W(V,+, K) Lo = lean oil rate, mols per unit time = [(q)(p)]/MW

VN+I = rich gas rate, mols per unit time K = K-value (yíx) of solute at absorber conditions N = number of theoretical stages in absorber

For most lean oil absorbers, Yo = O, and N = 6 to 8 The relationship is presented graphically in Figure 5.3.

The Kremser-Brown method is approximate only, and, since it does not consider irregular changes of WV or K, is limited to those cases where the changes in L/(VK) throughout the column are either small or regular.

One could redo the calculation in Step 3 based on an average L/V of about 0.245 to improve the accuracy, however little is gained by this more tedious approach given the computer methods available.

Rigorous Calculations

Data in the preceding example was used to model a lean oil absorber using rigorous tray-to- tray calculations. The lean oil stream was characterized with twelve pseudocomponents shown below.

Component NBP 194 NBP 223 NBP 251 NBP 276 NBP 298 NBP 323 NBP 347 NBP 368 NBP 400 NBP 428 NBP 451 NBP 478

mol YO 3.83 3.83 8.02

16.95 24.13 18.99 11.45 5.87 1.72 1.73 2.05 1.44

BP, O F

194.5 223.8 25 1.7 276.7 298.9 323.3 347.5 358.8 400.9 428.1 451.1 478.9

Mw 97.5

104.8 112.2 119.3 125.9 133.7 141.8 149.4 161.4 172.2 181.8 194.0

Y 0.717 0.733 0.746 0.757 0.766 0.775 0.784 0.791 0.800 0.808 0.814 0.821

I 100.00

BJohn M. Campbeii & Company


5.5

ABSORPTION

I @John M. CUnpbeU & Company


Example 5.1: Predict the performance of a 20 tray lean oil absorber using the Kremser-Brown method and the information below.

Composition, mol o/o

0.65 0.41

91.09 5.32 1.95 0.20 0.3 1 0.04 0.03 0.01

100.00

MW = 130 y = 0.7703

IBP 5% 10% 20% 30% 40%

60% 70% 80% 90% 95% EP

50%

Distillation Data, OF

242 270 283 294 300 306 3 12 319 328 340 36 1 386 435

Step 1: Calculate L&”+I -

= 1925 lb-mol/hr

100 O00 O00 x f ) ( 1 lb-mol )[ 1 day) = 980 lb-molkr 379.5 scf 24 hr VN+I =

1925 = 0.175 -- - LO VN+I 10980

Step 2: Look up K-values for each component at Tavg = 0°F and P = 700 psia (use 3000 psia convergence pressure charts from GPSA Engr. Data Book).

K 6.00 1.3 3.5 0.48 O. 14 0.054 0.039 0.016 0.012 0.0013

(based on nC,) - @John M. CampbeU & Company


5.7

ABSORPTION

Kample 5.1 (Cont’d.):

Step 3: Calculate the absorption factor (A = LNK) for each component and determine E, from Figure 5.3. Complete the material balance.

mol Y’ 0.65 0.4 1

9 1 .O8 5.32 1.95 0.20 0.3 1 0.04 0.03 0.01

100.00

Feed Gas, lb-moi/hr

71.37 45.02

10000.00 584.10 214.10 2 1.96 34.04 4.39 3.29 1.10

10979.36

K-Value 6.00000 1.30000 3.50000 O .48 O00 O. 14000 0.05400 0.03900 0.01600 0.01200 0.00130

A 0.0292 0.1346 0.0500 0.3646 1.2500 3.2407 4.4872

10.9375 14.5833

134.6154

Ea 0.0292 0.1346 0.0500 0.3640 0.9337 0.9994 0.9999 1 .o000 1 .o000 1 .o000

Rich Oil, lb-moyhr

2.08 6.06

500.00 212.64 199.90 21.95 34.03 4.39 3.29 1.10

Residue, Lb-moi/hr

69.28 38.96

9,500.00 37 1.47

14.20 0.01 0.00 0.00 0.00 0.00

l I I 985.44 I 9993.92 I

ote that the L N at the bottom and top of the tower are greater than the L N based on LO and VN+I.

Top L/V = - 1925 - - 0.193 9994

Bottom 1925+985 = 0.265 LN = 10979

Results of the simulation are tabulated below:

Stage No. Temperature, O F 1 -2.8

Feec mol YO

0.65 0.41

91.08 5.32 1.95 0.20 0.3 1 0.04 0.03 0.01 0.00

0.6 1.9 1.9 0.1

-5.1

Gas lb-molihr

7 1.37 45.02

10001 .o1 584.16 214.12 21.96 34.04 4.39 3.29 1.10 0.00

Residi mol ‘YO 0.73 0.33

95.97 2.94 0.02 0.00 0.00 0.00 0.00 0.00 0.01

! Gas lb-moiihr

69.61 31.28

9 141 .O6 279.67

2.15 0.00 0.00 0.00 0.00 0.00 1 .O7

Ricl mol ‘YO 0.05 0.41

25.44 9.01 6.27 0.65 1.01 0.13 0.10 0.03

56.91

1 9524.83 I 100.00


Oil lb-moYb

1.75 13.73

860.06 304.44 211.94 2 1.96 34.04 4.39 3.29 1.10

1923.94 3380.64

3

i

”,


4 Note that the average absorber temperature is 4 . 6 " F . This is considerably warmer than the -20°F and -25'F incoming gas and lean oil temperatures. This is due to the latent heat of vaporization (heat of absorption) given up by the NGL components as they are absorbed into the lean oil.

Also note the residue gas contains a small fraction of lean oil components (mol fr = 0.0001). This is due to vaporization of the lean oil components on the top tray of the absorber. This lean oil is lost to the sales gas and cannot be recovered. Lean oil losses depend on lean oil MW, IBP, and rate as well as absorber temperature and pressure. The IBP of the oil and absorber top tray temperature are probably the two most important variables in determining lean oil losses.

Presaturation In referring to the previous example, note the quantity of methane absorbed into the rich oil.

This methane (as well as the COZ and sometimes C2) must be stripped from the rich oil before it reaches the still. As we have seen, the absorption of these light components significantly increases the average absorber temperature due to the heat of absorption effects.

Presaturation is a process which allows the lean oil to be "presaturated" with light components by mixing the oil with a lean gas stream - usually residue gas or ROD overhead. This two-phase mixture is then sent through a chiller where the heat of absorption is removed. Any remaining vapors are separated from the oil and the cooled, presaturated oil stream is sent to the absorber. An example of a presaturation process scheme is shown in Figure 5.4.

Figure 5.4 Typical Process Flow Diagram - Absorber Overhead Presaturator



5.9

ABSORPTION

Presaturation has three benefits: 1. increases NGL extraction due to lower absorber temperatures 2. decreases lean oil losses due to significantly lower top tray temperatures 3. reduces compression requirements if a lower pressure vapor (such as ROD overhead) is

used to presaturate the lean oil The disadvantages of presaturation are: 1. increased lean oil chiller duty 2. more complex design - requiring additional equipment and higher operating costs Presaturation is common practice in refrigerated lean oil plants as well as in other physical

solvent processes such as Selexol.

Note the average absorber temperature is lowered to -12°F and in particular the top tray temperature is lowered to -18°F. NGL extraction is increased and lean oil losses are decreased. The chiller duty required to remove the heat of absorption from the lean oil (due to presaturation) is 3.0 MMBtu/hr.

:xample 5.2: Rework the previous rigorous example using a presaturation scheme similar to that shown in Figure 5.4.

Stage No. Temperature, O F

1 -18.3 -14.2 -1 1.3 -9.1 -7.8 -9.1

3

RichOil 7 Feed Gas Lean Oil Residue Gas Comp. lb-mol/hr mol % lb-moilhr mol YO mol % lb-moyhr mol %

0.65 0.4 1

91.08 5.32 1.95 0.20 0.31 0.04 0.03 0.01 0.00

71.37 45.02

1000 1 .o1 584.16 214.12 21.96 34.04 4.39 3.29 1.10 0.00

0.06 0.40

28.14 5.82 0.04 0.00 0.00 0.00 0.00 0.00

65.54

1.76 11.75

826.45 170.93

1.17 0.00 0.00 0.00 0.00 0.00

1924.86

0.68 0.40

94.86 4.04 0.01 0.00 0.00 0.00 0.00 0.00 0.01

71.23 41.93

9931.61 422.89

1.43 0.00 0.00 0.00 0.00 0.00 0.72

0.05 0.43

25.99 9.63 6.20 0.64 0.99 0.13 0.10 0.03

55.82 10980.47 100.00 10469.81 100.00 100.00 2936.93 100.00


5.1 O BP Exploration Company (Columbla) Ltd.

4. Tray Efficiencies Both the tray-to-tray and shortcut calculation methods are based on the theoretical plate con-

cept, which assumes that the vapor and liquid streams leaving the tray are in equilibrium. Equilibrium is not achieved on actual trays for a variety of reasons - primarily inadequate vapor/liquid contact time. Tray efficiencies are used to account for the lack of equilibrium conditions on the tray and may be of two kinds, local and overall. The local efficiency is the ratio of actual change in composition of vapor or liquid going through a tray to the change that would have occurred had the vapor or liquid been in equilibrium with the opposite phase on the tray. While local efficiencies are more rigorous, they are impractical to use in absorber calculations. Accordingly, it is necessary to resort to the overall tray efficiency which is the ratio of the number of theoretical trays or stages to the actual number to accomplish the absorption. If a 20-tray absorber performed in such a manner that its absorption was equivalent to that obtained by four theoretical trays, the overall tray efficiency would be 4/20~100% or 20%. Tray efficiencies are dependent upon:

1. Tray design 2. Properties of the liquid and vapor from the tray 3. The component absorbed

Absorber tray efficiencies are most often correlated in terms of liquid properties and K-values. One correlation which has been used successfully is that of O’Connell. The O’Connell correlation is shown in Figure 5.5.

Hp/P or KMNp

Figure 5.5 O’Connell Correlation for Determination of Actual Trays from Theoretical Trays

Absorber tray efficiencies vary from component to component. To date, it has been impractical to incorporate this phenomenon in calculation methods. Accordingly, it has become acceptable to estimate the overall tray efficiency based on the lightest component to be recovered. In general, with properly designed trays, the use of six theoretical trays gives reasonable and reliable results for most absorbers having 20 or more actual trays (efficiency = 30%).



5.11

ABSORPTiON

Factors Affecting Absorption Effects of changes in absorber operating variables on extraction levels may be predicted from

the Kremser-Brown relationships:

L V K

A = - - A

- 1

AN+ 1

AN+1 Ea =

Only changes in the independent variables will change the extraction in a column. Changes in any of these variables will affect the extraction through the terms L, V, and K. The subsequent effect of increased or decreased extraction levels on the stripping or distillation section of the plant must also be evaluated. A few of the most important dependent variables are discussed below.

1. Temperature and Pressure

The effect of temperature and pressure is accounted for in the K-value. Lower K-values result in higher A's which yield higher extraction levels. Refrigeration is the most common method of minimizing absorption temperature. The choice of a temperature level must be a trade off between higher recovery levels and increased refrigeration costs. Re- frigeration also results in more selective absorption of heavier components. Methane absorption is essentially proportional to lean oil circulation, and since refrigeration lowers the oil circulation requirements for propane and ethane extraction, the extraction of the methane is reduced relatively.

As we have seen, another method of minimizing absorber temperatures is through the use of a presaturated lean oil. Large quantities of methane are absorbed in an absorber generat- ing a tremendous amount of heat, raising the absorber temperature, hence decreasing recovery. However, if the lean oil is presaturated with methane prior to entering the absorber tower, the resulting heat release will be reduced, considerably minimizing absorber temperatures.

Absorber pressure is typically not a variable which can be set independently, since the absorber will usually operate at sales gas pressure. However, a review of the K-value charts in the GPSA Engineering Data Book (use 3000-5000 psia convergence pressure) indicates that ethane and propane K-values actually go through a minimum at a pressure of 600-1 O00 psia and whereas the K-value of methane always decreases with increasing pressure. Therefore, in order to maximize selectivity (C2 and C3 over C1 and COZ) pressure above 1000 psia or below about 600 psia are to be avoided. This is not always possible due to equipment limitations or contractual obligations but should be the goal.

2. Lean Oil Quantity and Quality

Lean oil rates have a profound effect on absorber operation. Generally, the higher the lean oil rate, the higher the extraction levels. It should also be recognized, however, that increased oil rates carry associated costs: (1) higher heating, cooling, and pumping loads, and (2) increased relative recovery of methane which may overload stripping equipment. The optimum rate is determined recognizing the effect of these factors. To insure that the maximum molal rate of lean oil per gallon of circulation is being achieved, it is necessary to circulate the lightest MW oil which is economically feasible. The economic evaluation weighs the advantages of higher "L" values per gallon of circulation vs. the lean oil vaporization losses which become excessive if the oil MW is too low. Lean oil vaporization losses can be estimated by computer simulation or determined by test. Figures 5.6 and 5.7 show the trade-off between higher plant revenues and the cost of lean oil vaporization losses. The source of most of the equilibrium vaporization oil loss is in the first 20 vol%

@John M. Campbeil& Compauy

3

3

~

5.1 2 BP Exploration Company (Columbia) Ud.

I I I I I I I I I I

A0 110 IZO ib 140 IS0 160 170 180 190 200 LEAN O1 L MOLECULAR WE1 GHT

0 Lighter lean oil results in increased absorber recovery and increased lean oil vaporization losses.

1 Absorber Recovery Efficiency oil mol. wt.

1 lean oil mol. wt. Lean Oil Vaporization -

Figure 5.6 Effect of Lean Oil Quality on Absorber Performance \

1 O00

Molecular Weight

Change in Change in Lean

Molecular Weight Maximum Point

Figure 5.7 Determination of Optimum Lean Oil Quality

Ojoha M. Campbell & Company


5.13

ABSORPTION

of the lean oil fraction. Improper still operation is the primary cause of lean oil vaporization losses.

It is obvious that lowering the temperature of the last oil-gas contact will reduce lean oil losses. However, in some lean oil plants such a low molecular weight lean oil is used with low temperatures and good still operation and the equilibrium losses are still unreasonable. In these situations it sometimes becomes economical to install a sponge oil system. In the sponge oil process the top 2 or 3 trays of the absorber are isolated from the rest of the tower and the last oil-gas contact is with an oil having a higher molecular weight. This contact removes most of the light (bulk) absorption oil components from the residue gas. The sponge oil is stripped continuously to remove the light oil and return it to the system. The sponge oil circulation rate is usually 3 to 5 percent of main oil rates.

3. Inlet Gas Quality

Inlet gas quality is important in that the gas must be clean prior to entering the absorber. Carryover into the absorber of sand, drilling mud, glycol, corrosion inhibitors, and crude oil not only reduces absorber efficiency but also fouls the lean oil system with impurities. These impurities (especially heavy crude oil fractions) must be removed from the oil system. This is normally accomplished using a reclaimer. The importance of good front end separation cannot be overemphasized.

4. Mechanical Design

Mechanical design of absorbers is important in two ways. It affects tray efficiency which in turn affects recovery levels, and it can be a cause of lean oil losses due to entrainment (physical carryover of liquid as opposed to equilibrium vaporization). Any mechanical design should be based on sound engineering concepts and should recognize the existence of special factors such as surging, foaming, etc.

Calculations Based on Operating Data

Since most present day engineering is on absorbers already in operation, calculations which are based at least in part on operating data and not on design data should be of interest.

In design work an estimate of overall tray efficiency must be made. This may be based on fairly good data, but it is still an estimate. For an absorber in operation, the overall tray efficiency may be determined. Normally, only those components which appear in appreciable quantities in both the residue gas and the rich oil can be used to determine the tray efficiency with any degree of accuracy. Analytical procedures are less reliable at low concentrations and at low and very high extraction levels, extraction is independent of the number of theoretical trays. Therefore, the lightest component which is being absorbed in the 30-90 percent range is the best guide to overall tray efficiencies. In calculating the efficiency, the actual measured conditions should be used wherever possible. From the absorption equations, the number of theoretical plates may be calculated. The ratio of the number of theoretical to actual plates is the overall tray efficiency.

When estimating changes in extraction levels or other dependent variables resulting fiom changes in independent variables, comparison between actual operating results and theoretically estimated results should be avoided. Instead, two calculations should be made, one for the base condition and one for the changed condition. The percentage change in the dependent variable from these calculations may then be safely applied to actual operating data with a reasonable assurance of accuracy.



1 Distillation and Stripping Equipment This section of a lean oil absorption plant is the area where:

1. The light impurities (methane, C02, and possible ethane) are removed from the rich oil.

2. The NGL is separated from demethanized rich oil stream.

The equipment used in these processes is tabulated below:

Although calculations around this equipment will not be specifically discussed, it is important to remember that any change of independent variables in the absorption section must be evaluated in light of its effect on the stripping or distillation section. For example, an increased L N in the absorber will result in higher C4C2 or (Cl + C2)/C3 ratios in the rich oil, making operation of the demethanizer more difficult. In addition, increased recovery of ethane relative to the other products may prevent condensation of the NGL product, resulting in product losses to flare from the feed tank (still reflux accumulator).

Heat to ROD reboiler is usually supplied by the hot lean oil returning from the still bottoms. As Cl/C2 or ( C , + C2)/C3 ratios increase, additional reboil heat is required. Frequently in these cases

tion. In this case, the ROD pressure must be lowered or NGL product must be artificially injected to the bottom of the ROD to provide stripping vapors. An alternative solution is to lower the lean oil rate at the absorber and reduce light component absorption.

. the heat supplied to the reboiler by the hot lean oil is insufficient to meet the ROD bottom specifica-

Perhaps the most important equipment in the entire lean oil plant is the still. This is where the NGL product is separated from the lean oil. In refrigerated lean oil plants, the natural gasoline portion of the NGL product and the lean oil contain significant quantities of the same components. The still is that point in the system where the separation between these components is achieved. In most lean oil plants, lean oil is actually manufactured in the still by adding a portion of the C7+ components from the feed gas. Some of this "manufactured" lean oil is stored for future requirements, the rest is sent to the NGL product stream.

Still operation requires control of three variables:

lean oil MW lean oil IBP lean oil FBP

1. Lean oil MW

The molecular weight of the lean oil is controlled by manipulating the still bottoms temperature. This is typically done by raising the temperature of the still bottoms heaters which serve as the still reboiler. As we have seen in an earlier discussion, there is an optimum molecular weight for the oil which balances the income from additional NGL extraction against the expense of lean oil losses at the absorber.



5.1 5

ABSORPTION

400 -~ LL 350

2. Lean oil IBP

The IBP of the lean oil is controlled by manipulating the still reflux. Increasing reflux increases the lean oil IBP. The IBP of the oil is probably the most important variable in determining lean oil losses. Even if the lean oil MW is at the optimum, excessive lean oil losses can be experienced when the LBP is too low. In general, the IBP of the oil should be kept within about 25°F of the natural gasoline product final boiling point (FBP). This is illustrated in Figure 5.8. The efficiency of still operations is checked by performing an ASTM D86 distillation on the oil and NGL product.

LEAN O11 --------

LEAN OIL ASTM.

3

g 250 z I- 200

t 50

W

I O 0 I 1 I I I 1 b

"Good" Stiü Operations

1. Lean oil initial (IBP) is within 25°F of product end point (FBP). 2. Lean oil mol weight is the desired system mol weight.

Figure 5.8 Checking Efficiencies of Still Operations

3. Lean oil FBP

The final boiling point of the lean oil is controlled by use of an additional piece of equipment called a reclaimer. The reclaimer is a small atmospheric still which is used to remove heavy non-volatile components from the lean oil. The light lean oil components are vaporized in the reclaimer and sent back to the lean oil system. The heavy components are removed from the bottom of the reclaimer and blended with crude oil or sold as "waste oil." The reclaimer operation should be targeted to yield a lean oil FBI which is about 100-120°F above the IBM.

@Joan 1. CampbeU & Company


Section 6

f-

n

G 1Y CO 1 DEHYDRATION TABLE OF CONTENTS

PAGE # THE BASIC GLYCOL DEHYDRATION UNIT .... ................................... ........................................ 6.2 BASIC PROCESS DESIGN FACTORS ............................................................................ MINIMUM LEAN TEG CONCENTRATION ...................... ......................................................... 6.8

Absorber Design .........................

.....................................

Still Column .................. Glycol-Glycol (Lean-Ri

GLYCOL FLASH VESSEL ........................................................... ......................................................... 6.30

GLYCOL CIRCULATION PUMPS ............................................. ......................................... 6.31 ..............................................................

............................................ 6.33

........................................................................................... 6.34 Corrosion-Erosion .......... ................................................ ............................................ 6.34

REFERENCES .................................................. ............................................. 6.36 APPENDIX 2A ...... .............................................................................................................................. 6.37

..................

........................................................................................... 6.51

LIST OF FIGURES

FIGURE # PAGE # 6.1 6.2 6.3 6.4 6.5 .......................................................................................................................................................... 6.12 6.6 6.7

Basic Glycol Dehydration Unit ...................................................................................................... 6.3 TEG Regeneration Alternatives ...................................................................................................... 6.5 Flow Sheets for Two Different TEG Dehydration Systems ......................................................... 6.6 Equilibrium H20 Dewpoint vs. Temperature at Various TEG Concentrations ........................... 6.9

Eficiency of Absorption, Ea vs. Absorption Factor, A ................................................................ 6.14 Activity Coefficient for H20 Concentration at Various Temperatures ........................................ 6.15

0John M. Campbeii & Company


6.1

GLYCOL DEHYDRATION

LIST OF FIGURES (CONT'D.)

FIGURE # PAGE #

6.8 6.9 6.10 6.1 1 6.12 6.13 6.14 6.15 6A . 1 6A.2 6A.3 6A.4 6A.5 6A.6 6A.7 6A.8 6A.9

Equation 6.5 Mol Fraction H20 vs . TEG Concentration .............................................................. 6.16 Bubblecap, Random and Structured Packin ................................................ 6.18 Example Structured Packed Absorber ......... Effect of Stripping Gas on TEG Concentration for Regenerators with S Recommended Still Column Top Temperatures ............................................................................ 6.28 Cutaway and Operation of a Kimray Glycol Pump ....................................................................... 6.32 Solubility of Hydrogen Sulfide in TEG at Various Partial Pressures .................. C02 Solubility in TEG 3.5 wt% H20 ....................................... Densities of Aqueous Ethylene Glycol Solutions (percent by Densities of Aqueous Diethylene Glycol Solutions (percent by weight) ..................................... 6.38 Densities of Aqueous Triethylene Glycol Solutions (percent by weight) .. Densities of Aqueous Tetraethylene Glycol Solutions (percent by weight) ................................ 6.39 Densities of Aqueous Propylene Glycol Solutions (ipercent by weight) ..................................... 6.40 Total Pressure over Aqueous Ethylene Glycol Solutions vs . Temperature .................................. 6.40 Total Pressure over Aqueous Diethylene Glycol Solutions vs . Temperature ........... Total Pressure over Aqueous Triethylene Glycol Solutions vs . Temperature ............................. 6.41 Total Pressure over Aqueous Propylene Glycol Solutions vs . Temperature ................................ 6.42

6A . 1 O Viscosities of Anhydrous Glycols ...................... ..................................................................... 6.42 6A . 1 1 Viscosities of Aqueous Ethylene Glycol Solutio ......... 6.43 6A . 12 Viscosities of Aqueous Diethylene Glycol Solutions .................................................................... 6.43 6A . 13 Viscosities of Aqueous Triethylene Glycol Solutions ................................................................... 6.44 6A . I4 Viscosities of Aqueous Tetraethylene Glycol Solutions ............................................................... 6.44 6A.15 Viscosities of Aqueous Propylene Glycol Solutions ..................................................................... 6.45 6A.16 Heat Capacity of Anhydrous Glycols ............................................................................................. 6.45 6A.17 Water Vapor Dewpoints over Aqueous Ethylene Glycol Solutions ............................................. 6.46 6A.18 Water Vapor Dewpoints over Aqueous Diethylene Glycol Solutions ......... 6A . 19 ........................... 6.47 6A.20 Water Vapor Dewpoints over Aqueous Propylene Glycol Solutions ........................................... 6.47 6A.21 Vapor-Liquid Composition Curves for Aqueous Ethylene Glycol Solutions

........................... 6.48 6A.22 Vapor-Liquid Composition Curves for A

(760 minHg Pressure) ............................... ............................................................................... 6.48 6A.23 Vapor-Liquid Composition Curves for Aqueous Triethylene Glycol Solutions

(300 mmHg Pressure) ...................................................................................................................... 6.49 6A.24 Vapor-Liquid Composition Curves for Aqueous Triethylene Glycol Solutions

(600 mmHg Pressure) ...................................................................................................................... 6.49 6A.25 Vapor-Liquid Composition Curves for Aqueous Triethylene Glycol Solutions

(760 mmHg Pressure) ...................................................................................................................... 6.50 6B . 1 6B.2 6B.3 6B.4 6B.5

Water Vapor Dewpoints over Aqueous Triethylene Glycol Solutions ........

(760 mmHg Pressure) ...............................

Water Removal vs . TEG Circulation Rate at Various TEG Concentrations (N = 1.0) ............... 6.51 Water Removal vs . TEG Circulation Rate at Various TEG Concentrations (N = 1.5) ............... 6.52 Water Removal vs . TEG Circulation Rate at Various TEG Concentrations (N = 2.0) ............... 6.53 Water Removal vs . TEG Circulation Rate at Various TEG Concentrations (N = 2.5) ............... 6.54 Water Removal vs . TEG Circulation Rate at Various TEG Concentrations (N = 3.0) ............... 6.55

LIST OF TABLES

TABLE # PAGE #

6A.1 Physical Properties of Glycols ........................................................................................................ 6.37

@John M . Campbell & Company

r?

6.11 ~ ~~

BP Exploration Company (Columbia) Ltd .

Section 6

GLYCOL DEHYDRATION ehydration is the process of removing water from a gas and/or liquid so that no condensed water D is present in the system. Inhibition is the process of adding some chemical to the condensed

water so hydrates cannot form.

Dehydration is preferred, if economically and mechanically feasible, because it prevents water from condensing in the system. This system is the source of both hydrates and corrosioníerosion problems. In some instances, inhibition may be the preferred process, particularly in sweet systems employing moderate refrigeration (-40°C [-40°F] and above).

Natural gas is commercially dehydrated in one of three ways.

1. Absorption Glycol Dehydration 2. Adsorption Mol Sieve or Silica Gel 3 . Condensation Refrigeration with Glycol Injection

Glycol dehydration (absorption) is the most common dehydration process used to meet pipeline sales specifications and field requirements (gas lift, fiiel, etc.). Adsorption process are used to obtain very low water contents (0.1 ppm or less) required in low temperature processing such as deep NGL extraction and LNG plants. Condensation is commonly used as a dehydration process when moderate levels of refrigeration are employed or in pipeline transportation. An inhibitor such as ethylene glycol (EG) or methanol is used to prevent hydrate formation, but it should be noted that the actual extraction mechanism is condensation.

Four glycol are used for dehydration andor inhibition:

1. Monoethylene glycol (MEG) 2. Diethylene glycol (DEG)

3. Triethylene glycol (TEG) 4. Tetraethylene glycol (TREG)

Triethylene glycol (TEG) is the most common glycol used in absorption systems. Monoethylene glycol (MEG) is the most common glycol used in glycol injection systems. All glycol are hydroscopic, which means they have an affinity for water.

The basic properties of these glycols are shown in Appendix 6A. The major properties govern-

1. Viscosity 2. Vapor pressure 3. Solubility in hydrocarbons

ing the choice of glycol for a given application are:



6.1

R 1

GLYCOL DEHYDRATION

In absorption dehydration systems the solvent (glycol) should be hydroscopic, noncorrosive, non-volatile, easily regenerated to high concentrations, non-soluble in liquid hydrocarbons, and unreac- tive with hydrocarbons, COZ and sulfur compounds.

7 Several of the glycols come close to meeting all of these criteria. Diethylene (DEG),

triethylene (TEG) and tetraethylene (TREG) glycols all possess suitable traits. However, almost 100% of the glycol dehydrators use TEG.

DEG is somewhat cheaper to buy and sometimes is used for this reason. It is also sometimes used as an inhibitor in addition to an absorbent. However, by the time it is handled, stored, and added to the units there is often no real savings. Compared to TEG, DEG has a larger carry-over loss, offers lower dewpoint depression, and regeneration to high concentrations is more difficult. For these reasons, it is difficult to justify a DEG unit, although a few are built each year.

TREG is more viscous and more expensive than the other processes. The only real advantage is its lower vapor pressure which reduces absorber carry-over loss. It may be used in those relatively rare cases where glycol dehydration will be employed on a gas whose temperature exceeds about 50°C [122"F].

In recent years some units have employed propylene glycol (PG). PG is the least toxic glycol and has a lower affinity for aromatics, but PG has a much higher vapor pressure than TEG, and a much lower flash point.

This chapter will concentrate on TEG, even though property data are shown in Appendix 6A for several glycols. Some of the system characteristics apply also for all glycols.

THE BASIC GLYCOL DEHYDRATION UNIT Figure 6.1 shows the basic glycol unit, regardless of the glycol used. Not shown is any cooling

equipment that may be a part of the dehydration processes. When it is possible to cool the entering wet gas with air or suitable water ahead of the absorber, do so. Such cooling is the least expensive form of dehydration.

The entering wet (rich) gas, free of liquid water, enters the bottom of the absorber (contactor) and flows countercurrent to the glycol. Glycol-gas contact occurs on trays or packing. Bubble cap trays have been used historically but structured packing is more common today. The dried (lean) gas leaves the top of the absorber.

The lean glycol enters on the top tray or at the top of the packing and flows downward, absorbing water as it goes. It leaves rich in water.

It is convenient to use the word "rich" to describe the bottom of the absorber and the word "lean" for the top. At the bottom, both the entering gas and glycol leaving are rich in water; at the top end they both are lean in water.

The rich glycol leaves the bottom of the absorber and flows to a reflux condenser at the top of the still column. The rich glycol then enters a flash tank where most of the volatile components (entrained and soluble) are vaporized. Flash tank pressures are typically 300-700 kPa [44-102 psia]. Leaving the flash tank the rich glycol flows through the glycol filters and the lean-rich exchanger where it exchanges heat with the hot lean glycol. The rich glycol then enters the still column where the water is removed by distillation.

7 @John M. Campbeii & Company


THE BASIC GLYCOL DEHYDRATION UNIT

I-

2 2 rn

rn rn

B a n

Q

t a

- - - I I I I I

I



6.3

GLYCOL DEHYDRATION

1 Glycol

7 The still column and reboiler are often called the regenerator or reconcentrator. This is where the glycol concentration is increased to the lean glycol requirement.

The regeneration unit shown is designed to operate at prevailing atmospheric pressure. The initial thermal decomposition temperatures of the glycols are shown in the table below:

I Decomposition Lean Glycol Temperature Cone., wt%

EG DEG TEG

TREG PG

165°C [329"F] 164°C [328"F] 206°C [404"F] 238°C [460"F]

96.0 97.1 98.7 -

These are the temperatures at which measurable decomposition begins to occur in the presence of air. DEG is no more stable than EG because it pyrolyzes in contact with carbon steel.

In the normal unit containing no air (oxygen), it has been found that one can operate the reboiler very close to the above temperatures without noticeable decomposition. The composition of the lean glycol is set by the bubblepoint composition at the regenerator pressure. Maximum concentrations achievable in an atmospheric regenerator operating at decomposition temperature are also shown in the above table.

If the lean glycol concentration required at the absorber (to meet the dewpoint specification) is higher than the maximum concentrations above, then some method of increasing the glycol concentration at the regenerator must be incorporated in the unit. Virtually all of these methods involve lowering the partial pressure of the glycol solution either by pulling a vacuum on the regenerator or by introducing stripping gas into the regenerator.

9

A typical stripping gas system is shown in Figure 6.2(a). Any inert gas is suitable. A part of the gas being dehydrated, or exhaust from a gas-powered glycol pump (if used), is suitable. The quantity required is small. The stripping gas may be introduced directly into the reboiler or into a packed "stripping column" between the reboiler and surge tank. In theory, adding gas to a packed unit between the reboiler and surge tank is superior and will result in lower stripping gas rates. If introduced directly to the reboiler, it is common to use a distributor pipe along the bottom of the reboiler.

A second stripping gas alternative is close the stripping gas loop and use a material like iso-octane, as shown in Figure 6.2(b). It vaporizes at reboiler temperature but can be condensed and separated from the water in a 3-phase separator. The stripping solvent is then pumped back to the regenerator to complete the stripping loop. Sold under the trade name DRIZO@ , this unit has the advantage of providing very high stripping gas rates with little or no venting of hydrocarbons. Glycol concentrations in excess of 99.99 wt% have been achieved with the DFUZO@ process. It has an added advantage of condensing and recovering aromatic hydrocarbons from the still column overhead. In fact, these units often operate with a stripping solvent which is not iso-octane but a mixture of aromatic, naphthenic and paraffin hydrocarbons in the Cs-Cg range.

Figure 6.2(c) shows a third regenerator alternative called a COLDFINGER@. The COLDFIN- GER' process achieves glycol enrichment by passing rich TEG through a cool "finger" inserted in the surge tank vapor space. This condenses a water-TEG mixture which is very rich in water. This mixture is drawn out of the surge tank by means of a trough below the "coldfinger" and is recycled back to the regenerator. The H2O partial pressure in the vapor space is thus lowered and the lean glycol

7



THE BASIC GLYCOL DEHYDRATION UNIT

Vent Gases to Flare

a) Stripping gas Flue Gas

Glycol Pump

b) DRIZO'

Vent Gases lo Rare orRecycle FlueGas

1

--El Glycol Pump

C) COLDFINGER'

Coding Medium

Waier Rlch TE0 Mixhire To still Column

Figure 6.2 TEG Regeneration Alternatives

concentration increased. Lean TEG concentrations of 99.5-99.9 wt% have been achieved in COLD- FINGER@ units without the use of stripping gas, although a small amount of gas is introduced into the surge tank for pressure balancing.

The unit shown in Figure 6.1 is typical. Figure 6.3 shows examples of two systems using TEG that incorporate additional features.

The upper flow sheet is for an offshore unit. The inlet scrubber is in the bottom of the absorber. Three-phase separation is required. The gas rises through a "chimney tray" to the absorber. The hydrocarbon and water are separated as shown. Three-phase separation saves on deck space and is less expensive, but many of the existing units are unsatisfactory because they provide inadequate separation. This is particularly true when the absorber utilized structured packing.

The rich TEG from the chimney tray goes to a degassing pot (flash tank) which is operated at a In some systems the pressure is high enough pressure to send the gas to fuel or recompression.

@John M. Campbell 8s Company


6.5

,.

GLYCOL DEHYDRATION

Pipeline 8 Wells to shore (50 MMscfd each)

(400 MMscfd)

Dried Gas

Entrainment Separator A

Wet Gas 4 o c)

c 8 s S

8 Separator-Conductors Sump Tank

16 Glycol Pumps (8 operating, 8 spares)

Blanket Gas

3 Glycol Regenerators

Water

I Heat t

Fuel I

Figure 6.3 Flow Sheets for Two Different TEG Dehydration Systems



BASIC PROCESS DESIGN FACTORS

r'

r"

6"

sufficient merely to enter the main flare system. The purpose of this is several fold: (1) use or dispose safely any volatile components picked up by the TEG in the absorber, and (2) minimize the presence of corrosive sulfur compounds and carbon dioxide in the high temperature reboiler.

The true solubility of paraffin hydrocarbons is very low in the glycols. But, separator carryover and entrainment does introduce hydrocarbons into the rich glycol. Many of these are "heavier" than air and can be a safety problem unless disposed of properly. In addition, aromatic components are very soluble in TEG. These can also be a safety concern when discharged to atmosphere at the top of the still column.

Both sulfur compounds and carbon dioxide are very soluble in water and react to some degree with the glycols. The degassing in the flash tank prior to the stripping column reduces their concentration and minimizes high temperature corrosion. This degassing is more efficient if the rich glycol is preheated first, as shown in the lower flow sheet in Figure 6.3.

The glycol cooler is a gas-TEG exchanger in the upper flow sheet, as opposed to cooling medium in the lower one. Offshore it is suitable to use sea water cooling in plate exchangers, provided treated sea water is available for other purposes. In temperate latitudes aerial cooling also is a viable alternative.

A forced convection rich-lean glycol exchanger is shown in the lower flow sheet and is preferred if fuel is expensive. Close temperature approaches reduce reboiler heat load. Plate exchangers are widely used in this service. In the upper (offshore) flow sheet the rich-lean exchanger is a coil in the surge tank. This is inexpensive to build but results in poor heat transfer and higher reboiler duties.

Both flow sheets show a gas fired reboiler. The use of hot oil, steam, waste heat or electrical resistance coils are all suitable if they are readily available at the site. Oftentimes the use of electrical resistance offshore is cost-effective and safe.

No filter is shown on the upper flow sheet. It is on the low pressure side of the flash tank in the lower one. Location, pressure-wise, obviously affects cost. I prefer locating the filter at some point ahead of the reboiler to minimize the "gunk" accumulating therein. For effective operation it is imperative that full-flow, glycol filters be installed in the system.

BASIC PROCESS DESIGN FACTORS All factors controlling the behavior of absorption systems also apply for TEG dehydration. In

fact, from a process viewpoint, TEG is one of the simpler absorption processes being employed in the petroleum industry.

In order to properly design a unit one needs to know maximum and minimum gas flow rate, maximum and minimum temperature and pressure, gas composition and required water dewpoint or water content of the outlet gas. From these one can calculate:

1. The minimum concentration of TEG in the lean solution entering the top of the absorber required to meet outlet gas water specification.

2. The lean TEG circulation rate required to pick up from the gas needed amount of water necessary to meet the outlet gas water content specification.

3. The amount of absorber contact required to produce the necessary approach to equilibrium required in (1) above at the chosen circulation rate.



6.7

GLYCOL DEHYDRATION

-! To obtain these answers it is necessary to have a vapor-liquid equilibrium correlation for a

TEG-water system. From this basic input, one can size equipment and develop mechanical specifications.

The procedure that follows is straightforward and can be performed manually. In all but a few exceptional applications, it will give results as reliable as more rigorous (process simulator) methods. Following an outline of the basic calculation procedure, each major equipment component will be reviewed.

MINIMUM LEAN TEG CONCENTRATION If water-saturated gas is placed in a static cell with a given concentration of TEG-water solu-

tion at a fixed P and T, equilibrium would be attained in time. Assuming the liquid had a sufficiently lower water concentration, water would transfer to this liquid from the gas. At equilibrium, the mol fraction water in the gas divided by its mol fraction in the liquid equals the K value for this system.

Figures 6.4(a&b) are based on equilibrium data published by Parrish, et. Several equilibrium correlations@ 3-6 5h7) have been presented since 1950. Previous editions of the Campbell texts and the GPSA engineering Data Book presented an equilibrium correlation based on the work of Worley(6.1). In general, the correlations of Worley(b.l), R o ~ m a n ( ~ ~ ' ~ and Parri~h('.~) agree reasonably well and are adequate for most TEG system designs. All are limited by the ability to measure accurately the equilibrium concentration of water in the vapor phase above TEG solutions, The Parrish correlation has been included in this edition because equilibrium water concentrations in the vapor phase were determined at infinite dilution (essentially 100% TEG). The other correlations use extrapo- lations of data at lower concentrations to estimate equilibrium in the infinite dilution region. The effect of pressure on TEG-water equilibrium is small up to about 13 800 kPa [2000 ~ s i a ] ~ ~ . ~ ' .

-I

A TEG absorber is essentially isothermal. The heat of solution is about 21 kJ/kg [91 Btu/lbm] of water absorbed in addition to the latent heat. But, the mass of water absorbed plus the mass of TEG circulated is trivial to the mass of gas. So, the inlet gas temperature controls. The temperature rise due to heat of absorption seldom exceeds 1-2°C [2-4'F] except when dehydrating at pressures below about 1000 Wa [145 psia]. In low pressure service some temperature adjustment may be desirable.

The diagonal lines represent weight % TEG in a TEG-water mixture entering the top of the absorber. What is the lowest water dewpoint one could attain with a given concentration at a given temperature?

Example 6.1: What equilibrium water dewpoint could be obtained at 40°C [104"F] with a lean glycol solution containing 99.5 wt% TEG?

In Figure 6.4 locate 40°C [104"F] on the abscissa, go vertically to the 99.5 wt% line and then horizontally to the ordinate. Read -19°C [-2"F].

This water dewpoint could be attained in a test cell but not in a real absorber. The gas and TEG are not in contact for a long enough time to reach equilibrium. In addition, the gas theoretically leaves the top tray of the absorber in equilibrium with the TEG leaving the tray, not entering. Numer- ous tests show that a well designed, properly operated unit will have an actual water dewpoint 5.5- 8.5'C [10-15'F] higher than the equilibrium dewpoint. This "approach" to equilibrium depends on the glycol circulation rate and number of contacts in the absorber and is used to speciSl minimum lean glycol concentration. The procedure is as follows.



r

r"

MINIMUM LEAN TEG CONCENTRATION

c>

a c S .-

3 al O

.- f L o - .- 3 U W

40

20

O

-20

-40

-60

20 30 40 50 60 Contactor Temperature, "C

Contactor Temperature, OF

Figure 6.4 Equilibrium H20 Dewpoint vs. Temperature at Various TEG Concentrations



6.9

w I I

GLYCOL DEHYDRATION

1. Establish the desired outlet water dewpoint needed from sales contract specifications or from minimum system temperature.

2. Subtract the approach from (1) to find the corresponding equilibrium water dewpoint. 3. Enter the value in (2) on the ordinate of Figure 6.4 and draw a horizontal line. 4. Draw a vertical line from the inlet gas temperature on the abscissa. 5 . The intersection of the lines in Steps (3) and (4) establishes minimum lean TEG concentra-

tion required to obtain the water dewpoint in Step (1). If water content is specified or calculated in mass per unit gas volume, a water content, pres-

sure, dewpoint temperature correlation is required. Note that the equilibrium water dewpoints on the ordinate of Figure 6.4 are based on the assumption the condensed water phase is a metastable liquid. At low dewpoints the true condensed phase will be a hydrate. The equilibrium dewpoint temperature above a hydrate is higher than that above a metastable liquid. Therefore, Figure 6.4 may predict dewpoints which are lower than can actually be achieved. The difference is a function of temperature, pressure and gas composition but can be as much as 8-12°C [15-2O0F]. When dehydrating to very low dewpoints, such as those required upstream of a refrigeration process, the TEG concentration must be sufficient to dry the gas to the hydrate dewpoint.

Example 6.2: The gas sales contract specifies an outlet water content of 100 kg/106 std m3 [6 lbm/MMscf3 at a pressure of 69 bar [lo00 psia]. The inlet gas temperature is 40°C. What minimum lean TEG concentration is required?

SI Units: For 100 kg/106 std m3 and 69 bar, the equivalent dewpoint from a water content correlation is -2°C. If we use an 8°C approach the equilibrium dewpoint is -10°C. From Fig. 6.4(a) at -10°C and 40°C contact tern- perature, wt% TEG = 99.0.

FPS Units: At 104°F and 1000 psia the dewpoint is 28°F for a water content of 6 1bMMscf. An approach of 14.4"F gives an equilibrium temperature of about 14°F. From Fig. 6.4(b), lean TEG concentration equals 99.0 wt%.

The dashed line in Figure 6.4 at about 98.5 wt% represents the concentration of lean TEG that can be produced routinely in a regenerator operating at standard atmospheric pressure and 204°C [400°F]. This is a safe value for design and specification purposes. Concentrations of 98.7-98.8 wt% are common; some to 99.1 wt% have been reported but represent a special case where incoming hydrocarbons provided natural stripping andíor the pressure was lower than standard atmospheric.

Since the initial capital cost of ordinary gas stripping accessories is trivial, they always should be included. Conditions can change to where they may be required.

It is necessary to fix a lean TEG concentration for subsequent calculations. For the first consideration, use the results from Figure 6.4. If the concentration obtained is less than 98.5 wt%, use 98.5 wt% for the calculation unless you plan to reduce the reboiler temperature below 204°C [400"F].

The minimum lean TEG concentration may not be the one used. A higher concentration than this may be specified to minimize circulation rate and optimize cost.


6.1 O BP Exploratlon Company (Columbla) Ltd.


Absorber Design The absorber (contactor) is where the water is removed from the gas by the process of physical

absorption. The amount of water removed depends on three factors:

0 glycol concentration 0 glycol circulation rate

number of contacts in absorber

The effect of glycol concentration on water removal is illustrated in Figure 6.4. As the TEG concentration increases, the equilibrium water dewpoint decreases. In an absorber this would mean a lower water concentration in the gas at the absorber outlet, hence higher water removal.

In a real absorber the gas leaving the absorber does not reach equilibrium with the incoming lean TEG. This is the reason for the "approach" discussed in the previous section and applied in Example 6.2. Approach depends on two parameters - circulation rate and number of contacts. In other works, the theoretical outlet water dewpoint (concentration) is set by the lean glycol purity, the actual water dewpoint depends on the circulation rate and number of trays or height of packing. This is illustrated in Figure 6.5.

As you can see in Figures 6.5(a-c), adding contacts or increasing the circulation rate narrows the approach. If we have an infinite number of contacts or an infinite circulation rate, the approach would be zero and the water removed would depend exclusively on the lean TEG concentration. Our goal, however, is to design a glycol system which is economically viable. We have several options. We can select a contactor with several contacts and a low circulation rate or one with few contacts and a high circulation rate. In fact, there exist an infinite number of choices to meet our outlet dewpoint specification.

Most designs use a circulation rate of 15-40 liters TEG/kg H20 absorbed [2-5 US gaVlb H20 absorbed]. This is near the economic optimum. Higher circulation rates result in a larger regeneration system and higher energy consumption. Lower rates require a taller contactor and may result in poor traylpacking hydraulics.

The relationship between circulation rate and the number of contacts can be quantified by use of the Kremser-Brown relationship, a shortcut absorber calculation.

Where: E, =

Y N + I

Y, =

Yo =

A = L = V = K = N =

efficiency of absorption mols of water in entering (wet) gas per mol of gas entering mols of water in leaving (dry) gas per mol of gas entering mols of water in equilibrium with the incoming lean glycol per mol of gas entering absorption factor, A= LNK glycol circulation rate, moldunit time gas flow rate, moldunit time vapor liquid equilibrium constant (Y/X) for water in a water-gas-TEG system number of theoretical constants in absorber

In TEG systems, the mol fraction, y, may be used instead of the regular absorption parameter, Y, due to the concentrations involved.

@John M. Campbell & Compmy


6.1 1

GLYCOL DEHYDRATION

@John M. Campbell & Cwipony



f"

r'

This means that the left hand side of Equation 6.1 becomes

YN+1 - Y1

YN+i - YO

where "y" refers to the mol fraction of water rather than a rate per mol of incoming gas.

Further, since the mol fraction of water can be converted to mass of water per standard gas volume by a fixed conversion factor,

SI units: W = 761 O00 YH,O

FPS Units: W = 47 400 YH,O

Where: W = kg/106 m3 (std) [IbmhíMscfJ

Equation 6.2 may be rewritten in terms of W

Where: WN+I = water content of entering (wet) gas, kg/106 m3 (std) [IbdMMscfl W, = water content of leaving (dry) gas, kg/106 m3 (std) [lbm/MMscfl Wo = water content of gas in equilibrium with incoming lean TEG, kg/106 m3 (std)

[lbm/MMscfl

WO is a function of the lean TEG concentration.

Figure 6.6 is a plot of Equation 6.1 and is convenient for manual calculations. This uses what could be called an overall absorption factor. LO is the rate of lean TEG entering the top tray and VN+I is the gas rate entering the bottom tray. Even though the (LN) changes slightly throughout the absorber the effect of these changes are essentially canceled by changes in the "K" value, so use of L f l ~ + l K has little effect on the accuracy of the method.

The left-hand ordinate of Figure 6.6 is called absorption eflciency of absorption, E,. It is the actual amount of water removed, divided by the maximum amount theoretically removable. The values of N encompass the range of theoretical stages usually employed in TEG contactors.

Calculation of Lean TEG Rate for a Given Absorption Efficiency and N -

1. Calculate yo (or Wo) 2. Determine absorption efficiency

3. Use Equation 6.2 or Figure 6.6 to find absorption factor A for a given value of N 4. Knowing V N + ~ and K, solve A for LO, the lean TEG circulation rate

Calculation of N for a Given Lean TEG Rate and Absorption Efficiency -

1. Calculate yo (or WO) 2. Determine absorption efficiency 3. Calculate absorption factor A 4. Determine N from Equation 6.2 or Figure 6.6

It is usual to repeat the calculation to obtain three lean glycols rate/absorber contact values that satisfy the required absorption efficiency. The final choice is economic. This usually involves selection of standard designs.



6.13

GLYCOL DEHYDRATION

0.99999 I

Figure 6.6 Efficiency of Absorption, E, vs. Absorption Factor, A

This calculation should be made at the lowest pressure and highest temperature anticipated for the entering wet gas, to obtain the maximum water loading. Unfortunately, the tendency has been to r\ choose a design temperature lower than that actually incurred.

The overall tray efficiency in a well-designed TEG unit will vary from 25-40%. It is recommended that 25% be used for most applications. This provides an affordable safety factor to help compensate for the inherent errors in the design specifications.

Equilibrium Relationships Various studies have been made of the equilibrium behavior of water in the TEG-water sys-

All provide rather consistent data. The use of an activity coefficient (y) is a convenient and reliable method for calculating K. Using this relationship

K = (Yw)(Y) = (B)(W)(y) (6.3)

Where: K = equilibrium constant for water in a TEG-water system yw = mol fr water in the gas at saturation over 100% liquid water (from regular water

y = activity coefficient for water in the TEG-water system as found from Figure 6.7 W = water content on a mass per volume basis, at saturation, as found from a regular

B = 1.31(E-06) when W = kg/106 std m3

content correlation)

water content correlation

2.1 1 (E-05) when W = IbmMMscf

? Notice that y and thus K vary with TEG concentration and temperature, which in turn varies throughout the absorber. An average K at average concentration cannot be found until the circulation rate is fixed. So, a simple trial-and-error calculation is involved. One can assume the inlet lean TEG




Figure 6.7

concentration as a first try. As stated previously, for most dehydration applications the increase in K from the lean to rich TEG is roughly proportional to the increase in L N , so the absorption factor (A =

LV/K) remains relatively unchanged.

Activity Coefficient for H20 Concentration at Various Temperatures

In the absorption efficiency term,

Yo = Kxo and wo = (W)(Y)(Xo)

Where: xo = mol fr water in the lean TEG entering the absorber

This may be calculated from Xgl, the weight percent TEG in the lean solution entering the absorber. This must be not less than the minimum value required from Figure 6.4.

(100 - Xg1)/18 xo =

[(loo -Xg,)/18] + (Xg1/150)

Equation 6.5 is shown graphically in Figure 6.8.

The above procedure using the Kremser-Brown approach is as accurate as a rigorous tray-to- tray balance around the absorber. This is true because 'A' is essentially constant throughout the tower.

The use of other equilibrium K values will have some effect on contactor design. The required lean glycol concentrations may differ but the difference is normally less than the random error in process specifications and is statistically insignificant.

When using the Kremser-Brown method, the terms V and L must be expressed in molar units. This requires calculation of the MW of the TEG solution. The molecular weight of a TEG-water solution may be calculated as follows:

I" = 1 8 ~ 0 + 1 5 0 ( 1 - ~ o ) (6.6) @John M. CampbeU & Company


6.15

GLYCOL DEHYDRATION

0.50

0.40

O X 0.30

I C O

o

0 .- c

2 LL 0.20 - 2

0.1 o

0.00 90 91 92 93 94 95 96 97 99 1 O0

TEG Concentration (Xgl), wt%

Figure 6.8 Equation 6.5 Mol Fraction H20 vs. TEG Concentration

The actual circulation rate used in the unit will be different from this because operating conditions always differ to some degree from those specified in the design. In addition, TEG circulation pumps come in discrete capacities. It is sound practice to choose other components based on the capacity of the circulation pump.

Example 6.4: Calculate the circulation rate of 98.7 wt% lean TEG needed to dry lo6 std m3/d [34.9 MMscfd] of gas at 7.0 MPa [ 1000 psia] and 40°C in a six tray absorber (1.5 theor. tray) to achieve an exit gas water content of 117 kg/106 std m3 [7 lb/MMscfJ. The inlet water content is 1100 kg/106 std m3 (saturated gas) [67 IbíMMscfl.

1. From Equation 6.5, q = 0.099 2. From Figure 6.7, y = 0.66 3. W is the water content of saturated gas at 7.0 MPa and 40°C or 1100 kg/106 std m3 in

4. From Equation 6.4, WO = (1100)(0.66)(0.099) = 71.9 kg/106 std m3 5. The left-hand side of Figure 6.6 is: (1 100 - 117)/(1100 - 71.9) = 983/1048 = 0.956 6. From Figure 6.6, for N = 1.5, A = 7.3

(In order to solve for L one must recognize that it is different at each point in the absorption tower. The conservative approach is to assume the gas volume is constant and solve for from A. This will yield a circulation rate sufficiently close to a more rigorous calculation.)

this case.



r'

á"

f-



7- LO = (A)(K)(vN+i) From Equation 6.3, K = (1.33 x 10-6)(1100)(0.66) = 0.000 967 V = 1739 kmolh SO, LO = (7.3)(9.67 x 104)(1739) = 12.3 km0Vhr

8. MW lean glycol = (0.099)( 18) + (0.901)( 150) = 137 9. kg TEGh = (12.3)(137) = 1681

1 O. Density of TEG is 1.1 2 g/cm3 = 1.12 kg/liter Circulation rate is 168M.12 = 1500 literh In one hour (1 1 O0 - 1 17)/24 or 41 .O kg H20 is absorbed. Circulation rate is 1500/41 = 36.6 liter/kg H20 absorbed.

:n traditional English units the calculation follows the same format. 1.,2. The same

3. W = 67 lb/MMscf 4. WO = (0.66)(67)(0.099) = 4.38 lb/MMscf 5 . (67 - 7)/(67 - 4.38) = 0.958 6. A = 7.5 7. K = (2.11 x 10-5)(67)(0.66) = 0.000931

V = (34.9)(110) = 3839 lb-molh Lo = AKVN+I = (7.5)(9.31 x 104)(3839) = 26.8 l b - m o b

8. MW = 137 9. lb TEG/hr = (26.8)(137) = 3672

10. Density of TEG is about 9.3 1bLJ.S. gal Circulation rate is 3672/9.33 = 394 U.S. gal/hr In one hour a total of 92 lb of water is absorbed. Circulation rate is 394/92 = 4.3 U.S. gal/lb water absorbed

For the previous example, would we buy a 1.5 theoretical tray absorber? Probably not! The circulation rate calculated is toward the high end of the economic range. A 7 or 8 actual tray (1.75-2.0 theoretical trays) might well be specified to provide valuable flexibility and inexpensive "insurance."

The shortcut calculation method presented in this section is still somewhat tedious to do by hand. Appendix 6B presents graphical solutions to this shortcut method taking into account the water removal, circulation ratio, lean TEG concentration and number of theoretical stages in the contactor.

Absorber Design

The design of the absorber (contactor) is based on two parameters:

1. gas rate which determines the contactor diameter, and

2. number of contacts, which determine the contactor height.

In some cases the contactor may also contain an integral scrubber designed to remove entrained drop- lets and solids from the gas prior to entering the absorption section. In addition, an internal gas-glycol exchanger is included at the top of the contactor in some designs.

OJohn M. Cimpbeli & Company

Technical Asslstance Service for the Design, Operation, and Maintenance of Gas Plants

6.17

GLYCOL DEHYDRATION

-I The contactor diameter depends almost exclusively on the gas rate and is virtually independent

of the glycol rate. This is due to the low liquid loadings employed in glycol contactors. Two types of contactor internals are used in glycol systems - trays (usually bubble cap) and packing (usually structured).

Bubblecap trays have been historically used in TEG systems. This was certainly the case until Since then, many contactors have been installed with structured packing. This is the mid-1980’s.

primarily due to the superior gas handling capacity of structured packing relative to trays.

When trayed contactors are used, bubblecaps are preferred over other types of trays (valve and sieve) due to their higher turndown ratio and generally better efficiency at low liquid rates. Likewise, when a packed contactor is employed, structured packing is preferred over random packing because of its higher capacity, better turndown and superior performance at low liquid rates. Random packing is sometimes used in small diameter contactors (less than 0.6 m [24 in]) for convenience.

Figure 6.9 shows a bubblecap tray, section of structured packing and random packings.

Raschig Rings

Pail Rings

Bubble Cap

Ceramic Bed Saddles Sbuctured Packing

Figure 6.9 Bubblecap, Random and Structured Packing




r" Contactor Diameter As stated earlier, the diameter of a glycol contactor is determined, almost exclusively by the

gas rate. If a contactor is operating near flood, changes in the glycol rate can have a noticeable effect on the glycol carryover, but for design purposes the liquid rate is typically not a factor in absorber sizing.

The calculation of diameter can proceed two ways. The first method employs the Souders Brown equation, frequently used to size separators.

Where: v = allowable gas velocity, m / s [ft/sec]

0.055 ; 0.09-0.105 O .3 0-0.34

K, = sizing parameter bubblecaps structured packings

pg = gas density, kg/m3 [lbm/ft3] pL = liquid density, kg/m3 [lbm/ft3] for TEG systems, pL = 1120 kg/m3 [69.9 lbm/ft3]

The calculation of the diameter follows

Where: d = contactor diameter, m [ft] qa = actual gas flowrate, m3/s [ft3/sec]

The actual gas flowrate can be calculated from either the mass flowrate or the standard volumetric rate.

m

PP 9 a = -

Where: m = mass flowrate, kg/s [lbmísec] pg = gas density, kg/m3 [lbm/ft3]

Where: qstd = gas flowrate in standard volumes, m3 (std)/d [scflday] Pstd = standard pressure, kPa [psia] Pa = actual flowing pressure, kF'a [psia] T, = actual flowing temperature, K ["R]

Tsid = standard temperature, K ["R] z = gas compressibility factor at flowing conditions


(6.10)


6.1 9

GLYCOL DEHYDRATION

A second sizing equation uses an F, value which is related to the kinetic energy of the gas (pv2). This is the most popular method for sizing packed towers.

(6.1 1)

Where: v = allowable gas superficial velocity, m / s [fiísec] F, = sizing parameter, P:.’ [1bm0~’/sec-ft0~’]

The sizing parameter, F,, depends on the type of packing but for most structured packings, F, = 3.0 P:’ [2.5 lbm0~5/sec-ft0~5]

It should be noted that for contactors containing structured packing the gas handling capacity may be limited by the mist extractor, not the packing. This is particularly true when the glycol viscosity exceeds 15-20 cp, the viscosity of TEG at temperatures of 30°C [86”F].

Example 6.5: A glycol contactor is to be designed to handle 1 x lo6 m3 (std)/d [35.3 MMscfd] of gas at 40°C and 70 bar [lo15 psia]. The gas compressibility factor is 0.85 and the MW = 19.0. Size the contactor for both bubblecaps and structured packing.

SI Units: 1) Calculate the gas density

2) Calculate allowable velocity, v II Bubblecaps, v = Ks [ pLipg = 0.055 [ l2:i 6o 7’ = 0.23 m / s

3) Calculate the actual volumetric rate, qa

qa = = ‘Ooo 86400 ‘Oo ( 0 ) ( % ) ( 0 . 8 5 ) 7000 = 0.154 m3/s

4) Calculate d

11 5 ) for structured packing

3 .O -- - 0.391nís

V. I

y (0.39) (3.14)



?-

f-


:xample 6.5 (Cont'd.):

FPS Units: 1) Calculate the gas density

= 3.75 1bdft3 1015) (19 pg = (0.8;) (10.73) I564)

2) Calculate allowable velocity, v

3) Calculate the actual volumetric rate, qa

35*3 'O6 ( -- 1407 )[ :;:)(0.85) = 5.46ft3/sec 86400 1015 qa'

4) Calculate d

d = (0.756) (3.14)

5 ) for structured packing

In an actual design it is sound engineering practice to size the contactor for a gas rate 20-30% higher than the expected rate. This contingency provides contactor capacity for changes in flow rate and pressure and for pessimistic reservoir engineers.

Contactor Height

The contactor height is determined by the number of equilibrium contacts required and efficiency of the mass transfer. For trayed contactors the conversion from equilibrium stages to actual trays is accomplished by using a tray efficiency. The tray efficiency is measure of the approach to equilibrium and can be calculated from either vapor phase or liquid phase compositions.

The overall tray efficiency is defined as follows:

No. of equilibrium stages No. of actual trays Eoverall = (6.12)

For glycol contactors Eoverall typically ranges from 25-30%. (This is equivalent to a Murphree plate Efficiency, EmG, of approximately 45-50%). For most engineering calculations on overall tray efficiency of 25% will yield satisfactory results.



6.21

GLYCOL DEHYDRATION

A minimum spacing of 24 inches is recommended. It is essential that a stable foam not fill the gas space between trays to prevent excess glycol loss. This spacing also allows a suitable liquid level in the downcomers.

Tray hydraulics design is critical because of the low circulation rate. Liquid can by-pass caps or valves in some areas of the tray, ineffective gas-liquid contact can occur with low gas rates, and liquid levels can be unstable. In situations like this, absorber performance can vary markedly with gas and liquid rate. A higher than calculated liquid rate may be necessary to provide the tray efficiency required. A minimum downcomer area of 11% of the cross-sectional area of the tower is required to ensure adequate liquid distribution across the tray.

For packed towers, equilibrium stages are converted to packing heights using an HTU (Height of a Transfer Unit) or HETP (Height Equivalent and a Theoretical Stage).

HETP and HTU are similar concepts and depend primarily on the gas and liquid properties, gas and liquid rates as well as the surface characteristics and density of the packing.

The mass transfer in a packed tower is continuous, and does not take place in discrete steps as in trayed towers. It is for this reason that many packing manufacturers prefer to work in HTUs. The HTü is multiplied by the Number of Transfer Units (NTUs) to arrive at the packing height. For overall mass transfer with resistance on the gas side. The from the following equation.

Y b

NTU = j"y_ Y - Y *

Yt

Where: NTU = number of transfer units

number of transfer units may be calculated

(6.13)

7 Yb = concentration of water in the bottom of the contactor yt = concentration of water at the top of the contactor y* = equilibrium water concentration

A similar concept is employed in heat transfer calculations. If we assume the operating and equilibrium lines to be straight,

Yb-Yt In - NTU = AYb - AYt (") (6.14)

Where: AYb = yb - y: AYt = Yt - y;

A typical HTU for structured packed towers depends on gas density and packing type but for packing with a specific area of 250 m2/m3 [76 ñ2/R3] (Sulzer Mellapak 250Y, Koch Flexipak #2, Glitsch Gempak 3A, Montzpak B1/250) an HTU of 0.8 m [2.6 Et] gives good results for preliminary calculations.

Despite the technical superiority of the HTU/NTU approach many companies still use the HETPNTS method.

NTS stands for the Number of Theoretical Stages and is identical to the "N" values in Equa- tions 6.1 and 6.2. Even though mass transfer in packed towers is continuous the HETPNTS method assumes the mass transfer takes place in discrete stages and the HETP is the height of packing equivalent to one of these equilibrium stages.

/cI\




Increasing Pressure

Increasing Gas Flowrate

Increasing Liquid Flowrate

r"

Increases HETP

Increases HETP

Decreases HETP

For TEG contactors containing structured packing with a specific area of 250 m2/m3 [76 R2/ft3], typical HETP values range from about 1.6-2.0 m 15.3-6.5 R]. The HETP of structured packing varies with contactor operating parameters as follows:

1) Increasing Specific Area of Packing I Decreases HETP I U

~~

Increases in HETP mean poorer mass transfer Decreases in HETP mean better mass transfer

Example 6.6: Calculate the height of packing required to provide 2 equilibrium stages in a TEG contactor.

Assume HETP = 1.75 m [5.74 R]

Packing height = (2)(1.75) = 3.5 m = (2)(5.74) = 11.5 ft

It is customary to add about 10% additional packing (usually 1-2 1 yers) to account for operating contingencies and allow for distribution of the gas and glycol at the top and bottom of the packings.

f- Distributor Design In packed contactors, the design of the liquid distributor is critical. The liquid distributor

ensures that the incoming lean glycol is uniformly distributed across the packing. Several different distributor designs are used but they generally all consist of a main header box and a series of rectan- gular radial flow channels, as shown below.

8 John M. Campbell & Company


6.23

GLYCOL DEHYDRATION

7 The glycol flows out of the radial distributors through weirs along the side of the box or

1. Distributor must be level 2. Weir openings or drip tubes must be resistant to plugging 3. Drip point density should be 80-100/m2 [8-10/ft2] 4. Minimum glycol circulation rate should be about 1 (m3/h)/m2] [0.4 US gpm/ft2]

5 . Area available for gas flow should be a minimum of 40-50% of cross sectional area 6. Drip points should be less than 10 mm (1/2 in) from the top of the packing to avoid

The distance between the top of the distributor and the mist eliminator should be a minimum of

through drip tubes. Although designs differ there are several critical design parameters.

splashing and droplet re-entrainment

0.6 m [2 ft]. A drawing of a typical structured packed tower is shown in Figure 6.10.

Gas Outlet

Demister Mat

Manhole Glycol Inlet

Liquid Distributor

Structured Packing

RiserCap -\

GasRiser -, Chimney Tray ------\

(used if integral scrubber installed in base of contactor)

Drain Pipe

Gas Inlet L

r

Manhole

Vortex Breaker

Glycol Outlet Note. Column inside diameter = D

Figure 6.1 O Example Structured Packed Absorber

\


0.15 D (minimum 0.15 m 16 in] 'O. 10 m (4 in]

0.60 m [2 it]

0.2-0.4 m [8-16 in] depends on distributor desgn

Pad& heighi 2.5-4.0 m [8-13ft] typical

0.4 m [15 in] minimum

h, (0.25 m (10 in] typid)

0.5 D or 1 m [3 it] minimum


TEG REGENERATION

rc'

TEG REGENERATION The required lean TEG concentration is produced in the regenerator. The regenerator consists

of a reboiler, still column and in some cases a gas stripping column. The lean TEG concentration is controlled by adjustment of reboiler temperature, pressure and the possible use of a stripping gas. So long as no stripping gas is used, the concentration of the lean TEG leaving the reboiler is independent of the rich TEG entering.

The concentration of rich TEG leaving the absorber is found by a water material balance around that absorber. By definition

Because regeneration takes place near atmospheric pressure, under essentially ideal gas conditions, the calculation is routine. Figure 6.1 1 (6.8) has been calculated to predict regenerator performance for various stripping gas rates at a pressure of 1 atm [14.7 psia] and 204°C [400"F].

100.0

99.8

99.6

99.4

99.2

oe o

o 98.8

I-

9 98.6

99.0

c

w C

J

98.4

98.2

98.0

97.8

97.6

Stripping Gas Rate, (std) rn3/m3 TEG

2 4 6 8 10 12 Stripping Gas Rate, scf/gai

1 Figure 6.11 Effect of Stripping Gas on TEG Concentration for Regenerators with Stripping Columns


Technical Assistance Service for the üeslgn, Operation, and Maintenance of Gas Plants

6.25

GLYCOL DEHYDRATION

Steam Hot Oil

Electric

( 100) mass lean TEG

mass leanTEG+ mass water absorbed+ mass water in lean TEG wt % richTEG =

24 kW/m2 [7600 Btu/hr-ft21 24 kW/m2 [7600 Btu/hr-ft2]

12.5 kW/m2 i4000 Btuihr-ft21 /L53

The mass quantities in this equation may be found per unit of time or per unit of gas or glycol flow. In any case, the values used depend on circulation rate. As we have seen in previous sections, TEG rate depends on dewpoint requirements, lean TEG concentration, number of absorber contacts and economics.

The rich TEG concentration may be calculated from Equation 6.15.

(p) (Lean TEG) RichTEG = P + (1/L)

SI Units Where: p = lean TEG density 1.12 kg/liter

L = lean TEG rate/mass H20 removed literdkg H20 Rich TEG = wt% TEG in rich TEG solution Lean TEG = wt% TEG in lean TEG solution

(6.15)

FPS Units 9.3 1bKJ.S. gal

U.S. gal/lb H 2 0

The crosshatched region in the Figure 6.11 represents the effect of stripping gas when introduced directly into the reboiler through a sparge tube. The lean TEG concentration achievable depends on the rich TEG concentration. The upper line defining the crosshatched region is for 98 wt% rich TEG while the lower line is for 95 wt%. Linear interpolation may be used to estimate the stripping gas rate for a rich TEG concentration between 95 and 98 wt%.

The concentration of the rich TEG also affects the stripping gas requirements when a packed stripping section below the reboiler is used; however, the concentration effects are small and a single line was used in Figure 6.1 1 to describe the relationship for both N = 1 and N = 2. ('N' refers to the number of theoretical stages in the stripping column).

Stripping gas rates seldom exceed 75 m3 (std)/m3 TEG [lo scf/gal] unless lean TEG concentrations in excess of 99.99 wt% are required. If these concentrations are required, an alternate design such as DRIZO@ should be considered.

Heat input to the regenerator is provided in the reboiler. The heat source is usually direct fired with the fire tubes immersed in a glycol bath. Other heat sources include hot oil (or other heat transfer fluid) steam, or electric resistance heating.

In any event, the temperature of the glycol in the reboiler should not exceed 204°C [400°F] due to the degradation of TEG at higher temperatures. However, in order to maintain the glycol bath at a temperature at or slightly below 204°C [400"F] it is necessary to maintain the heat transfer surface at a temperature above this value. This can lead to some degradation of the glycol in contact with the heat transfer surface. For this reason the following maximum flux rates should not be exceeded. These flux rates, in turn, will set the heat transfer area.

I Direct Fired I 19 kW/m2 [6000 Btu/hr-ft?] I

@John M. Campbell & Comppny


TEG REGENERATION

6"

r'

f"

The heat source (burner, hot oil or steam rate, electrical demand) should be a minimum of 30% higher than the above values. This is necessary for start-up conditions, or periods when the glycol is excessively loaded with water.

Most glycol reboilers maintain the bath temperature at 204°C [400°F]. Lower temperatures may reduce degradation but result in lower lean TEG concentrations which, in turn, necessitate higher circulation rates or higher stripping gas rates.

Pressure effects on regenerator operation are not often fully appreciated. Lean TEG concentrations are reduced by backpressure on the regenerator. Typical backpressure for venting to a LP flare system or condenser system may be as much as 3.5-7 kPa [0.5 to 1 psi]. At 204°C [400"F] and 7 kPa [1 psi] backpressure (108 kPa [15.7 psia] total pressure) will reduce the lean TEG concentration by 0.1 wt%. Conversely, the lean TEG concentration may be increased at high altitudes due to the lower atmospheric pressure.

The reboiler duty depends on the TEG circulation ratio (liters TEG/H20 [gal TEG/lbm H20]), the efficiency of the rich-lean TEG exchanger, the reflux ratio, stripping gas rates and effectiveness of the insulation. Heat balances indicate a required reboiler duty of 250-300 kJ/liter [900-1075 Btu/US gal]. The reboiler duty should actually be sized to deliver 120-125% of the expected duty to provide for start-up, insulation losses, etc. A design value of 350 kJ/liter [1250 Btu/US gal] will typically provide sufficient heat input flexibility to meet any expected operating condition. The reboiler duty should always be sized based on circulation pump capacity, not the expected circulation rates.

The still column is the "fi-actionator" portion of the regenerator. The column may be packed or trayed. Packed columns are more common and the packing is typically a random packing such as stainless steel slotted rings. Packing sizes range from 16 mm [5/8 in] to 51 mm [2 in] depending on the still column size with the larger packing used in the larger diameter columns. Structured packing has also been used and in very large units (diameters > 3 ft), trays have been installed.

The still column is sized based on standard packed tower sizing correlations. Since the vapor loading is often tied to the glycol circulation rate many correlation have been developed which estimate the still column diameter as a function of TEG circulation rate. One such correlation is shown below and is based on 25 mm [ 1 in] slotted ring packing.

d = (A) (m)0.5 (6.16)

SI Units FPS Units u US gaYmin

The reflux ratio employed in TEG systems is very small. L/D valves typically range from 0.1 to 0.2 moldmol. This is equivalent to condensing 10-20% of the total overhead vapor stream. The reflux rate should be the minimum required to maintain the still overhead temperature at the boiling point of water for the partial pressure of water at the top of the regenerator. In other words, when stripping gas is used, the water partial pressure will be less than 1 atm. In fact, this is the principle which results in lower water concentrations in the lean TEG. Consequently, when stripping gas is used, the partial pressure of H20 will be less than 1 atm, hence the boiling point will be lower as well.

Where: m d = = glycol diameter circulation of packed rate tower A = empirical constant (1 in. pall rings)

Figure 6.12 shows the recommended still column overhead temperature as a function of the TEG circulation ratio and stripping gas rate for a still column operating at 1 atm.



6.27

GLYCOL DEHYDRATION

Circulation Ratio, gal TEG/lbm H,O 1 2 3 4 5 7

1 O0

95

90

9 85 2- 2 80 3

a, (1

- E l- 75

70

65

60

Figure 6.1 2 Recommended Still Column Top Temperatures

Reflux is normally supplied by using the rich glycol stream circulating through a condensing coil inserted in the top of the still column. Cooling water or an air finned cooler could also be used. However, using the rich glycol stream is the preferred method for cost, simplicity and energy efficiency. Heat transfer coefficients in the reflux coil typically range from 110-230 w/m2K [20-40 Btu/hr-ft2-"F].

Glycol-Glycol (lean-Rich) Heat Exchanger This is the basic heat exchanger. Its efficiency has a direct effect on reboiler heat load. The

rich glycol from the absorber (Pt.1) enters at a temperature 513°C [9-18"F] warmer than the inlet gas due to the reflux condenser duty. The lean glycol from the regenerator (R. 3) enters usually at about 204°C [400"F]. The exchanger is designed so the temperature of the lean glycol at Pt. 4 should not be greater than 60-65°C [140- 149"FI.

In most cases a 1520°C [27-36"F] approach in the heat exchanger is desirable. If it is too high, the reboiler and glycol cooler duties will increase. In the lean-rich exchanger

4

(6.17)

-l



FILTERS

The easiest way to make this enthalpy calculation is to look up the average heat capacity of the glycol in Appendix 6A and multiply it by AT across the heat exchanger to find Ah for the lean mixture.

Double pipe type, with bare or finned tubes, or plate type heat exchangers are frequently used. Plate exchangers are preferred, especially for offshore or large units, because they are more compact, lighter and cheaper. They are, however, susceptible to fouling and plugging and it is imperative the glycol is clean and filtered.

r"

Some small dehydration units use a coil in the surge drum to exchange heat between the rich and lean glycol stream. Not only are these coils limited in surface area but the overall 'U' values tend to be low. As a result they may not be adequate for applications where the reboiler heat is limited. They also may not cool the lean glycol adequately to meet the pump temperature limit. They should only be considered for small pack units having a reboiler duty less than 150 kW [500 MBtu/hr].

Lean Glycol Cooler A final glycol cooler is required so that the lean glycol entering the top of the contactor is

cooled to within 510°C [9-18"F] of the gas temperature entering the top tray. The lean glycol may be cooled or with a lean glycol-gas heat exchanger. Gas-glycol exchangers are cheaper and more compact but the lean TEG temperature can increase to unacceptable levels at low gas rates. Although air-cooled glycol coolers risk under-cooling the lean glycol, they are recommended since control of the glycol temperature is then independent of the gas flow.

If a gas-glycol exchanger is used we do not recommend an integral coil in the top of the absorber column due to less efficient heat transfer and problems of inspection and maintenance.

F I LTE RS

I"

Good filtration is critical. The full-flow type is preferred. I recommend two filters in parallel, with no by-pass lines, so that full filtration is assured.

A cloth fabric element that is capable of reducing solids to about 100 ppm by weight is preferred. Paper and fiberglass elements generally have proven unsatisfactory. Filter size in a properly operated glycol system should be 5-10 pm. Larger sizes 25-50 pm may be required during startup and in dirty service.

It may be impossible to judge the effectiveness of filtration by color alone. Even well filtered glycol will often be black. But, removal of most of the solids will reduce corrosion, plugging and deposits in the reboiler, and may reduce foaming losses. Good filtration is critical for satisfactory performance. It is desirable to measure the pressure differential across the filter and change the element when it reaches about 170 kPa [25 psi].

The use of a carbon purifier downstream from the filter often is recommended. This will produce essentially water-white glycol. Maintenance of this color has proven desirable because it tends to increase dehydration efficiency and minimum foaming, a major source of glycol loss.

Aromatic hydrocarbons are often present in the rich glycol entering the regenerator and will be adsorbed on the carbon filter. These components will quickly reach equilibrium loading on the carbon filter although it is likely they are eventually displaced (to some extent) by heavier hydrocarbons. Because of this high aromatic content, chang out of carbon filters requires special precautions to avoid unnecessary exposure of workers to BTEX components.



6.29

GLYCOL DEHYDRATION

T"7 In some units, the carbon filter is installed in the lean TEG stream, upstream of the contactor,

Coal-based activated carbon should be used because wood-based charcoal tends to break up in use. This carbon can be placed in a metal canister or fill a vessel. In either case, good screens are needed to prevent carbon loss into the system. Said carbon particles, much like iron sulfide, tend to promote a stable foam.

to avoid significant aromatic loading.

Glycol filters are only effective when used. In especially dirty glycol systems, filters are often bypassed to avoid frequent filter change-out. The problems with this should be obvious. If filter plugging is excessive, try larger filter size and look for source of problem (e.g., poor inlet separation, degradation, corrosion products, etc.).

SURGE DRUM The surge drum should be sized to provide the following:

a retention time not less than 20 minutes between low and normal levels, based on design circulation rate;

0 hold-up capacity between normal level and high level; a reasonable length of time between glycol additions;

0 sufficient volume to accept the glycol drained from the reboiler to allow repair or inspection of the firetube or heating coil.

It is often vented to the reboiler but a small amount of N2 or dry fuel purge gas is sometimes needed to prevent water vapor from the reboiler flowing through the vent line and being absorbed by the glycol in the surge drum. If it is not located directly below the reboiler, it should be provided with a separate purge gas supply.

/1

Provisions should also be made to facilitate: 0 make-up of the glycol inventory from glycol storage by means of a simple hand or air-

driven pump with a check valve and filter; 0 the batch addition of chemicals to the glycol surge vessel, e.g. for pH control, corrosion

inhibition, etc.

GLYCOL FLASH VESSEL The glycol flash vessel is used to remove light hydrocarbons, COZ and/or H2S, that have been

absorbed or entrained with the glycol as well as recover the spent gas from gas-glycol powered pumps. It also serves to separate any liquid hydrocarbons from the glycol to prevent them from entering the reboiler and causing fouling and foaming.

Both sulfur compounds and carbon dioxide are very soluble in water and react to some degree with the glycols. Degassing in the flash vessel upstream of the regenerator reduced their concentration and potential high temperature corrosion. Degassing is more efficient if the rich glycol is pre-heated first. Pre-heating, to about 60-70°C [140-158"F] is often done to decrease viscosity and facilitate degassing.

compatible with the allowable pumpback pressure of 15 percent of the contactor operating pressure. When gas-glycol powered pumps are used, the operating pressure of the flash vessel must be /7



GLYCOL CIRCULATION PUMPS

Thus, 70 bar [IOIO psia] contactor pressure the flash vessel must operate at a pressure lower than f- 10 bar [I45 psia].

The flash vessel should not contain liquid hydrocarbons. Unfortunately this is often not the case due to poor inlet gas separation upstream of the contactor. Therefore, it is prudent to install a hydrocarbon skim nozzle, bucket or trough and weir to collect and separate the liquid condensate from the rich glycol.

GLYCOL CIRCULATION PUMPS Glycol circulation pumps may be electric driven gas, or gas-glycol powered. The pumps

should be sized to provide a minimum of 25 percent excess capacity, and in critical service two glycol pumps shall be provided each designed for 100 percent duty.

Pumps utilizing conventional electric motor drives are normally reciprocating multiplex type. A conservative, slow piston speed (0.6 m / s [120 ft/min]) pump should be used since the lubricating properties of glycol are poor. Due to the expected turn down of glycol systems, variable speed drives are often used on larger units. They can provide the flexibility to increase the glycol circulation rate if needed to meed dewpoint requirements and reduce the glycol rate to operate more economically. Low speed centrifugal booster pumps are sometimes used when NPSHA to the reciprocating pumps is marginal.

r"

The gas-glycol powered pump (Kimray pump) utilizes the rich glycol under pressure plus sup- plemental gas from the contactor to furnish the driving energy. Since the pumping rate is proportional to the volume of the return glycol and gas, the pumping rate is controlled by adjusting this flow and no level control of the surge drum is required.

A cutaway of a Kimray pump is shown in Figure 6.13.

INLET SEPARATION Without a doubt, most operating problems with glycol dehydration systems are a direct result

of inadequate inlet gas treatment upstream of the contactor. The upstream separator should remove liquid hydrocarbons, liquid water, solids, corrosion inhibitors, etc. prior to the glycol unit. Campbell says "You cannot afford the dehydration if you cannot afford to place an effective separator on the gas inlet." A glycol unit is a closed system. Nonvolatile contaminants remain in the glycol and must be removed by filtration or blowdown.

Some of the more deleterious contaminants are salt and high boiling point hydrocarbons. These remain in the glycol. Salt precipitates in the reboiler, still column and rich-lean exchangers. This can cause plugging increasing pressure drop and decreasing flowrates. In the reboiler, salt can coat the firetube causing hot spots and eventual firetube failure.

Heavy hydrocarbons can "coke" in the reboiler causing hot spots on the firetube and plugging in the still and stripping columns. In addition, these burnt hydrocarbons cause the glycol to become black and promote foaming in the contactor, and frequent filter change outs. If sulfur compounds are present in the gas, these can combine with the heavy hydrocarbons to form a corrosive "sludgetf which accumulates in the system.



6.31

GLYCOL DEHYDRATION

OPERATION: The Kimray glycol pump is double acting, powered by Wet Glycol and a small quantity of gas at absorber pressure. Wet Glycol from the absorber flows through port #4 and is throttled through the SPEED CONTROL VALVE to the left end of the Pump Piston Assembly, moving this assembly from left to right. Dry Glycol is being pumped from the left cylinder to the absorber while the right cylinder is being filled with Dry Gly- col from the reboiler. At the same time Wet Glycol is discharging from the right end of the Pump Piston Assembly to a low pressure or atmospheric system. As the Pump Piston Assembly nears the end of its stroke, the POSITION RING on the PISTON ROD contacts the right end of the ACTUATOR. Further movement to the right moves the ACTUATOR and PUMP "D" SLIDE to uncover port #1 and communi- cate ports #2 and #3. This exhausts Wet Glycol from

the left end of the PILOT PISTON through ports #2 and #3 to the low pressure Wet Glycol system. AI the same time port #1 (which was communicated with port #3) admits Wet Glycol to the right end of the PILOT PISTON. This causes the PILOT PISTON and PILOT "D" SLIDE to be driven from right to left. In its new position the PILOT "D" SLIDE uncovers port #5 and communicates ports #4 and #6. This exhausts Wet Glycol from the left end of the Pump Pis- ton Assembly through ports #4 and #6 to the low pressure Wet Glycol system. Port #5 (which was communicated with port #6) now admits Wet Glycol through the right hand SPEED CONTROL VALVE to the right end of the Pump Piston Assembly.

The Pump Piston Assembly now starts the stroke from right to left. It follows the above procedure with reversed directions of flow.

Figure 6.13 Cutaway and Operation of a Kimray Glycol Pump



OPERATING PROBLEMS

Certainly, a properly sized impingent separator is the first step in preventing unwanted carryover. Many units also utilize coalescing filter separators or superheaters downstream of the primary separator. Frequently, a heat exchanger designed to increase the gas temperature by 5°C [9'F] is installed immediately upstream of the contactor. This serves to vaporize any entrained particles in the gas.

r'

Many designs utilize an integral scrubber installed in the base of the contactor. While this may provide some removal of entrained particles, these "scrubbers" are typically under sized, particularly when the contactor is packed with structured packing. These integral scrubbers should never be used as a primary separator - only for secondary scrubbing.

OPERATING PROBLEMS The glycol unit operation should be essentially trouble-free. It seldom is. Many of the prob-

lems stem from inadequate design and/or operational faults. The basic simplicity of the unit and the availability of "standard" units tends to obscure the need for attention to mechanical design details. The glycols are very reactive chemically and need to be protected from contamination.

r"

One common symptom of many problems is excess glycol loss. This loss is due to one, or a combination, of the following:

1. Foaming

2. Degradation

3. Salt plugging the regenerator still column 4. Inadequate mist extraction

5 . Inadequate absorber design for flow conditions

6. Loss of glycol from pinholes in a gas-glycol coil in the top of the absorber or in a chimney tray above a separator section in the bottom of the absorber

7. Spillage of glycol or pump leakage

8. Lean glycol to absorber is too hot

9. Inadequate reflux (temperature too high at top of still column)

Glycol likes to foam. It will foam whenever allowed to. Ordinary foaming may not be critical if the unit is carefully designed. Any foam tends to be more stable when aromatics and/or sulfur compounds are present. Metallic sulfides and sulfites, and degradation products, all contribute to the problem.

Foams are only broken using surface and time, or chemicals. Tray spacing must be large enough so that foam cannot fill up the space between trays and form a continuous liquid phase. A mist extractor does not break foam effectively. Once foam fills the absorber, there is a continuous liquid phase for glycol to go out overhead.

Use of an antifoam agent can reduce the problem. There are many antifoam agents available. One that works in one unit may not work equally well in another. Some trial-and-error testing of an antifoam agent, and concentration of that agent, is often necessary.

Avoid adding too much antifoam agent. If too much is added it may accelerate foaming. Set up a careful control policy so operators keep unit concentration within the limits specified.

0 J o b M. Campbell & Company


6.33

GLYCOL DEHYDRATiON

-I Degradation is a natural occurrence and is accelerated in the presence of sulfur compounds.

The answer is effective filtration. Degradation products contribute to foaming but they also are major sources of corrosion problems.

Salt is a continuing problem. Good separation ahead of the absorber is mandatory. Any salt arriving at the regenerator deposits either in the still column or in the reboiler. It is common for packed still columns to plug up to the point glycol is lost overhead. If this does not occur, salt can plug the reboiler and cause failure. Not providing good separation is inexcusable,

The water vapor in gas is relatively fresh but is slightly saline. NaCl is soluble in TEG to some degree. At 50°C about 3.3 kg will dissolve in 100 kg of TEG. So, some salt is always present. The soluble salt hydrolyzes to HCI and lowers the pH of the glycol.

Corrosion-Erosion Glycols are very reactive with sulfur compounds. The resultant materials tend to polymerize

and form "gunk" which is very corrosive. Also, the glycol pH becomes lower. Corrosion inhibitors alone cannot solve the problem satisfactorily. The real solution is good mechanical design and good filtration supplemented by a corrosion inhibitor.

Good design involves factors like control of fluid velocities, long radius ells and a host of little details that too seldom are done properly. In many cases, good mechanical design will eliminate the need for expensive alloy steels.

If feasible to do so, the glycol pH should be maintained above 6.0. Some become so preoccu- pied keeping it at 7.0 or above that they add copious amounts of caustic, sodium carbonate and the like to the unit. The result is seldom satisfactory. Adjustment of pH is proper but the cure can be worse than the disease if it is overdone.

/7

Corrosion inhibitors which plate out on metal surfaces and form a film can be effective in minimizing corrosion. Judicious use of stainless steel can also be effective in m i d z i n g corrosion. This is particularly true for dehydration of high C02 gases. When the COZ partial pressure is high, the rich TEG pH can be less than 5. It is often more effective to use stainless steel on the "rich" side of the unit than to try to combat corrosion with inhibitors.

In a corrosive environment, the total elimination of corrosion is an unrealistic goal. The proper goal is reducing it to economically tolerable levels.

Figure 6.14 shows the solubility of H2S in TEG.(6,9) This is true absorption that takes place in the absorber. It lowers pH and provides a mechanism for reactions. Figure 6.15 shows solubility of CO2 in a 96.5 wt% TEG solution.@.'') Solubility of CO2 in pure TEG is approximately 20% higher. Reference 6.1 1 provides additional data on the solubilities of H2S and C02, as well as C1, C2 and C3 in TEG.

AROMATIC ABSORPTION The affinity for aromatics by TEG has long been recognized. The UDEX process was used for

many years in refineries and chemical plants to extract aromatic hydrocarbons from paraffins with TEG.

In gas dehydration service, TEG will absorb limited quantities of aromatic hydrocarbons (ben- 7 zene, toluene, ethylbenzene and xylene) from the gas. These components are often abbreviated as



if-

f-

AROMATIC ABSORPTION

Temperature, “C O 10 20 30 40 50 60 70 80 90 100

20 40 60 80 100 120 140 160 180 200 220 Temperature, O F

Figure 6.14 Solubility of Hydrogen Sulfide in TEG at Various Partial Pressures

Pressure, kPa 1000 2000 3000 4000 5000 6000 7000 8000

‘0 100 200 300 400 500 600 700 800 900 1000 1100 1200 Pressure, psia

Figure 6.15 COP Solubility in TEG 3.5 wt% H20



6.35

GLYCOL DEHYDRATION

7 BTEX. QuantiSring the absorption levels has become more important in recent years due to increased restrictions on aromatic emissions from glycol units. To date, little published equilibrium data is available.f6.12,6.13)

Based on data from Reference 6.13, predicted absorption levels for BTEX components vary from 5-10% for benzene to 20-30% for ethylbenzene and xylene. Absorption is favored at higher pressures and lower temperatures, increased TEG concentration and circulation rate.

The bulk of absorbed aromatics will be vented with the water vapor at the top of the regenerator. In some cases these aromatics can be condensed and recovered, however the effectiveness of condensation declines rapidly with increasing stripping gas rates.

In these cases, many operators use a partial condenser to condense the water and pipe the volatile hydrocarbons to an incinerator or to the reboiler as fuel. Back pressure on the regenerator is minimized by the use of an eductor - which is often the fuel gas valve on the reboiler.

Another mitigation measure which looks promising is to strip the aromatics from the TEG with a small stripping gas stream at the flash tank. The stripping vapors are then at a sufficiently high pressure to allow recycling to a compressor or use in a remote fuel gas system.

Others have proposed using membranes or activated carbon but these methods are not yet commercial.

This problem is one which requires careful attention in the design phase. Environmental considerations are increasingly driving the selection and operation of process equipment. In some cases the use of a dry desiccant unit (albeit at a higher capital cost) may be a lower cost alternative to glycol given the environmental impact of these BTEX emissions. "7

REFERENCES 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8

6.9 6.10 6.11 6.12

6.13

Worley, M. S., Gas Cond. Conf., Univ. of Oklahoma (April 1967). Perry, C. R., personal communication. Townsend, F. M., Ph.D. Thesis, Univ. of Oklahoma, Norman, Oklahoma (1955). Scauzillo, F. R., Jour. Petr. Tech. (July 1961), p. 697. Rosman, A,, /bid. (Oct. 1973), p. 297. Worley, M. S., Cdn. Petr. (June 1967), p. 34. Parrish, W. B., et al., Proceedings GPA 65th Annual Meeting, San Antonio, TX, 1986. Hubbard, R. A,, Lawrence Reid GC&P Conf., "Recent Development In Gas Dehydration and Hydrate Inhibition,"

Blake, R. J., Oil Gas J. (Jan. 9, 1967), p. 105. Takahashi, S., et al., GPA Technical Publícation TP-9 (1982). Jou, F.Y., et al., Fluid Phase Equilibrium, 36 (1982), p. 121. Fitz, C. W. and R. A. Hubbard, "Quick Manual Calculation Estimates Amount of Benzene Absorbed in Glycol Dehy- drator," O&GJ (Nov. 23, 1987), p.72. Robinson, D.B., Chen C.-J., Ng, H.-J.; RR-131, "The Solubility of Selected Aromatic Hydrocarbons in TEG", Gas Processors Association, May 1991.

(1 994), p. 1.



APPENDIX 6A

Formula

Molecular Weight Boiling Point' at 760 mmHg OF

APPENDIX 6A

C2H602 C4H1003 C6H 1 4 0 4 C3H802

62.1 106.1 150.2 194.2 76.1 387.3 473.8 550.0 618.1 369.3

TABLE 6A.l Physical Properties? of Glycols

Diethylene Triethylene Tetraethylene Propylene I *;y"' 1 Glycol I Glycol 1 Glycol I Glycol2

Boiling Point' at 760 mmHg "C I 197.4 I 245.5 I 287.8 I 325.6 I 187.4

?Note: These properties are laboratory results on pure compounds or typical of the product, but should not be confused with, or regarded as, specifications.

' Pure compound * Available in industrial and U.S.P. grades

Penskey-Martins Closed Cup Cleveland Open Cup


Technical Assistance Service for the üeslgn, Operation, and Maintenance of Gas Plants

6.37

GLYCOL DEHYDRATION



APPENDIX 6A



6.39

GLYCOL DEHYDRATION

e! a ii cI>

@John M. Campbd & Company


r"

APPENDIX 6A

O u7 u7 O O u7

s 8 O W m O m aD

s!+ !!!

a p

cu

cu o 7

a Q

5 ; O

O : c O co

o o o o o m o O o m w b l c w W

O d N c9


6.41

H

GLYCOL DEHYDRATION

0John M. CampbeU & Cornpony


6"

f-

APPENDIX 6A

O co

O <D

O d

c> @ O =i(u

E Q O E

c

al

O (u I

ioi

7 7

O 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a ) t D b O C u -cowyqcy O 0 O 0 o O 0 0 0 0 0 0 0 O a ) ( O b O ( u 7

O 0 O c o ( O * O ( u - c 9 C u -


O O (v

O m Y

O z !+ @

%$a P E ii

O

O M I

O O


6.43

GLYCOL DEHYDRATION


rf"

APPENDIX 6A

k O



6.45

GLYCOL DEHYDRATION

I Solution Temperature, "c

$3 John M. Campbell & Company


APPENDIX 6A

20c

18C

I 60

I& 140 - c I 20

0 loa

5 80

3 a,

O 0. m

a, - s 60

40

20

9 l 40 60 80 1

Solution Temperature, "C 10 60 80 1 O0 120 140 160

90

80

70

609

8 a. zl

- c 50 'g

40 e 30 3 20 g 10

O

-1 o

D 120 140 160 180 200 220 240 260 280 300 320 Solution Temperature, "F

Figure 6A.19 Water Vapor Dewpoints over Aqueous Triethylene Glycol Solutions

Solution Temperature, "C

Solution Temperature, O F

Figure 6A.20 Water Vapor Dewpoints over Aqueous Propylene Glycol Solutions



6.47

R I

GLYCOL DEHYDRATION

4


6.48 BP Exploration Company (Coiumbla) Ltd.

APPENDIX 6A



6.49

GLYCOL DEHYDRATION



APPENDIX 6B

6'

f-

APPENDIX 6B 0.95

0.90

0.85

0.80

0.75

0.70

0.65

0.60

0.55

0.95

0.90

0.85

0.80

0.75

0.70

0.65

0.60

0.55

99.5 wt%

10 15 20 25 30 35 40 45 50 55 TEG Circulation Rate, liters TEWkg H 2 0

l i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i ! ! i i i i i i i i i i ( 1 .o 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5

TEG Circulation Rate, gal TEGilbrn H 2 0

Figure 6B.1 Water Removal vs. TEG Circulation Rate at Various TEG Concentrations (N = 1 .O)



6.51

GLYCOL DEHYDRATION

’5

z 1

v 2 < 0.85

Figure 6B.2

I I I

1 .o 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5

TEG Circulation Rate, gal TEGllbm H20

Water Removal vs. TEG Circulation Rate at Various TEG Concentrations (N = 1.5)



APPENDIX 6B

9

2 I g 0.90

9

2 I

v 3-

1 .o 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5

TEG Circulation Rate, gal TEGílbm H20

g 0.90

Figure 6B.3 Water Removal vc. TEG Circulation Rate at Various TEG Concentrations (N = 2.0)



6.53

R

GLYCOL DEHYDRATION

3 TC

0.94

-g 0.92

10 15 20 25 30 35 40 45 50 55

TEG Circulation Rate, liters TEG/kg H20

h

2 IC

2.

0.94

s’ 0.92



r"

h

I C g

6"

P c 3

APPENDIX 6B

h P IC 3: 2

1 .o0

0.99

0.98

0.97

0.96

0.95

0.84 10 15 20 25 30 35 40 45 50 55

TEG Circulation Rate, liters TEG/kg H20

1 .o0

0.99

0.98

0.97

0.96

0.95

0.94 1 .o 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5

TEG Circulation Rate, gal TEG/lbm H20

Figure 6B.5 Water Removal vs. TEG Circulation Rate at Various TEG Concentrations (N = 3.0)



6.55

GLYCOL DEHYDRATION

NOTES:



7

GAS SWEETENING / SULFUR RECOVERY LIST OF SLIDES

ACID GAS REMOVAL PROCESSES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.1 TYPICAL ECONOMIC RANGE FOR PROCESSES TO TREAT SOUR GASES . . . . . . . . . . . . . . . . . . 7.2 AMINES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.3 TYPICAL AMINEISULFINOL TYPE SWEETENING PROCESS . . . . . . . . . . . . . . . . . . . . . . . . . 7.4 PHYSICAL SOLVENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.5 FACTORS AFFECTING PROCESS SELECTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.5 SOME SIMPLE PROCESS CALCULATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.6 CLASS EXERCISE #3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.7 MOLECULAR SIEVES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.8 NON-REGENERABLE FIXED BED PROCESSES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . SIMPLE LO-CAT/SULFEROX PROCESS SCHEME . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . SCAVANGER / NON-REGENERABLE LIQUID PROCESSES . . . . . . . . . . . . . . . . . . . . . . . . . .

7.9 7.10 7.11

AMINE SYSTEM AND SRU OPTIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.12 SULFUR RECOVERY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.13 CLAUS PLANT - FLOW SCHEMES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.14 CLAUS PLANT - FLOW SCHEMES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.15 SULFUR RECOVERY REQUIREMENTS FOR ALBERTA SOUR GAS PLANTS . . . . . . . . . . . . . . . . . 7.16

COST INDEX - SULFUR RECOVERY 99+% . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.18

EXAMPLE SCOT" PROCESS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.19 EXAMPLE FLOWSHEETS FOR AMOCO CBA PROCESS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.20 PRODUCT TREATING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.21 OVERVIEW OF LIQUID PRODUCT (NGL) TREATING PROCESSES . . . . . . . . . . . . . . . . . . . . . . 7.22 APPLICATION OF LIQUID PRODUCT TREATING PROCESSES . . . . . . . . . . . . . . . . . . . . . . . . 7.23 GAS MEMBRANE SEPARATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.24 GAS MEMBRANE SEPARATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.24 RELATIVE PERMEATION RATES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.24 GAS MEMBRANE SEPARATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.25 GAS MEMBRANE SEPARATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.26

TAIL GAS CLEAN-UP PROCESSES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.17

Technical Assistance Service for the Design. Operation and Maintenance of Gas Plants


This page left blank intentionally!


rc

f-

GAS SWEETENING I SULFUR RECOVERY

O

O

ACID GAS REMOVAL PROCESSES

Ph

Chemical Absorption (H2S & C02)

Amines: MEA, DGA, DEA, DIPA, MDEA

Carbonate: Benfield, Catacarb

1 Absorption (H2S & C02)

Sulfinol

Propylene Carbonate (Fluor Solvent)

Se 1 ex o 1

Rectisol

Purisol

IFPEXOL

0 Fixed Bed (usually H2S)

Molecular Sieve

Iron Sponge

Sulfatreat

Zinc Oxide

0 Direct ConversiodRedox (H2S)

Stretford

LO-CAT

SulFerox

Chemsweet

Sulfa-Check

Slurrisweet

Sulfa-Scrub

Numerous proprietary scavangers

0 Other (mainly C02)

Membranes

Extractive Distillation

Technical Assistance Service for the Design, Operation and Maintenance of Gas Plants

7.1


TYPICAL ECONOMIC RANGE FOR PROCESSES TO TREAT SOUR GASES

inn I I100

Solid Bed Primary Tertiary Hot Physical Membranes Bulk and and and Amines Potassium Solvents Extractive

Scavenger Secondary (Note 6) Carbonate Distillation (Note 8) Amines Processes

1.

2.

3a.

3b.

4.

5 .

6. 7.

8.

9.

10. EuropeaníNorth American economic basis. I1

PrimaryiSecondary Amines may be economic up to approximately 20% acid gas (H2S + COZ).

Physical solvents (e.g. Selexol) may be economic down to 510% acid gas (50 psi [3 bar] partial pressure). Membranes may be economic down to 8-12% acid gas or even lower - strongly dependent on power costs (for compression) and plant capacity. Smaller membrane units economic down to 3-5% COZ. Lower economic limit for Bulk and Extractive Distillation not yet well-defined.

COZ recovery unlikely to be economic above 90% feed COZ content. Tertiary Amines: MDEA (also TEA, DMEA). Range for combination solvent processes (such as Sulfinol) felt to be roughly similar to that of Hot Potassium Carbonate and Physical Solvent Processes. Solid bed (mol sieve, iron oxide, zinc oxide, Sulfatreat, etc.) usually only economic for significantly less than 1% acid gas and/or < 100-200 kg/d sulfur. LO-CAT/SulFerox processes not addressed by this diagram.

rl

T7


r'

r"


v,

f 3

I-

L

E s 2 5 -= o Br, L '5


7.3


TYPICAL AMiNE/SULFINOL TYPE SWEETENING PROCESS

AlMER

A simplified schematic of an amine/sulfinol process is shown above. The flow is similar to that of a glycol dehydration or lean oil absorption plant. The sour gas is contacted with the lean amine/sulfinol solution in an absorber, the rich amine/sulfinol solution is flashed to liberate any coabsorbed hydrocarbons, heated, and then sent to a regenerator. In the regenerator the absorbed sour gas components are driven off, the overhead is cooled. The condensed stream (water) is pumped back to the regenerator as reflux. The bottom product, the lean amine/sulfinol, now stripped of the acid gas, is cooled and pumped back to the absorber. The absorber operates at sales gas pressure and regenerator at slightly above atmospheric pressure.

The flow schemes for the potassium carbonate and physical solvent processes are similar to that shown above. They differ in volume of circulation of solution, temperature levels, and methods for removing coabsorbed hydrocarbons.

The Sulfinol process is a solvent process utilizing three chemicals, a physical solvent, a chemically reactive component and water. The physical solvent is sulfolane - tetrahydrothiophene 1, 1 dioxide. The chemically reactive constituent is an amine: DIPA or MDEA. The mixture of the three components is tailored based on the duty. For example if H2S and COZ are to be removed, a formulation might be 45% amine, 40% sulfolane, and 15% water. A 50% amine, 25% sulfolane and 25% water mixture is applied for deep COZ removal. Other formulations are derived when complete H2S removal and only partial C02 removal is required. The sulfinol process is able to treat gas down to H2S and total sulfur pipeline specifications and to treat down to 50 ppm (m) level of COZ. ,-?



f" PHYSICAL SOLVENTS

0 Advantages Good for bulk removal (high acid gas partial pressure)

Selective

Lower energy requirements

Less corrosive

Can remove mercaptans

0 Disadvantages

Expensive

Proprietary-license required

Co-absorption of heavy hydrocarbons

More difficult to meet tight effluent specs (COZ and H2S)

FACTORS AFFECTING PROCESS SELECTION

0 Effluent Gas Specifications

0 Feed Gas Composition

Acid gas partial pressure

H2S-C02 ratio

Mercaptans, COS, CS2, etc.

Hydrocarbon content

0 Energy Costs

0 Selectivity

0 Feed Gas Rate


7.5


MEA

DEA

MDEA

SOME SIMPLE PROCESS CALCULATIONS

15-22 std m3/m3 amine solution



[2-3 scf/uS gal]

[3-6 scf/uS gal]

[5-8 scf/uS gal]

-

T/d ofsulfur= (13.6)(106 stdm3/d)(H2S conc, %)

= (0.38)(MMscfd)(H2S conc, %)

Typical amine regenerator duties: 250-330 MJ/m3 solution [900-1200 Btu/U.S. gal]



r"

CLASS EXERCISE #3

The gas in Class Exercise 1 is to be treated in an amine system to remove H2S and COZ.

1) The solvent is DEA (non-selective) which will remove all of the H2S and COZ.

a) What is the acid gas rate in std m3/d [scf/d] and composition?

b) Estimate the solvent circulation rate in m3/h [US gpm].

c) Estimate the regenerator reboiler duty in MW [Btu/hr].

2) The solvent is MDEA and is designed to slip 50% of the COZ.

a) What is the acid gas rate in std m3/d [scf/d] and composition?

b) Estimate the solvent circulation rate in m3/h [US gpm].

c) Estimate the regenerator reboiler duty in MW [Btu/hr].

~~~ ~~ ~ ~


7.7


MOLECULAR SIEVES

LEFT TOWER DRYING RIGHT TOWER HEATING

STPIJU: I S U

-_--

WET REGENERATION

OR FUEL (WTDRECYCLE

---- DRY REGENEFiATlON

REGENERAMN GAS HEATER

0 Used for simultaneous removal of HzO and H2S

BP-Amoco Saaja plant in Sharjah

Esso Australia Longford plant in Victoria, Australia

0 0

Effective for low H2S concentrations (< 1000 ppm)

Effective loading for H2S, C02, and RSH is lower than water

Lower equilibrium loading

Longer MTZ

Competition from other components

0 5A sieve most common for natural gas sweetening

13X sieve most common for product treating

0 Also removes light mercaptans

0 Regeneration gas is sour

Flared

Fuel

Amine system

0 COS formation



P NON-REGENERABLE FIXED BED PROCESSES

Gas in Y

1 A

Y

Adsorption Vessels

(Zinc Oxide, Sulfatreat)

1 - Gas Out

Zinc Oxide (Puraspec)

Iron Oxide (Sulfatreat, Iron Sponge)

0 Generally favorable when sulfur recovery less than 100 kg/d E220 lbdday]

IJ LeadAag bed operation

0 Disposal of spent material

0 Pressure drop

~ _ _ ~ ~~


7.9


A

2 I

+ í r + $ 0

+ f i +

7.1 O BP Exploration Company (Columbia) Ltd.

r"

r'


SCAVANGER / NON-REGENERABLE LIQUID PROCESSES

Chemsweet

Sulfa-Scrub

Sulfa-Check

Sulfa-Treat

Several others

Generally favorable economically when sulfur removal is less than 50-100 kg/d [110-220 lbdday] low Capexhigh Opex

Solvent cost

Solvent disposal

Water/oil soluble

Contact method (bubble tower, static mixer, etc.)

Mercaptan removal


7.1 1


L

B n 7 4

a c E!! > '0

C -

L

dl cn

I

c (d a

w c a > O cn -



SULFUR RECOVERY

H2S+1XO2 * H 2 0 + S 0 2 2 H,S+ SO2 3 3 S + 2 H2O

When H2S has been removed from a gas stream it must be disposed of. The easiest method of disposal is by burning the H2S to SO2 and venting the incinerated gas. This is typically not possible because of environmental restrictions. Alternatively, the amount of H2S may be such that an economically attractive option is to convert the H2S to elemental sulfur, a marketable product. r" The most common method of converting H2S to sulfur is the Claus Process. This process is about 100 years old and is used widely throughout the world. The basis for this scheme is to pass H2S and SO2 (in the vapor phase) at a 2:l ratio over an alumina (A1,0,) catalyst where the following reaction takes place:

2H,S+SOZ*3S+2H,O

A schematic of a Claus plant is shown in the above figure. There are a number of variations to this scheme involving various methods for temperature and H2S/S02 ratio control, but the principles applied are the same.

The acid gas (from the top of the amine regenerator) at a pressure slightly above atmospheric is mixed with enough air in reaction chamber to bum one third of the H2S to SO2 (to satisfy the above equation) and all the hydrocarbons to COZ. This is highly exothermic and the temperature out of the reaction chamber is typically in the range of 650" to 1000°C. High pressure steam may be generated in a waste heat boiler at the outlet of the reactor furnace.

Some sulfur is produced in this waste heat boiler and it is drawn off as liquid from the sulfur condenser and sent to storage. The vapor from the condenser is reheated to raise its temperature to around 230°C and sent to the first catalytic reactor. The gas leaving the reactor is cooled to around 140°C to condense the produced sulfur. The non-condensed gas is again heated up to about 230°C and sent to a second reactor. The process of condensation and liquid sulfur draw-off is repeated.

r" Two or three converter steps will be applied. Depending on the H2S concentrations in the acid gas, with two beds about 92 to 94% of the H2S may be converted to sulfur and about 95 to 97% with three beds.


7.1 3


CLAUS PLANT - FLOW SCHEMES

f S

Condenser

S S

(a) Straight-Through Generally used when acid gas is 50% HzS or higher

S

40% min.

S

Generally used when acid gas is:

4

Condenser

(b) Split-Flow S S 20-50% H2S without preheat or IO-25% HzS with preheat

1 Condenser Air

Sulfur Recycle S (excess)

(c) Sulfur Recycle Generally used when acid gas is 10% H2S or lower

B Converter

S

denotes Reheater S

t

S S

(d) Direct Oxidation-Selectox Generally used when acid gas is: 2 or 3% H2S up to 5% (up to approx. 25% H2S with recycle)


'7

f-


CLAUS PLANT - FLOW SCHEMES

Acid Air Gas

Steam Reaction Furnace

(a) Two Converter Claus

Reheater Incinerator

Steam u n

Acid Air Gas (b) Superclaus 99

n

r Steam

Reaction Furnace

Gas Coolei

I I

Acid i r Gas (c) Superclaus 99.5

Incinerator I I

t S


7.1 5


u, I- 4 e ü 0

4-



-

f-

or - React

PROCESSES Extend Claus

DRY BED

SULFREEN

r in

Liquid Phase with

Catalyst

Oxidize to

WET SO2 - SCRUBBING and

PROCESSES Absorb or

React

TAIL GAS CLEAN-UP PROCESSES

TOWNSEND

ASR SULFOXIDE

WELLMAN-LORD

ATS (Ammonium Thiosulfate:

UCAP

CLAUS

PROCESSES CLAUSPOL Claus

Reaction

\ - u Reduce to H2S and

Absorb ^I

SCOT (Shell) BSRP (BeavoNStretford) SU LFTEN BSWMDEA BS WSELECTOX LO-CAT SULFEROX MODOP


7.1 7

I

ir 1


COST INDEX - SULFUR RECOVERY 99+%

240

220

200

% 180

8 160 o

14C

12c

1 oc $

X

c - w

Clam (2R) + SUPERSCOT

I I I

Claus (2R) + LS SCOT

Claus (2R) + Add-on SCOT

Claus (2R) + Casc. SCOT

SUPERCLAUS (3+1) I I

i 97 98 99 1 O0 Sulfur Recovery (%)

Recovery System ~~

Two-Stage Claus with Bypass Reheat Two-Stage Claus with Indirect Reheat Three-Stage Claus with Indirect Reheat Two-Stage Claus (96% Recovery) with One CBA Stage in Tail-Gas-Cleanup Unit Two-Stage Claus (96% Recovery) with Two CBA Stages in Tail-Gas-Cleanup Unit Two-Stage Claus (95.5% Recovery) with Sulfreen Process Tail-Gas-Cleanup Unit Two-Stage Claw (95% Recovery) with IFP Process Tail-Gas-Cleanup Unit Claus Plant (94% Recovery) with SCOT Process Tail-Gas-Cleanup Unit

Claus Plant with Beavon Process Tail-Gas-Cleanup Unit

Claus Plant with Welman-Lord Process Tail-Gas-Cleanup Unit

Overall Ciaus plus TGCU Recovery

Capability, %

90-94 93-96.8 95-97.8

Up to Approx. 99 Up to Approx. 99.5 Up to Approx. 99

98.1-99.4 99.9+ 99.9+

99.9


rs)


r"

F"


7.19


-;;<:- I I I I I I

1 4$ I I I I

-' 5 <:+I - " I

I I I I I -

-

I w 4 I

I I - -

A

- S O .-


I


r"

W Z + 3 a m s

W z

t I v


7.21


Amines

Bulk Removal

DEA LiquidíLiquid Contactor, also use DGA, MEA

Malaprop-Huntsman

ADIP (DIPA) Shell

Molecular Sieve

Dry and Sweeten at Same Time

Fixed Bed

Usually Removal of Trace Amounts

Zinc Oxide

Mixed Metal Oxides

iron Sponge, Iron Oxide

1 Doped Aluminas

Other Processes

OVERVIEW OF LIQUID PRODUCT (NGL) TREATING PROCESSES

Caustic

NaOH (Caustic Soda)

Can be used for Bulk Removal

Regenerativernon-Regenerative

Merox (UOP)

Mericat (Merichem)

Caustic-Free Merox

KOH

Withiwithout Alcohol

Miscellaneous


P

(XI = YCS, scc notes

Liquid Processes

Glycol


COZ HZS HZO carbon hydrogen

water dioxide sulfide

X

APPLICATION OF LIQUID PRODUCT TREATING PROCESSES

MEA

DEA

DlPA (ADIP)

DGA"

X X

X X

X X X X

Hot Potassium Carbonatc (not liquid tcsting)

Caustic Soda íNaOH

X X

regenerative (circulatc)

non-rcgcncrativc (batch)

MeroxSM

swcctcn

extraction

X X

Zinc Oxide íZnO)b I I 1 x

Pcrco (Coppcr) 1 I

I

Doped Alumina' ; 1 ~ ; Molecular Sicvc' I

Alumina

Silica Gel I

~

Solid Processes

Iron Spongc

Sulfatreat

little data

X X

RSH Low MW

mercaptans Le. C4-

Potassium Hydroxidc (KOH)

X

X

X

X

X

X

X

(X)k

High MW

I I

X 1 I

X I I

only at 150-200°C with ZnO; 8O-10O0C with mixed oxidesb

degraded by COS - need reclaimer

some removal possible above 50°C [120aF], longer I

contact time J not for NGL with significant CSt

some removal may be possible

Menchem DrOcesses similar


7.23


Where:

GAS MEMBRANE SEPARATION

ni =l lAApi

ni = permeation rate of component i, kmolsh [lbmoih] ni = permeability of component i, kmols/s.m*.kPa [lbmol/hr-ft2-psi] A = membrane area, m2 [ft2]

Bpi = partial pressure difference, kPa [psi]

Molecules w/a high permeability are said to be fast molecules e.g. HzO, HzS, COZ

Molecules w/a low permeability are said to be slow molecules e.g. C,, Cz, C3, etc.

The higher the ratio of fast molecule permeability to slow molecule permeability the higher the membrane selectivity.

Membrane materials: cellulose acetate, polyamide, polysulfone

RELATIVE PERMEATION RATES

Fast Slow I

GAS MEMBRANE SEPARATION TJ Increasing membrane area increases purity of slow components in residue stream, but de-

creases purity of fast components in permeate stream

D Decreasing membrane area increases purity of fast components in permeate stream, but de- -I creases purity of slow components in residue stream


r'


I I cu I I

Q Q o 0

II o" II o 0 Q Q

8 a

\ \ \ \ \ \ \ I

cci v) Q O O O

T J F

LLa

.-

g I I

\

al a

II o" II o 3 Q Q

u-a


7.25


COZ, mol%

C , . mol%

GAS MEMBRANE SEPARATION

Residue Permeate

Feed A B A B

9.0 2.0 5.0 45.3 55.6

76.6 81.7 79.7 50.0 39.6

Carbon Dioxide

~~

C2+, mol%

H20, 1bmiMMscf

T, O F

P, psia ~

Rate, lbmoííhr

I Permeate b

__

14.4 16.3 15.3 3.6 2.8

80 0.6 8.2 1 .o* 1.9*

110 101 105 101 1 o5

990 985 985 20 20

1 O0 83.8 92.0 16.2 8.0

Low Pressure Carbon Dioxide, Water (and Some Methane)


Section 8

f-

ADS0 RPTl O N PRI N CI PLES TABLE OF CONTENTS

PAGE #

DESIGN AND OPERATION ........................................................................ Process Characteristics .......................... Regeneration and Recovery ............................................................ General Specifications ......................................................................................

LIST OF FIGURES FIGURE # PAGE # 8.1 8.2 8.3 8.4 8.5 8.6

Process Flow Diagram - Typical Adsorption Plant ....................................................... Typical Process Flow Diagram for Short Cycle HRU .......................................... 8.3 Regeneration Gas Specific Gravity vs. Time on Heati th Pecan Lake HRU) ....... 8.4 Hydrocarbon Equilibrium on Silica Gel ........................................................................................ 8.5

Typical Regeneration for Hydrocarbon Adsorbers ....................................................................... 8.8 Illustration of the Adsorption/Desorption Process .........




/1 NOTES:


8.ii ~ ~~ ~ ~ ~


Section 8

ADSORPTION PRINCIPLES Adsorption can be applied to NGL recovery in much the same way as in dehydration. These

units are called Hydrocarbon Recovery Units (HRU's) or Short Cycle Units (SCU's). The dry desiccant most commonly used in such units is silica gel, but activated carbon is also used. An example of activated carbon application is hydrocarbon removal from acid gas feeding a sulfur recovery unit (SRU). Silica gel has a strong affinity for the heavier hydrocarbons (C,+) as well as for water.

The process equipment utilized for SCU's is virtually identical to that used for adsorption dehydration (Figure 8.1). However, the dry desiccant units designed to recover hydrocarbons as well as dehydrate gas use a much shorter cycle time than for dehydration only, hence the term "short cycle unit." Typically the cycle time will be from 30-40 minutes (as opposed to cycle time of greater than 8 hours for a dehydrator). Although such units have been designed for hydrocarbon dewpoint control, complete heavy hydrocarbon removal is not always achieved, which makes dewpoint control difficult.

To completely remove the adsorbed hydrocarbon regeneration gas, temperatures of around f-

260°C [500"F] are applied.

At one time, when gas prices were low compared to liquid prices, such facilities were common in the United States. Early Canadian gas processing industry used this technology due to the lack of a LPG market. One early application was the Btu control of fuel gas used in pipeline compressors since overly-rich fuel can cause detonation (backfire) in the power cylinders of large integral-type reciprocating compressors. However, they are very energy intensive since the beds and towers must be heated from ambient temperature to 260°C [500°F] during each cycle. As the value of gas has risen, the economics of such units have deteriorated and they are not commonly used at the present time.

The desiccant materials have an enormous amount of surface area per unit volume (170 O00 O00 ft2/ft3 for silica gel). They also have a high physical attraction for hydrocarbon molecules which "condense" on the surface of the adsorbent. The degree of "condensation" or adsorption is related to the volatility of the component being adsorbed. The less volatile components are more readily adsorbed than the highly volatile components. The adsorption process is identical to that employed to separate and analyze materials in a chromatograph.

Adsorption is a batch process. Once the adsorbent bed is saturated with hydrocarbons it must be regenerated. Regeneration is accomplished by the application of heat. A regeneration gas stream (usually 5-10% of the total inlet gas stream) is heated to 400-600°F and passed through the bed. Those components which had previously condensed on the adsorbent are vaporized and carried away by the regeneration gas stream. Since the volume of the regeneration gas stream is much smaller than the inlet gas stream, the recoverable components are concentrated in the regeneration gas. When the regeneration gas stream is cooled these components are condensed. Normally, the condensed stream of extracted products must be stabilized before being sold. So that the gas may be continuously

r"



8.1


5

' I

:e



DESIGN AND OPERATION

processed, more than one adsorption tower is employed. Two and three tower systems are the most common in these hydrocarbon recovery units or HRU's. r'

A typical process flow diagram for a three tower SCU is shown in Figure 8.2. One tower is heating, one cooling, and one adsorbing in the arrangement. SCU's differ only in the number of towers and the type of regeneration loop employed. An open system utilizes a portion of the inlet or residue gas as the regeneration gas stream. An artificial pressure drop (usually 30-50 psi) is created across a control valve to provide the necessary driving force for the regeneration gas to flow through the system. This method, while simple, wastes energy. A closed system utilizes a compressor (usually a rotary type with electric motor drive) to recirculate the regeneration gas stream in a closed loop. This method is more energy efficient but generally requires additional maintenance. The choice between the two is based on operating costs, location, etc.

+ Residue Gas Inlet Desiccant Desiccant Regeneration Desiccant Regeneration

Scrub be r Tower Tower Gas Heater Tower Gas Scrubber I Figure 8.2 Typical Process Flow Diagram for Short Cycle HRU

DESIGN AND OPERATION The most important design and operating variables in the hydrocarbon recovery application are: 1. Cycle length

Cycle time must be sufficient to fully saturate the bed but not so long as to allow breakthrough of recoverable hydrocarbons. This timing is normally best determined by analysis of the residue gas.

2. Regeneration gas temperature

The regeneration gas temperature and cycle time must be sufficient to drive the hydrocarbons off the bed. Insufficient time or temperature will leave the bed partially saturated, which cuts into unit capacity. One method of checking the regeneration cycle time is shown in Figure 8.3.

3. Regeneration gas cooler temperature

The regeneration gas must be cooled to allow condensation of recoverable hydrocarbons. Obviously, the colder the condensing medium the more hydrocarbons which can be condensed and hence recovered.

OJobn M. Campbell & Company

Technical Assistance Senrice for the Design, Operation, and Maintenance of Gas Plants

8.3


~

0.67

0.66

0.65

0.64

0.61

0.6C

Time, min I

Figure 8.3 Regeneration Gas Specific Gravity vc. Time On Heating Cycle (South Pecan Lake HRU)

SCU’s typically result in low recoveries relative to other extraction processes. Design recoveries are usually 70-80% of the pentanes and heavier and 10-20% of the butanes. Actual operation normally yields 50-60% of the pentanes and heavier and essentially no butanes. These units have low capital cost and can be operated unattended. The product can usually be blended into a crude oil or condensate stream which makes the application ideal for remote locations. In addition, most adsorbents (activated charcoal is an exception) also dehydrate the gas stream. SCU’s are typically used for dewpoint control. With increased NGL prices, adsorption type recovery systems will seldom be the optimum choice.

The basic mechanism for hydrocarbon recovery is similar but describing the multiple zone behavior is difficult.

In a dehydrator, the purpose of the condenser is just to remove the desorbed liquids from the gas stream. In the short cycle plant, condenser operation has a critical effect on recovery. The adsorption bed simply serves to concentrate the recoverable components so that condensation is more efficient. The temperature and pressure of condensation is a critical parameter governing plant economics. A carbon plant with a refrigerated condenser is capable of good ethane recovery in some instances. A gel plant with ambient condensation is limited to some butane recovery and 75-90% of the pentanes. There are many alternatives in between.

Consider that the hydrocarbon deposited on the bed is picked by 10-15% as much regeneration gas. The net effect is to make the gas 6-10 times richer in condensable hydrocarbons, thus making recovery easier. A refrigerated condenser may be used to further enhance recovery but the condensa-



DESIGN AND OPERATiON

F"

/-

tion temperature should exceed the hydrate temperature. In some cases enough liquid may be recovered to reduce simultaneous dehydration cost below that of glycol.

Ambient cooling plants have been used primarily on lean gas streams where other methods of processing were not economically attractive. The untapped potential for hydrocarbon dewpoint control and as an adjunct to refrigeration appears large particularly at higher pressures. The biggest obstacle to more widespread use of the process would appear to be the rather unimaginative design methods used to date.

Process Characteristics The capacity of most desiccants is about the same for hydrocarbons as for water. Activated

carbon, of course, has no effective capacity for water.

Figure 8.4 shows the equilibrium capacity of silica gel for various hydrocarbons in a two component gas where the second component is methane. The figure shows both static equilibrium (from cell tests) and dynamic equilibrium (from flow tests). It is apparent that capacity is not affected very much by flow.

0.500

0.400

0.300

0.200

0.100 0.080

0.060 0.050 0.040

0.030

0.020

0.010

0.008

0.006 0.005

0.004

0.003

0.002

n nni Y."" I

0.02 0.03 0.05 0.07 0.1 0.2 0.3 0.5 0.7 1.0 2.0 3.0 5.0 7.0 10.0

ye, mol% Component in Inlet Gas

Figure 8.4 Hydrocarbon Equilibrium on Silica Gel



8.5


/1 Notice that all curves are approaching the monolayer capacity of the gel. The monolayer capacity is found by assuming that only one layer of molecules is held to the solid surface. Knowing both surface area and molecule size, one may compute the capacity. This same characteristic has been found for all gels, aluminas and molecular sieves. This means that ultimate capacity for any component is fixed by surface area - provided that the component is small enough to enter the interior of the adsorbing particle.

Actual capacity for any component is fixed by the zone movement previously described, bed geometry, equilibrium capacity and gas flow rate. The surface of the adsorbent is always occupied by some molecule. As the zone of a given component progresses down the bed it must displace the molecules already there. The rate of displacement depends on their relative wettability.

Theoretically, the zone for any component cannot move any faster than it can completely displace the materials ahead of it. In actual practice, at commercial flow rates, the zone tends to "over- run" said displacement. Therefore, true chromatographic separation does not occur. This is illustrated in Figure 8.5.

1 .o C - c o

Figure 8.5 Illustration of the AdsorptionlDesorption Process

1.2

1 .o

0.8

0.6

0.4

0.2

O

Time, min

Example of breakout cutves for adsorption of multicomponent mixtures, showing displacement of lighter by heavier component.

A 0 Velocity, Run %CE %CE fümin No.

- 0.616 0.541 22.2 106 0.447 0.395 20.7 108 1.61 1.42 22.9 112

(b)

Time

As shown in Part (a), once the front of a given zone reaches the outlet of the bed the ratio of outlet to inlet concentration (C/C,) starts to increase. When this ratio reaches unity, all primary adsorption ceases for that component. Desorption now begins because the zone behind it is replacing the material adsorbed. The concentration ratio rises above unity but again approaches unity when this next zone starts breaking out the end of the bed. This process continues until the cycle is terminated.

Area A is representative of the amount of Component 1 adsorbed and Area B of that amount desorbed by the next zone. The latter is smaller than the former. At times (3B for Component 2 (Area A - Area B)/Area A is generally about 0.35-0.40. Thus, immediate displacement has not occurred. 7




r"

Inlet Mol% (C,) Time, min.

-

2 12 22 32 42

c3 1 .o20 0.804 0.976 0.938 0.962 0.970

iC5 0.120 0.03 1 0.074 0.075 0.077 0.155

nCc 0.085 0.012 0.03 1 0.043 0.065 0.130

Even if one recognizes the 6-10% error in sampling and analysis, no sharp separation has occurred. The propane has broken through in less than two minutes. This is not a very efficient plant because even good pentanes recovery is not obtained early in the cycle. Beyond this point, the exit stream is being enriched.

If liquid recovery is the goal, some net amount of component is available even after its zone passes from the tower. The recovery will simply be less. If hydrocarbon dewpoint control is the goal, such enrichment is probably intolerable for components heavier than the butanes. In such case, the iso-pentane OB represents maximum cycle time. For these reasons, liquid recovery plants tend to use longer cycles than dewpoint control plants since the loss of efficiency is not so apparent. Unfortu- nately, the bulk of the recovery plants use cycle times too long for most efficient recovery. Recovery is not limited by the process but by the way it is applied.

Part (b) of Figure 8.5 shows the adsorption-desorption process for a test run. The gas in question contained methane, ethane, negligible propane and butanes, and the amounts of pentanes and hexanes shown. There was no heavier component present to displace the hexanes. Water content was negligible.

f"

With activated carbon the zones tend to move slower. For one thing, water does not promote displacement. Basically though, the zone speed is lower because of the greater affinity for the lighter hydrocarbons.

Regeneration and Recovery Figure 8.6 summarizes the regeneration behavior of a short cycle plant. Notice that the materi-

als do not desorb at a constant rate. The pentanes and lighter start desorbing almost immediately. The hexanes and heavier concentration in the exit gas peak after a finite time.

Obviously, the composition of the stream to the condenser varies continually with time. For this reason a series of flash calculations must be made with time to accurately represent the liquid recovery to be expected. The minimum necessary amount of regeneration gas should be used to enhance the concentration of recoverable components in the condenser. The use of refrigeration will likewise enhance recovery. Even at refrigeration levels as high as 16-20°C [60-68'F], marked recovery efficiency may be obtained.

General Specification Proper detailed design is complicated enough that a computer is necessary. A design method

will not be outlined herein although the principles governing it have been. It is sufficient to say that the general empirical methods to date fail to utilize the full potential of the adsorption process.

r"



8.7


SI Units mímin

Time, rnin

a) With Butane, Pentane, Hexane

Y FPS Units

Wmin

Time. min

b) With Pentane, Hexane, Heptane

Notes: 1. Concentrations for n-butane and n-hexane in Figure a) are actually one-tenth of the values shown on the y-axis. 2. Concentrations for n-pentane and n-hexane in Figure b) are actually one-tenth of the values shown on the y-axis. 3. Composition measurements were taken at 3 minute intervals.

" O 20 40 60 80 100

- E 0.6 - -

a

Time, % of Cycle Length Bed Outlet Temperature, "F

c) Outlet Bed Temperature Produces d) Characteristic Pentane Content of Regeneration Characteristic Curve

Typical Regeneration for Hydrocarbon Adsorbers

Gas vs. Bed Outlet Temperature

Figure 8.6

Most obvious problems stem from improper specification. A liquid recovery guarantee is useful for planning purposes but is more likely to be a basis for controversy than a measure of plant performance.

I recommend the following minimum specifications:

1. Gas flow limited by

Where: vg = gas superficial velocity pg = gas density D, = partial diameter (average) C = constant

kg/m3 m

1200

lb/d ft

j 540

-l The particle diameter, Dp, is found from the mesh size of the desiccant used. A typical silica gel will have a size like 3-8 mesh or 4-8 mesh. The first number is the size of screen all particles pass through and the second number is the size

@John M. C í ~ ~ ~ b e l i & ComppnY


r'


opening all particles are returned on. Therefore, the particle diameter range is the following.

4.699 O . 185 2.362 0.093

rf"

Velocities up to 18 d m i n [60 ft/min] have been used successfully in some installations.

2. Cycle length should be fixed by the following considerations: a. Not less than 15 minutes for gas containing pentanes and heavier. b. Breakthrough time for the component for which recovery is desired or

which must be removed for dewpoint control. 3. Bed length should be at least 15 feet. 4. Regeneration should be to at least 230°C [446"F] and preferably 260°C [500°F]

when processing gases containing pentanes and heavier. 5. The alternative of using refrigeration instead of ambient cooling in the condenser

should be considered. Items 1-4 are not independent; each affects the other. If the cycle length is less than 15 min-

utes it is almost impossible to remove the "heel" properly, which ultimately affects the economics adversely. If this 15 minutes is greater than the breakthrough time for key component, some compromise is needed. Breakthrough time depends on gas velocity and bed length (for a given gas composition and adsorbent. Economics and/or process needs will govern the compromise. As a matter of information - so one may make an intelligent decision - it is wise to also specifi that the vendor furnish you with adsorption efficiency as well as condenser recovery. Adsorption efficiency is simply that fraction of the component entering during the proposed cycle length that is retained on the adsorbent. This enables you to not only compare the relative merit of competitive bids but to make necessary changes prior to purchase. Any design approach that uses an overall method of calculation from inlet gas to stock tank may yield a workable plant but seldom an optimum one.

The adsorbent bed may contain more than one adsorbent. Such composite bed will behave like two towers in series and should be treated as such.

F OJohn M. Campbeli & Company


8.9


7 NOTES:



Section 9

r"

P

SOLID BED DEHYDRATION TABLE OF CONTENTS

PAGE # BASIC PROCESS FLOW ......

Two Tower Unit ........ ............................................................................................. 9.2 Three Tower Units ................................................................................................................................. 9.3

DESICCANT TYPES .................................................................................................................................... 9.5 Activated Alumina ........ ....................................................................... 9.5 Silica Gel ................................................................................................................................................ 9.6

CHOICE OF DESICCANT ........................................................................................................................... 9.9 THE NATURE OF ADSORPTION .............................................................................................................. 9.11 ADSORBER DESIGN ................................................................................................................................... 9.13 REGENERATION DESIGN ......................................................................................................................... 9.17

Steel Shell ...... ............................................................................................................................ 9.21 .......................................................................................... 9.21

Calculation of QD ....... ...................................................................................................... 9.21 Regeneration Gas Heater ......................................................................................................... 9.23 Regeneration Gas Coo ............................................................... 9.23

MECHANICAL DESIGN CONSIDERATIONS ......................................................................................... 9.26 Sieve Loading ......................................................................................................................................... 9.27 Mol Sieve Unit Operations .................................................................................................................... 9.27

SPECIAL ISSUEShiEW DEVELOPMENTS .............................................................................................. 9.29 LIQUID DRYING

FIGURE #

9 . I 9.2 9.3 9.4 9.5 9.6 9.7 9.8 9.9 9.10 9.1 1

LIST OF FIGURES PAGE #

Typical Two-Tower Solid Desiccant Dehydration Units .............................................................. 9.3 3-Bed Dry Desiccant Dehydration System .................................................................................... 9.4 Schematic View of a Typical 3-Tower Plant Using Cooling and Heating in that Order ........... 9.5 Part Way Through Adsorption Cycle ............................................................................................. 9.11

............................... 9.12 ............................................. 9.15

Silica Gel Capacity as a Function of Time in Service .................................................................. 9.16

................................................. 9.18 Types of Internal Insulation ............................................................................................................ 9.20 Approximate Pressure Vessel Wall Thickness .............................................................................. 9.22

Representation of the Adsorption Zone Concept Hz0-4A Isothenns - Intermediate H 2 0 Pressu

Adsorber with Floating Screen and Balls for Improved Gas Distribution .................................. 9.17 Temperature Curves for a 2-Tower Adsorber Plant .....

@Job M . Campbell & Company

Technical Assistance Service for the Design. Operation. and Maintenance of Gas Plants

9.1


LIST OF TABLES FIGURE # PAGE # 9.1 Molecular Sieves .............................................................................................................................. 9.7 9.2 Molecular Diameters ........................................................................................................................ 9.8 9.3 Properties of Various Adsorbents ................................................................................................... 9.8 9.4 Application Guides for Molecular Sieves and Silica Gel .............................................................. 9.10


9.ii BP Exploration Company (Columbia) LM.

Section 9

I"

SOLID BED DEHYDRATION lycol dehydration is used for most applications where dehydration of natural gas to pipeline speci- G fication is required because glycol units cost less initially and are cheaper to operate. Solid bed

dehydration (also called dry desiccant or adsorption dehydration) is often the superior alternative in applications such as:

1.

2.

3.

4.

5 .

Dehydration to water dewpoints less than 4 0 to -50°C [ A 0 to -58"F], such as those required upstream of NGL extraction plants utilizing expanders and LNG plants.

Hydrocarbon dewpoint control units where simultaneous extraction of water and hydrocarbon is required to meet both sales specifications. This is sometimes used for hydrocarbon dewpoint control on high pressure lean gas streams.

Simultaneous dehydration and H2S removal of natural gas.

Dehydration of gases containing H2S where the, H2S solubility in glycol can cause emis- sion problems.

Dehydration and sulfur compound (H2S, COS, CS2, mercaptan) removal for LPG and NGL streams.

Adsorption is defmed as the adhesion of a layer of molecules to the surface of a solid or liquid. It is distinguished from absorption, which is the transfer of molecules through an interface into a bulk solid or liquid.

Two types of adsorption on solids exist.

1. Chemical adsorption (or chemisorption) is a chemical bonding of molecules to surface atoms.

2. Physical adsorption (or physisorption) is a concentrating of molecules near a surface by molecular attractions, similar to the attractions that cause a vapor to condense under appropriate conditions.

For dehydration and desulfurization of natural gas and NGLs, we are concerned primarily with the physical adsorption of water and sulfur compounds and secondarily with the chemical adsorption of reactive hydrocarbons that over time can drastically reduce the effectiveness of the adsorbent. Desulfu- rization feasibility depends on what is done with the regeneration gas - if commingled with residue gas it may sour it again. If sweetening liquids, regeneration gas could sour the residue gas also. This is a problem, even though sulfur compounds will only come off for a short period of time during the heating cycle. This sulfur "spike" can cause plant shutdowns andor purchaser curtailment.

F- Adsorption processes occur in both fixed and moving beds. In moving beds, the operation is

continuous with fluid and adsorbent contacted countercurrently. The analysis of such an operation is @John M. Campbell & Company


9.1


-l analogous to the analysis of an absorption tower. However, in spite of the relative simplicity of the analysis, moving beds are rarely used because of particle attrition and mechanical complexity.

In contrast, fixed-bed adsorption is a batch process and is, therefore, not amenable to a steady- state analysis. In fixed-bed adsorption, the fluid to be purified is passed through a large vessel packed with adsorbent.

At the end of the adsorption cycle, the bed is taken off stream for regeneration.

These notes are concerned particularly with the dehydration of gases in fixed beds. Dehydra- tion of liquids and product treating are covered briefly.

BASIC PROCESS FLOW Solid bed dehydration plants use a solid desiccant described in detail in later sections. The

desiccant is contained in two or more towers. Two or three towers are most common, although plants exist with four or more. The two-tower plant can be considered the basic unit and will be considered first.

Two Tower Unit Figure 9.1 shows a typical two tower solid desiccant dehydration unit. During dehydration,

wet gas from the inlet gas separator enters the top of the tower in adsorption service. Downflow drying of gas is used almost entirely since it has been found that upflow, even at very low velocities, causes some lifting and bouncing of the bed which pulverizes the desiccant particles. Dehydrated gas leaves the bottom of the bed to sales or NGL extraction. Flow continues through one bed for a finite period, perhaps 8-24 hours. The time is usually set by the capacity of the desiccant bed and the amount of water to be removed or the time required for regeneration and cooling.

rl

While one bed is "on line" drying, the second bed is being regenerated. A side stream, usually amounting to approximately 5-10% of the total gas stream, is heated to about 204-316°C [400-60O0F]. Direct or indirect (salt bath) heaters are the most common sources of heat. Hot regeneration gas from the heater enters the bottom of the bed being regenerated, passes up through it and out the top of the bed. As the desiccant heats up, water is driven off. Most of this water removal takes place at a temperature of about 1 10°C [230"F]. Hot regeneration gas from the tower is cooled to condense most of the water and then following separation, the stream is returned either to the main wet gas stream entering the dehydrator or, if feasible, to the sales gas stream. In the former case, regeneration gas flow can be created by holding a small amount of back pressure on the inlet side of the unit and using the control valve A€' to force the regeneration gas through the loop. Alternatively, a small in-line high speed centrifugal compressor can be installed in the regeneration loop. The latter requires higher maintenance but is more energy efficient.

At the end of the heating cycle, the temperature of the bed will typically be between 204-288°C [400-55O0F], depending on the desiccant and the service. This bed must be cooled before it can be placed back on-stream. In the simplest system, the heater is bypassed and the regeneration gas is passed through the hot tower. Cooling in this manner is continued until the outlet gas temperature falls to a temperature 14-17°C [25-30°F] above the inlet gas temperature. Once the tower is switched to dehydration, the bed cools rapidly to inlet gas temperature.

In a typical gas dehydration unit, adsorption flow will be down, regeneration heating flow will be up and cooling gas flow will be up. The primary reason for downflow during adsorption is to allow


r7

-


BASIC PROCESS FLOW

Fuel Gas

* To Liquid Disposal Tower 1

(Regeneration)

Temperature Cnntrniiar I

Process Regeneration Gas Gas Inlet - Regeneration *

I I I 1 I cqs Separator

1c To Liquid Dlsposai

7 - - - Regeneration Gas 4 O J nimai

Temperature Outlet Gas Filter

4 Regeneration

Gas Heater Regeneration Gas from

Outside Source

Figure 9.1 Typical Two-Tower Solid Desiccant Dehydration Units

operation at higher mass velocities without lifting or fluidizing the mol sieve bed. If the bed were fluidized, breakthrough would be very rapid, mol sieve attrition would be excessive, and the resulting dust would cause plugging in downstream process equipment.

Generally, desorption (regeneration heating) flow should be countercurrent to the direction of adsorption flow. If desorption is carried out concurrent to adsorption all of the water andor other contaminants must move through the entire bed requiring a longer regeneration time.

Cooling flow can be either concurrent or countercurrent to the adsorption flow. When cooling with dehydrated or treated gas, the flow is normally done in the same direction as the heating cycle (countercurrent to adsorption) for simplicity of piping. However, countercurrent cooling can be highly detrimental during the subsequent adsorption cycle if the cooling gas contains significant quantities of water and/or contaminants since these are adsorbed on the desiccant as the bed cools. This increases the residual loading in the adsorption cycle and results in a high water content or a high contaminant level in the effluent gas. If inlet or wet gas is used for regeneration and cooling, cooling should be concurrent with adsorption even though costs increase due to additional valves and more complex piping.

Three Tower Units In three bed systems, two beds are usually connected either in parallel or in series for adsorp-

tion, while the third bed is regenerated. In parallel, the cycle times are staggered so that the beds are



9.3


4

Knock- - , out Drum

3

Liquid Waste

T

L

Heater - Filter

I I

-I taken off stream for regeneration and returned for adsorption one at a time. In series, the first (lead) bed becomes completely saturated while the second (trim) bed removes the balance of the adsorbable component. When the first bed is taken off stream for regeneration, the trim bed is put in the lead spot with the regenerated bed in the trim spot downstream. The series arrangement is rarely used because of the higher pressure drops and the higher capital costs for the same capacity.

Figure 9.2 is a flow sheet for a 3-tower unit. At any one time it uses three towers, two in adsorption in parallel, with the third being regenerated and cooled. The shaded area inside towers 1 and 2 shows the progress of water adsorption in the bed or the portion of the bed which is essentially saturated with water. Below this area, the desiccant is capable of adsorbing more water, The bottom of this area represents the position of the adsorption front as it moves down through the bed with time.

The water front in bed 1 is lower than in bed 2 because it has been on stream longer. When the leading edge of this front reaches the outlet, bed 1 will be switched to regeneration and beds 2 and 3 will be on stream. Thus, at any one time, the two dehydrating towers possess different degrees of saturation. By the time bed 2 is ready for regeneration, bed 1 must be ready to go back on stream.

I

In Figure 9.2, the operating sequence of the towers on stream is:

1 and 2, 2 and 3, 3 and 1, 1 and 2, ad infinitum

In Figure 9.2, the dry residue (tail) gas is used for regeneration and cooling. This dry gas has more capacity for water than the inlet wet gas and will not partially saturate the bed during the cooling operation.

’ Panametncs Outlet Probe

Inlet Gas

NGL Recovery

Residue Gas

t

I I Cooling Bypass

Liquid Product

Figure 9.2 3-Bed Dry Desiccant Dehydration System

@John M. Cnmphli & Company


-

DESICCANT TYPES

The on-stream beds could be arranged in series, with the fresher bed of the two downstream of the partially saturated bed. This allows longer adsorption cycles; however, at a given flow rate, larger diameter towers would be required.

r"

Another variation on the 3-tower system is shown in Figure 9.3. In this system one bed is in adsorption, one in heating and one in cooling. This scheme is often used in Hydrocarbon Recovery Units (HRUs) when simultaneous adsorption of Cg+ hydrocarbons and water is required. The adsorption cycle in these facilities is often less than one hour. This does not leave sufficient time to completely heat and cool one bed during the regeneration cycle, hence the addition of the additional bed.

Figure 9.3 Schematic View of a Typical 3-Tower Plant Using Cooling and Heating in that Order

DESICCANT TYPES A commercial desiccant must possess an affinity for water, a large surface area per unit vol-

ume, high mechanical strength, resistance to abrasion, chemically inertness, and be reasonable in price. Three basic materials are used most commonly because they possess these characteristics in a satisfactory manner.

Activated Alumina Activated alumina is used to dry both gases and liquids. It is usually made by heating alumi-

num trihydrate (A1203.3 HzO) to remove most of the water. The structure of the product is amorphous

r"



9.5


rather than crystalline. Alcoa claims that activated aluminas can be used in natural gas dehydrators to obtain cryogenic dewpoints (-73°C [-100"F]). However, this claim has not been proven in the field. Activated alumina is rarely used in natural gas plants. Alumina has a higher equilibrium water capacity than does molecular sieve.

Silica Gel Silica gel is amorphous silicon dioxide (Si02). It is manufactured by adding aqueous sodium

silicate to sulfuric acid. The resulting chemical reaction produces a gel of silicic acid which is washed in alcohol and dried in an autoclave to form the hard glassy adsorbent.

Silica gel is used primarily as a desiccant. In general, it is less catalytic than either activated

1. Because it is amorphous, it will adsorb all molecules. This, it would have a reduced capacity for water if used to dry a gas saturated with heavier hydrocarbons.

2. Most dry silica gels shatter when contacted with liquid water. Special gels have been developed to protect the beds from entrained water in the feed. These are installed in a "buffer zone" at the top of the bed.

Because of its ability to adsorb several types of molecules, silica gel is often used for hydrocarbon dewpoint control on lean, high pressure natural gas streams. The silica gel will adsorb most of the Cg+ molecules as well as the water, effectively meeting the two dewpoint specifications simultane-

alumina or molecular sieves. However, it has two drawbacks.

ously.

Molecular Sieve Molecular sieves are manufactured in two crystal types, a simple cubic or Type A crystal and a

body-centered cubic or Type X crystal. The Type A sieve is available in sodium, calcium, and potassium forms. The Type X sieve is available in sodium and calcium forms. The sodium forms of the sieves are the most common and are shown below in oxide formulas.

Type 4A: Na20-A1203.2 SiQ.YH20

Type 13X: Na20.Al203.2.5 Si02.YH20

The value of Y depends on the extent of activation.

A sieve is manufactured by crystallizing the proper crystal type in sodium form from a solution of sodium silicate, aluminum trihydrate, and sodium hydroxide. If the sodium form is not the desired product, then either calcium or potassium ions are substituted for the sodium ions by soaking the crystals in a solution of the appropriate chloride salt. The ion exchange is never complete. The small crystals (1-4 microns in size) are then mixed with a clay binder and either extruded into cylindrical pellets or rolled into spherical beads. The sieve is finally activated (calcined) by heat at 650°C [ 1 200"FI to remove most of the hydrated water.

Table 9.1 contains data on the standard molecular sieves. Openings or ports into the crystal cavities are 3, 4, 5 , and 10 Angstroms (1 A = 10-lo meter). These dimensions are of the same order of magnitude as the diameters of small molecules. Type 3A is not as stable as Type 4A. It is used in cracked gas dehydrators to dry light olefin streams, e.g., ethylene, propylene, etc. Rarely is a sieve other than Type 4A, 5A or Type 13X needed in a natural gas plant. Type 4A is used for dehydration of gases and liquids. Type 5A is frequently used for simultaneous H20 and H2S removal from natural gas. Type 13X is used for NGL product treating.



r7

DESICCANT TYPES

Type 3A

Type 4A

TABLE 9.1 Molecular Sieves

Opening Metal Oxide Use

3A 0.6 K20, 0.4 Na20

4A Na20 Dry gases and liquids; remove H2S.

Dry olefins, dry methanol and ethanol.

r"

r"

Type 5A 1 5A 1 0.8 CaO, 0.2 Na20 Separate normal paraffins from branched- chain and cyclic hydrocarbons; remove H2S.

Type 13X 10A Na20 Remove mercaptans and H2S; remove H20 and C02 from air plant feed.

Davison Type Z- 1 O0 0.8 CaO, 0.2 Na20

Binderless Sieves for sharp separations.

Table 9.1 also gives data on special sieves. Many of these sieves are manufactured solely for the purpose of drying without forming significant amounts of carbonyl sulfide, which is difficult to remove from both gas and liquid streams. Others are designed to be resistant to low pH (acidic) streams which can occur when dehydrating high H2S/C02 content gases.

Molecular diameters of some common molecules are listed in Table 9.2. This data can be used to predict whether or not a particular molecule will be adsorbed. For example, the 5A opening into the crystals of the Type 5A sieve is large enough to permit entry of straight-chain hydrocarbons (n-butane to n-C22H46) but small enough to exclude branched-chain hydrocarbons (i-butane to i-C22H46). One use of Type 5A sieve is to separate a mixture of straight and branched chain hydrocarbons.

The size of openings into crystals given in Table 9.1 and the diameters of molecules given in Table 2 should be used with caution. Molecules and crystals are both somewhat elastic and the dimensions are slightly temperature dependent. For example, water is adsorbed by Type 3A sieve and methanol by Type 4A sieve.

This can lead to an erroneous prediction.

A comparison of the general characteristics/properties of commercial desiccants is shown in Table 9.3. Sorbead and Alumina Gel are trade names for silica gel which serves as a generic name for this general class of desiccants. Table 9.3 will be referred to in later calculations.

Davison Type SZ-5 UOP Type FK-33

UOP Type RK-29

Davison Type SX-9

UOP Type AW-300 UOP Type AW-500 Davison Type 700

@John DI. Campbell & Company

5A 5A >0.8 CaO, <0.2 Na20 minimal COS formation 5A

>0.8 CaO, <0.2 Na20

>0.8 CaO, <0.2 Na20

>0.8 CaO, <0.2 Na20

Remove H 2 0 and H2S from gas with

Remove H20, H2S and light mercaptans from liquid with minimal COS formation.

Remove H20, H2S and mercaptans from liquid with minimal COS formation.

4A Mostly Na20 Dry and purifi acidic streams containing 5A Mostly CaO compounds such as HCl, SO2, and NOx. 5A Mostly CaO (Naturally occurring zeolites, 4A is

8A

mordenite. 5A is chabazite mined in New Mexico.)


9.7


Helium Hydrogen Acetylene Carbon Monoxide Carbon Dioxide Nitrogen Water Ammonia Hydrogen Sulfide

Methane Ethylene Ethylene Oxide Ethane Methanol Ethanol Methyl Mercaptan Propane

~ Argon

TABLE 9.2 Molecular Diameters

~

Sorbead R Sarbead H

Specific Area, m2/g [IO6 ft2/lbm] 750 [3.66] 750 [3.66]

rs)

H-151 F-1 4A-5A Mal Sieve Aluminal Gel Alumina

400 [1.95] 250 [ 1.221 600-800 [2.93-3.911

Molecule

Pore Diameter, A 21 28

Pore Volume, ccig [ft2/lbm] 0.35 [0.042] 0.49 [0.059]

Avg Partical Diameter, mm [in.] 3.5 [0.145] 3.5 [0.145]

Bulk Density, kg/m3 [ibmift3] 785 [49] 721 [45]

Heat Capacity, k.J/kg.OC [Btu/lbm-"F] 1.05 [0.25] 1.05 [0.25]

Relative Saturation (@ 60% RH, % 33 36

43 26 4-5

0.35 [0.042] 0.21 [0.025] 0.27 [0.032]

6.4 [0.25] 3-12 [0.125-0.51 1.6-3 [0.06-0.125]

833 [52] 833 [52] 689-721 [43-45]

- 1.0 [0.24] 0.96 [0.23]

21 14 22

Diameter (4 2.0 2.4 2.4 2.8 2.8 3.0 3.2 3.6 3.6 3.8 4.0 4.2 4.2 4.4 4.4 4.4 4.5 4.9 4.9 5.0 5.1

~~

M o I e c u 1 e Butene- 1 Butene-2 Trans 1,3-Butadiene Chlorodiflouromethane (R-22) Thiophene i-Butane to i-C22&,

Dichlorodiflouromethane (R- 12) Cyclohexane Benzene Toluene p-Xylene Carbon Tetrachloride Chloroform Neopentane m-Xylene o-Xylene Triethylamine

TABLE 9.3 Properties of Various Adsorbents

Diameter (4 5.1 5.1 5.2 5.3 5.3 5.6 5.7 6.1 6.7 6.7 6.7 6.9 6.9 6.9 7.1 7.4 8.4

@John M. Campbell 8s Company


CHOICE OF DESICCANT

Desiccant Activated Aluminas

P

Max. Inlet Min. Outlet Dewpoint Temperature (Approximate) 52°C [125"F] -68°C [-90"FI

r"

The choice 1. 2. 3. 4. 5 . 6. 7.

CHOICE OF DESICCANT of desiccant depends on the following major factors: Inlet conditions Molecular size/gas composition Required outlet water dewpoint Possible chemical reactions Is simultaneous hydrocarbon recovery also desired? Co-adsorption Cost of desiccant

The aluminas have the lowest cost per unit of dehydration capacity. The silica gels are next. Molecular sieves are the most expensive and must be justified by their special characteristics.

A comparison of the molecular diameters in Table 9.2, with the pore diameters in Table 9.3, shows that all substances listed can be adsorbed into alumina and silica gel. Not so for 4A and 5A molecular sieves. Thus, molecular sieves have a degree of selectivity for the molecules they can adsorb.

The table below offers a general comparison of the common desiccants for dewpoint and inlet temperature considerations.

11 Silica Gel I 49°C [120"F] I -51°C [-60"FI I

These are guideline numbers only, The capacity of all desiccants decreases with increasing temperature. The above are the inlet temperatures at which a given desiccant normally is not economical. The dewpoints shown are the minimum values normally achieved with a properly designed and operated unit.

In actual practice, it is not always necessary to dry a stream to obtain a dewpoint below the coldest temperature in the system. Below -62°C [-80"F] ice will form very slowly and may not be a serious problem. Water content correlations in the hydrate region show that there is not a significant amount of water left in the gas to freeze below this temperature. Some expander plant operators have used silica gel or alumina for bulk dehydration and have successfully inhibited the trace residual water with methanol.

Small amounts of sulfur compounds may be tolerated by all three desiccants, however, aluminas and silica gel are generally not effective in removing these components. Hydrogen sulfide and carbon dioxide can be removed by molecular sieves. Commercial units are built for this specific purpose.

The aluminas possess little capacity for hydrocarbons. The silica gels are very effective for recovery of the C5+. Molecular sieves will recover any hydrocarbon molecule small enough to enter its crystalline structure which for a 4A sieve would be only methane and ethane in the paraffin series. F"

Table 9.4 is an application guide for molecular sieves and silica gel.



9.9


Application Hydrogen Rich Gas Drying Reformer Recycle Gas Drying Hydrocracker Gas Drying Inert Gas Ammonia Removal Low pressure Air Drying

Cracked Gas Dring (Olefin Drying)

Recycle Gas Sweetening

Air Plant Feed Drying and C02 Removal Natural Gas Drying Natural Gas HIS Removal

Natural Gas Mercaptan Removal Natural Gas COZ Removal

LNG Plant Feed Pretreatment (HzO, COZ,

TABLE 9.4

Molecular Sieve Remarks 4A 4.4 4A 4A 4A

3A

Effluent dewpoints below -75°C [-100"F] Effluent dewpoints below -75°C [-100"F] Effluent dewpoints below -75°C [-100"F] Effluent ammonia concentrations as low as 1 ppm. Very effective, since the large heat release which effects the performance of other desiccants does not reduce the effectiveness of molecular sieve systems. A 3A sieve is required to prevent entrance of olefins inside the micropore structure. Because of the high olefin and diolefin content, a careful regeneration procedure is used. Beds are stripped at room temperature with a purge gas to remnve olefins from macropores prior to the heating period. Eftiuent dewpoints below -75°C [-IOO"F]. Removal of HzS to below 4 ppmv [U4 gn'scfl. Used in conjunction with gas drying. High capacity for water and COZ because of negligible competition from Nz or 0 2 .

Effluent dewpoints as low as -IOO''C [-15O"F] can be otained. Requires high regeneration temperatures (260°C [50OoF]) to assure adequate HzS desorption. Large pore sieve required because of mercaptans size and rate considerations. 4A sieve if propane concentration is high in order to eliminate competition between propane and COZ. If the propane concentration is low, a 5A sieve provides faster rates and higher capacities. Choice of system would depend on specifics. A typical system could be a main 4 or SA

5A

13X 4A

4A-5A

13X 4A-5A

4A, 5A, 13X

MOLECULAR SIEVE APPLICATION GUIDE

H2S, mercaptan removal) 1 bed for H20, COZ and HIS removal. A 13X trimmer bed for mercaptan removal. Hydrocarbon Separation 5A I Separates straight chain hydrocarbons from is0 compounds. Systems may be designed for

Liquid (Jet Fuel, Kerosene, Akylation Feed, Hexane, Benzene, Cyclohexane. Etylene, Propylene, Methanol, Ethanol,

both gas and liquid phases. Liquids can be dried to water contents below I ppm. If the liquid contains polymenzable olefinic compounds, a 3A sieve is required. 3A sieves also used to dry methanol.

3A, 4A


LPG Sweetening (H2S and Mercaptan

-l 13X 1 Large pore sieve required to assure mercaptan removal and reasonable adsorption rates.


industry rlatural Gas iefining

~heinmical/Petrochemical

industrial Gases

Refrigeration

Insulating Glas

Commercial and Miilitary Packaging Telecommunications

Application Low dewpint drying, hydrocarbon dewpint control, hydrocarbon recovery of gasoline fractions. Drying of Reformer recycle gas, Kerosene jet fuels, fractionated hydrocarbons, hydrogen, inert atmospheres, and a wide variety of other hydrocarbon liquids and gases. Low dewpoint drying of innumerable gases and liquids such as Hz, Nz, CCh, HIS, HCI, SOX, Refrigerants, Ethylene, Benzene, etc. in addition, purification of gases and liquids are easily accomplished. A few examples are: Removal of Benzene, Toluene and other aromatics from paraffinic streams, purification of monomer feed streams, adsorption and recovery of misc. chemicals such as SOL chloroform, acetic acid, methanol and various solvents. Drying of inlet air, and hydrocarbon removal from the rich liquid in LOX plants. In addition, silica gel is used in diying product gases (Hz, N2. Argon). Because of its high water and acid capacity silica gel is used for refrigerant dehydrators in automotive and commercial refrigeration systems. Used alone or in combination with Davison molecular sieves, silica gel assures maximum protection against water and hydrocarbon dewpoints under the widest range of climatic conditions. Numerous package configurations using silica gel are used for protection against moisture, mildew and odor in pharmaceuticals, foods, packaging of metallic producis, delicate instruments and equipment. Underground Telephone cables are protected by drying equipment containing Silica Gel. In addition, above ground cable splices utilize silica gel in the splice to eliminate cable failure.

THE NATURE OF ADSORPTION

THE NATURE OF ADSORPTION Figure 9.4 illustrates the basic behavior of an adsorbent bed in gas dehydration service. During

normal operation in the drying (adsorbing) cycle, three separate zones exist in the bed:

1. equilibrium zone,

2. mass transfer zone (MTZ), and

3. active zone.

In the equilibrium zone the desiccant is saturated with water. It has reached its equilibrium water capacity based on inlet gas conditions and has no further capacity to adsorb water.

Virtually all of the mass transfer takes place in the MTZ. A concentration gradient exists across the MTZ. This is illustrated in Figure - 9 3 b ) for various times throughout the cycle. Curves 1-3 show the formation of the MTZ; curve 4 reflects the concentration gradient for the MTZ position in Figure 9.4. Curve 6 shows the concentration gradient at breakthrough. Notice the adsor- bate bed saturation is 0% at the leading edge of the MTZ and 100% at the trailing edge. 14

The third zone is the active zone. In the active zone the desiccant has its full capacity for water and contains only that amount of residual water left from the regeneration cycle.

When the leading edge of the MTZ reaches the end of the bed, breakthrough occurs. If the adsorption process is allowed to continue, the water content of the outlet gas will increase

Feed

Equilibrium Zone

Mass Transfer

Zone

Active Zone

* Product

following the traditional "S" curve. Breakthrough Figure 9.4 curves are illustrated in Figure 9.5(c) for three MTZ lengths.

Part Way Through Adsorption Cycle

Figure 9.5(d) shows the location of MTZs in multicomponent adsorption typical of hydrocarbon and water adsorption on silica gel. As the gas enters a dry desiccant bed, all of the adsorbable components are adsorbed at different rates. After the process has proceeded for a very short period of time, a series of adsorption zones will appear. These zones represent the length of tower involved in the adsorption of any component. Behind the zone all of that component entering has been adsorbed on the bed. Ahead of the zone, the concentration of that compound is zero (unless some is leR from a previous adsorption or regeneration). These zones form and move down through the desiccant bed. Water would be the last zone formed. On all materials except carbon it will displace the hydrocarbons if enough time is allowed to do so. If molecular sieve is used, adsorption of the C&6+ fractions will not occur because the bulk of these molecules cannot fit in the desiccant structure.

rF



9.1 1

T


Equilibrium Zone Mass

Transfer Zone

Active Zone

(a)

Feed

Product

I Bed Length

" o Bed Length h,

(b) Variation of Adsorption Zone Front with Time

(c) Character of the Breakthrough Curve

Time I (d) Schematic View of Bed Saturations (e)

(1) 14-20Mesh (2) 8-10Mesh (3) 5-6Mesh (4) %Mesh

~~

Figure 9.5 Representation of the Adsorption Zone Concept

100% of any component is adsorbed on the desiccant until the fiont of its zone reaches the outlet of the bed. When the back of its zone reaches the outlet of the bed, no more adsorption of that component will occur. It will furthermore be displaced almost entirely by the component in the zone following it down the bed if the cycle is continued. If the process continues long enough, no effective amount will remain on the bed.

With silica gel, at commercial flow rates and tower configurations usually employed, pentane will have a breakthrough time of from 12-20 minutes. Methane and ethane break out almost instanta- neously. If the process cycle proceeds beyond 30-40 minutes, all but the heaviest hydrocarbons will have been displaced out of the bed. From this time on, primarily dehydration is taking place. Thus, the performance of a given unit is dependent on the cycle length used.

For very short cycles both hydrocarbon adsorption and dehydration occur. The hydrocarbon recovery will substantially reduce dehydration costs charged to the unit or eliminate them. Such units may be used to simultaneously control water and hydrocarbon dewpoints at minimum cost. Their potential application is far greater than their use to date would indicate.

Figure 9.5(e) shows the effect of desiccant size on the length of the zone. The steeper the zone, the sharper the separation, the better the process. Therefore, the desiccant used should always be the smallest compatible with the pressure drop limitations. The smallest size will seldom be greater than 12 mesh (Tyler Screen Scale), in most commercial natural gas installations. This is equivalent to an average particle size of 1.6 mm [ 1/16 in.].

Other factors which affect the length of the MTZ include gas velocity, contaminants, water Increasing gas velocity increases MTZ length. content and relative saturation of the inlet gas.



ADSORBER DESIGN

rc" Contaminants are particularly insidious because they can slow the mass transfer process (lengthen MTZ) by providing additional resistance.

The length of the MTZ has a significant effect on the useful capacity of the desiccant since the MTZ is left in the bed at the end of the adsorption cycle. Remember, the desiccant in the MTZ is only partially saturated with water!

ADSORBER DESIGN

The design of fixed bed adsorption systems is complex. The desiccant manufacturers will typically perform these calculation with their own proprietary software. They are very reluctant to share details of the design with the client. Unlike many other processes, these calculations cannot be done a process simulator. This text will concentrate on the design techniques and the calculation

r"

r"

principles. studies.

1.

2.

Where:

The manual calculations presented here can be used for preliminary designs and feasibility

The first step in planning and specifying a dry desiccant unit is to establish the number of adsorbers and the cycle time. For small units (< 1.5 x 1 O6 m3/d [50 MMscfd]) a two tower system is generally a good assumption. For larger systems a 3 or 4 tower system may be more economic. Increasing the number of adsorbers allows for better bed geometry (tall thin beds) and increases operating flexibility. The drawback is higher capital costs and reduced time for regeneration since the regeneration time, t , is equal to the adsorption time, t, divided by n - 1 where n = number of beds.

t,=- ta n-1

Shorter regeneration times increases the regeneration gas rate and the size of the regeneration equipment.

Once the number of beds has been decided, the cycle time will determine the mass of desiccant per tower. If the feed gas is saturated with water a good assumption for cycle time is 8-16 hours. If the feed gas is undersaturated, a cycle time of 24-30 hours may be feasible. There are a limitless number of beds and cycle times which result in a technically feasible design, the goal is to find the one which minimizes capital and operating costs.

For the unit configuration established in 1), the second step is to size the adsorbers. The diameter of the adsorbers is set by gas velocity and allowable pressure drop. Several methods of determining the bed diameter are presented in the literature. Equation 9.2, below is based on the Ergun equation and is also presented in the GPSA Engineering Data Book.

SI Units AP/L = pressure dropllength

p = gas viscosity pg = gas density vg = superficial gas velocity

C P

kg/m3 d m i n

FPS Units psilft

C P lbm/ft3 Wmin

@John M. CampbeU 8s Company


9.13


nits C

0.00 135 0.00 189 0.00207 0.00319

FPS Units B C

0.0560 0.0000889 0.0722 0.000124 0.152 0.000 1 3 6 0.238 0.000210

Particle Type l/8" bead 1/8" extrudate 1/16" bead 1/16" extrudate

Most designs are based on a APIL of about 7-10 kpdm [0.3-0.44 psi/&].

When equation (2) is written in terms of velocity it becomes

SI 1 B

4.16 5.36

11.3 17.7

-Bp+[(Bp)2+4Cpg(F]]o.5 v =

SI Units m3 (std)/d

2 CP,

FPS Units scff d

(9.3)

For preliminary calculations, the following equation gives a useful approximation of gas velocity.

Where: vg =

PP - A =

-

(9.4) A vg = - Jpg

SI Units FPS Units d m i n fümin

l b d d superficial gas velocity gas density constant 1/8" spheres

1/16" spheres 48

will be For most moderate to high pressure designs using 1/8" desiccant the superficial gas velocity 9-12 ndmin [30-40 ft/min].

Calculation of tower diameter follows:

Where: d = tower diameter q = actual gas flowrate

vg = superficial gas velocity

Where:

SI Units FPS Units

m3/min ft3/min

The actual flowrate, q, can be calculated from Equation 9.6

= *(+j(+jZ

q, = standard gas flow P, = standard pressure Wa P = actual pressure kPa T = actual temperature K

T, = standard temperature K z = gas compressibility factor at T and P -


psia psia OR

OR

'1


ADSORBER DESIGN

The next step is to calculate the total amount of adsorbent required for the water loading per adsorption cycle. Once this is known the bed height may be calculated from the bed diameter and bulk density of the desiccant

r"

Establishing the "effective" or "useful" capacity of the desiccant is where science ends and black art begins. The static equilibrium capacity of most molecular sieves is about 21-22%. This can be seen from the isotherms for 4A sieve in Figure 9.6. The dynamic equilibrium loading will generally be 50-70% of the static equilibrium value from the isotherm. For molecular sieves, 15% is a reasonable value for dynamic equilibrium loading.

i"

20

16

12

a

4

C 0.001 0.01 0.1 1 .o

H,O Partial Pressure, rnm Hg 10

Figure 9.6 H20-4A Isotherms - Intermediate H20 Pressure

The effective capacity will be less than the equilibrium capacity because, at breakthrough, the MTZ must remain in the bed (see Figure 9.4). The desiccant in the MTZ is only partially loaded with water (Figure 9.5(b)) so on an aggregate basis the bed will not be loaded to dynamic equilibrium at breakthrough. If the length of the MTZ was known, one could determine the bed height by adding the MTZ height to the height of the equilibrium zone. Unfortunately, calculation of the MTZ height is very complex and is typically left to the vendor.

It should be obvious, however; that the height of the MTZ, relative to do the total bed height, has a significant effect on the useful bed capacity. The longer the MTZ the lower the effective desiccant capacity.

A second factor which impacts useful capacity is age and deterioration of the desiccant. Figure 9.7 shows a typical capacity decline curve for silica gel. The rate of capacity loss depends on several factors including gas velocity, number of regeneration cycles, presence of contaminants, etc.



9.1 5


- - - Desiccant Type Loading Mass WaterMass Desiccant

Alumina 4-7% Silica Gel 7-9% Mol Sieve 9- 12%

I

SI Units m hB = bed height

X = useful desiccant capacity, %

d = bed diameter

m, = water loading/cycle kg

pB = bulk density of desiccant

-

kg/m3 m

Months in Service

Silica Gel Capacity as a Function of Time in Service Figure 9.7

FPS Units fi

lbm -

~ f i 3 ft

Where:

The bulk density, PB, can be found in Table 9.3. Values for useful capacity, X, can be found in the table above.

O Jobn M. Campbell & Company


REGENERATION DESIGN

150 rnm [6 in]

75 mm [3 in]

75 rnm [3 in]

I Figure 9.8 Adsorber with Floating Screen and Balls for Im-

I"

The actual tower height (SS,TT) will be the bed height plus the height of bed supports and sufficient space to insure good flow distribution at the top of the bed. This additional height is typically 1- 1.5 m [3.3-5 ft]. A typical adsorption tower is shown in Figure 9.8.

Before proceeding to the regeneration calculations a quick reality check is in order. The desired bed length to bed diameter ratio (hB/d) should fall between about 2.5-6. A value less than 2.5 can result in lower useful desiccant capacity due to the relatively large MTZ/hB ratio. A value greater than 6 can result in an excessive AP. The total AP across an adsorbent tower should not exceed 55-70 kPa [8-10 psi].

If the bed is too short, the cycle time or number of beds should be increased. I f the bed is too long the opposite is true.

Once a viable design has been established, the regeneration calculations can

be made. proved G a s Distribution

REGENERATION DESIGN Once the adsorber towers have been sized, the next step is to determine the amount of regen-

eration gas needed as well as the heating and cooling loads.

The heating load is most critical because this is the primary operating cost unless waste heat is available. Heating must accomplish all of the following:

1. Heat the desiccant to at least 204-288°C [400-55O0F] (depends on desiccant).

2. Heat and then vaporize the adsorbed water.

3. Heat and then vaporize any hydrocarbons on the bed.

4. Heat the vessel shell and steel internals (if not insulated internally).

5 . Heat the valves and piping in the line between the regen heater and the towers.

6 . Supply heat lost through the insulation.

If the desiccant is a 4A molecular sieve, it is safe to assume that at the end of the heating cycle the bed temperature will be 260-288°C [500-55O0F] with 3 15-343°C [600-65O0F] entering gas; with gels and alumina these temperatures may be 204°C [400°F] and 260°C [500°F], respectively. The actual regeneration temperature should be the minimum required to adequately regenerate the bed.



9.17


O

-? Figure 9.9 shows a regeneration temperature profile for a silica gel system. The temperature, TH, is the hot regeneration gas into the bed. The temperature profile TI - T4 is the outlet gas temperature leaving the bed. In this case, when the bed outlet temperature (T4) reaches approximately 176°C [350°F], the heating cycle is finished and the cooling cycle begins. The temperature profile T4 - T5 shows the bed outlet temperature during the cooling cycle.

- Start of curve 3 End of Cycle Cycle

I

The entire regeneration cycle can be divided into four (4) specific time intervals. Interval A (QA) is virtually all sensible heat. It represents the time required to heat the bed, steel and adsorbed water from TI to T2. At T2, water begins to boil off the desiccant. For commercial desiccants T2 is about 1 10°C [230°F] and T3 is about 127°C [260"F].

Interval B (QB) is where the most of the water is driven from the bed. This requires sufficient heat to not only revaporize the water but also to break the attractive forces which bind the water to the surface of the adsorbent. This is often called the heat of wetting. The sum of the latent heat and the heat of wetting is the heat of desorption. This value is approximately 4200 kJkg [1800 Btdlbm] for sieves and 3260 kJ/kg [1400 Btdlbm] for alumina and gels. The temperature profile for QB is T2 to T3 *

Once the bulk of the water has been driven from the bed, Interval C (Qc) represents the time required to remove heavy contaminants and residual water. This is sometime referred to as driving the "boot" off the bed. The temperature profile for Qc is from T3 to T4.

When the bed reaches T4, cooling gas is introduced to the bed. The cooling gas is usually dry processed gas. Interval D (QD) represents the cooling cycle. This is all sensible heat and the temperature profile is from T4 to Tg. The cooling cycle is typically ended when T5 - TI = 14-17°C [25-3O0F].

@John M. Campbell & Company ~~


REGENERATION DESIGN

The regeneration gas flow rate is established by heat balance. The regeneration gas rate must be adequate to deliver the required heat input in the time available. Likewise it must also be sufficient to deliver the cooling in the time available. In equation form, this heat balance is summarized below.

r"

Qc = mCpATcOc (9.8~)

SI Units

Where: Q = m = c, =

AT =

e =

heat load in a time interval kJ mass flow of regeneration gas heat capacity of regeneration gas effective temperature difference, e.g. (TH - TB) for

kgfh kJ/kg."C

"C a time interval length of time interval h

Equations 9.8(a-d) can be solved for m subject to the constraint

FPS Units BtU

lbmh Btu/lbm-"F

"F

hr

e A + OB + Oc + < time available for regeneration (9.9)

For a two tower system, the time available for regeneration is equal to the adsorption time. For the three tower system shown in Figure 9.2, the time available is one-half the adsorption time. For the three tower system in Figure 9.3, the time available is twice the adsorption time.

Equations 9.8 and 9.9 can only be solved by trial and error and are not suitable for manual One problem is the inability to easily determine the effective temperature differences calculations.

ATA, ATB , etc.

An alternative is to calculate the total heating load, QH, (QH = QA + QB + Qc) and cooling load, QD. An overall AT for the heating and cooling cycles can then be estimated by log mean temperature difference equation.


heating (9.1 O)

cooling (9.11)


9.19


The calculation of the regeneration rate then follows:

(9.12)

To simplify the calculation further, many references, including the GPSA Engineering Data Book, suggest that QH be multiplied by 2.5 and ATH be taken as TH - TI. This eliminates the calculation of a mean temperature difference and gives results surprisingly close to the more complex methods. This is summarized below in Equation 9.13.

(9.13)

The calculation of the regeneration gas rate is still iterative since + 0~ < time available. For first guess, the following guidelines can be used:

a. Assume m = 10% of the process gas rate

b. Allocate 60-70% of the total regeneration time to heating and 30-40% to cooling

C. OH should not be less than 1 hour or greater than 8 hours

d. The minimum heating gas velocity should exceed that which gives a AP = 0.23 kPdm [0.01 psi/ft] to insure good flow distribution and water removal. This velocity can be calculated from Equation 9.2, or estimated from Fig 20-47 in the GPSA Engineering Data Book.

The heat loads QH and Qc can be calculated by heat balance. QH is the sum of the sensible heat required to heat the steel, desiccant, support balls, valves, piping, and water’ to T4 plus the heat of desorption.

.i

Can Type Cast Type

Figure 9.10 Types of Internal Insulation

1 Water is heated only to an average temperature of 116°C (240”Fl where the bulk of it is desorbed.



REGENERATION DESIGN

The heat required for the steel shell will depend on whether internal or external insulation is used. Internal insulation is of two types: (1) a steel "can" inside the shell that provides a stagnant gas space between the bed and shell or (2) cast or sprayed internal insulation. These are illustrated in Figure 9.10. With internal insulation the bed diameter usually is about 15 cm [6 in] less than the shell I.D..

Internal insulation is usually required on towers when the regeneration time is limited. It may be desirable on longer cycle units to save on fuel costs. For internal insulation the shell sensible heat load is about 25% that of an externally insulated tower.

The approximate wall thickness of adsorber is shown in Figure 9.11. This is based on ASME Section VI11 Div. 1 and A-516 Gr 70 plate steel. The diameter should be the nearest commercial size available that will allow a desiccant bed diameter at least as large as that calculated. The mass of the shell and heads may be estimated from Equation 9.14.

m, = A h d t (9.14)

SI Units Where: m, = mass kg

h = vessel length (HS-HS) m d = vessel I.D. mm t = shell thickness mm

A = weight factor 0.0347 HS-HS = head seam to head seam

FPS Units lbm ft in. in. 15

The following values are suitable for the heat balance calculation:

Heat Capacity, C, Steel Liquid Water Desiccant see Table 9.3

H20 on sieves H20 on gel or alumina Hydrocarbons 465 W/kg [200 Btu/lbm]

0.50 H/kg . C [0.12 BtuAbm-OF] 4.19 H/kg . C [1.00 BtuAbm-OF]

Heat of Desorption 4187 W/kg [1800 Btu/lbm] 3256 kJ/kg [1400 Btu/lbm]

Calculation of QH

QH = (mass of steel) (C,) (T4 - TI) + (mass of desiccant) (C,) (T4 - TI) (9.15)

+ (mass of water) (Cp) (TB - TI) + (mass of water) (heat of desorption)

+ heat losses and misc heat load (piping, valves, support balls, etc.)

The last item in Equation 9.15 is often estimated to be 25-30% of the calculated value for the steel, desiccant and water.

Calculation of Qo QD = (mass of desiccant) (Cp) (T4 - T5) + (mass of steel) (Cp) (T4 - T5) (9.16)



9.21

[r I


Vessel I. D., in I Figure 9.1 1 Approximate Pressure Vessel Wall Thickness



REGENERATION DESIGN

SI Units QRH = regeneration heater duty kW

m = regeneration gas mass flow kg/s C, = average regen gas heat capacity kT/kg.oc TH = heater outlet temperature "C TI = heater inlet temperature "C

/". Regeneration Gas Heater

Where: FPS Units

BWhr l b m h

Btu/lbm-"F OF "F

Regeneration Gas Cooler One needs to calculate the cooler load for all three intervals to find the highest load. It will

normally occur in interval B for long cycle units. The latent heats of water and hydrocarbons must be known. Knowing the time for the interval and assuming the desorption is uniform during it, one can find the latent heat load. To this one must add the gas sensible heat load. The normal temperature approach will be 16-20°C [29-38"F] for air cooling and 8-10°C [15-18"F] for water cooling.

Example 9.1: Perform a preliminary design calculation on a molecular sieve dehydration system processing the gas shown below:

Feed Gas - flow = 2.0 x lo6 std m3/d [71 MMscfd]

T = 30°C [86"F] P = 6000kF'a [870psia] z = 0.86 y = 0.65 p = 0 . 0 1 4 ~ ~

z = 1.00

y = 0.59 P = 6000 H a

Regen Gas -

C, = 2.3 kJ/kg."C [ O S 5 Btullbm-OF]

TH 1 310°C [590"F] T4 = 290°C [554"F]

Assume a 3 bed system with 2 beds in parallel and a 12-hour adsorption cycle. Step 1 : Calculate the superficial gas velocity, vg

('1 (MW) - (6000) (28*97) = 52.1 kgm3 = ZRT (0.86) (8.314) (303)

m - 9.3- v 67

g - rn min

Step 2: Calculate the bed diameter, d

d = dz= d s = 1.2m



9.23


xample 9.1 (Cont’d):

Step 3: Calculate the bed height for a useful loading, X, of 10%. The water loading per cycle, Mw is:

H20 1 x lo6 std m3 360 kg ( d I(%)[ 1i:;z3] = cycle

= 4.5 m (400) (360) (TC) (10) (705) (i.2)2

hB =

length to diameter ratio = - 4’5 - - 3.8 OK! 1.2

Check AP -

-- AP - B p vg + C pg vg = (4.16) (0.014) (9.3) + (0.001 35) (52.1) (9.3)2 L

= 6.6 kPa / m

Total AP = (6.6kPa/m) (4.5m) = 30kPa OK!

Step 4: Regeneration Calculations -

Calculate mass of steel in tower - 1.2 m dia x 6.0 m S-S

For a design pressure of 7000 Wa, t = 28 mm (use 30)

= (0.0347) (6.0) (1200) (30) = 7495 kg

Calculate the heating load, QH

MJ Steel Desiccant Water (sensible)

(7495) (0.50) (290 - 30) (3600) (0.96) (290 - 30) (360) (4.19) (116 - 30)

Water (desorption) (360) (4187)

= 974 = 899

130

= 1507

- -

3510

Add 25% to cover piping, valves, etc. + losses = 878 QH= 3510 + 878 = 4388 MJ use 4400 MJ

The available regeneration time is 6 hours. Assume 4 hours for heating and 2 hours fo1 cooling.

2.5 QH - - ( 2*5) (4 400 ‘Oo) = 4270 kg/h m =

‘’ = [ h ’ )[ (0.59) (28.97) kg

C, 0~ (TH - TI) (2.3) (4) (310 - 30)

4270 k 1 kmol

= 142 000m3/d

= 7% of process rate



REGENERAION DESIGN

Example 9.1 (Cont'd):

Check cooling load, QD

Steel Desiccant

MJ (5621) (0.5) (290 - 45) = 689

(3600) (0.96) (290 - 45) = 47 Qc = 1536

Estimate ATc from ATm

290

260- 15 = 86 "C

30 30 0

for a regeneration flow rate of 3979 kg/h

Q = mCp AT& = (3979) (2.3) (86) (2) = 1 574 O00 kJ OK!

Check to see if regeneration rate meets minimum velocity criteria using Equation 9.3

AP kPa L m for - = 0.23 -, v, = 1.8 d m i n

ford = 1.2m A = 1.3m2,soq = Av, = (1.13)(1.8) = 2.04m3/min

9s = (1440)(2.04) [ ~ ;:::)[ - :::)( - l lo) = 116000stdm3/d OK!

Size regeneration gas heater

= [ 397; kg )( 1 h 1 2.3 kJ (310- 3O)"C ) = 712kW

3600s kg."C



9.25


MECHANICAL DESIGN CONSIDERATIONS Once a unit configuration suitable to your conditions has been selected and the vendor has

furnished a process design, mechanical design can proceed. While a lot of work has been done, it is still necessary to design the vessels including distribution devices, associated valving, the regeneration gas heater, and the necessary piping layout.

1. Vessel Design. As discussed in an earlier section, the vessel sizing is largely set once superficial gas velocity and mass of desiccant required have been determined. To this one must add the space required for support members, screens, inert balls, gas distribution devices, and velocity reduction. Typical spacing for these items is shown in Figure 9.8. Bed weights are substantial and, coupled with the Al' forces incurred during adsorption, adequate support must be provided. Additional length of vessel shell to ensure adequate room for these system elements and good flow distribution can significantly improve operation. Remember that the mol sieve must be loaded and unloaded several times during project life, so provide adequate openings for these operations. Traditionally, wire mesh and grating assemblies have been used to support desiccant beds. These assemblies have served well with the larger desiccant particles. To minimize some of the problems in the sealing and maintenance of the wire mesh and gratings, Johnson screens (UOP) have been used, particularly for the finer mol sieve grades. A 0.84 mm r0.033 in] slot opening will retain the common 4-8 mesh and 8-12 mesh mol sieve materials. The annular space between the vessel wall and the edge of the bed support screen must be sealed to prevent loss of mol sieve. Asbestos rope packing forced into this space will prevent sieve loss. Internal or external insulation of the dehydration vessels may be selected. Internal insulation reduces the regeneration energy requirements somewhat. Small cracks in internal insulation can contribute to bypassing gas around the sieve bed. It is estimated that only O. 1 ?'O bypassed gas can prevent successful expander or LNG plant operation. Internal insulation may also extend the dryout or curing time required for plant startup. "Can type" insulation is not recommended. Gas channeling may result from several causes. One cause is poor flow distribution at either the inlet or outlet of the molecular sieve beds, Poor flow distribution may also contribute to impingement of high velocity gas on the bed surface. One of the simplest, lowest cost methods for control of gas flow distribution is to provide simple baffle plates over the inlet and outlet nozzles. A screen wrapped slotted pipe with gas exiting radially into the vessel can also be used. A moisture sample probe should be located in the dehydrator desiccant bed about 0.5 m [ 1.6 fi] from the outlet end of the bed and extending to the center. This probe, when used in conjunction with the outlet gas moisture probe, offers a source of valuable information for troubleshooting dehydration problems - particularly if a possible channel exists down the wall of the vessel. It also permits capacity tests for optimizing time cycles to be run with reasonable safety because movement of the water front can be detected prior to breakthrough. The probe can be a long thermowell drilled with 0.8 mm [1/32 in] holes on the sides near the end of the probe.

2. Heat Loss Management. There are several items that contribute to the regeneration gas heater load, e.g., heat required to heat desiccant, to heat the water, to vaporize the water, to

7

",



MECHANICAL DESIGN CONSIDERATIONS

t-

heat the steel, to heat the insulation, to heat piping and finally to make up heat losses from the vessels and piping. This last item can be of significance, especially for small units. At very low regeneration gas flow rates, it is conceivable that gas leaving the heater will not reach the bed at an adequate temperature or that heat losses around the vessel will be great enough to interfere with regeneration effectiveness. The regeneration gas system must be designed to ensure that heat losses are not excessive or that the heater and gas flow rates are large enough to provide for such losses.

Sieve Loading The loading of molecular sieve dehydration and treating vessels is an important step in assuring

that the unit will perform as designed. All openings for liquid level switches or other openings in or near the molecular sieve bed in liquid dehydrators and liquid product treaters will have to be equipped with screens to prevent sieve entry. The following procedures are recommended during loading:

1. Inert balls (three to four inches in depth) are usually placed on top of the bottom support screen to minimize adsorbent nesting in the screen opening and to aid in gas distribution. A rule of thumb says that the inert balls can be two to four times the size of the sieve particles. If necessary two sizes of inert balls may be used. The larger balls near the screen must be retained by the screen. Where available materials permit, graded sieve may be used instead of inert balls. Either balls or graded sieve must be placed evenly across the screen.

2. After the support balls have been loaded, the moisture probe should be inserted. Installa- tion of probes is very difficult after the sieve has been loaded.

3. The sieve should be loaded through a plastic or cloth sock that reaches to the support balls from the top loading nozzle. The sock must be moved in a random manner to level the bed and prevent coning and is pulled out as the vessel fills. An equivalent method to sock loading would be to lower the sieve into the vessel and place it at random locations using a dump bucket. If the sieve is simply dumped from the top or at the center from a sock or bucket it will segregate (large particles on sides,. small in center) and the bed will have more flow capacity at the sides, thus causing early adsorbent breakthrough.

4. A layer of inert balls (13 mm [OS in] diameter) should be placed on top of the molecular sieve bed. This layer of balls will prevent movement of sieve on the surface caused by high local velocities and may aid in gas distribution. A floating stainless steel screen should be placed between the inert balls and the sieve to prevent migration of the denser balls down into the sieve bed.

Mol Sieve Unit Operations Mol sieve units are reliable and usually require very little operating attention. Mol sieve units

can also be unforgiving. Some of the things that can cause serious operating problems are:

Feed Gas Conditioning

Without exception the most frequently encountered operating problem in molecular sieve systems is feed gas conditioning. The gas entering the bed should be free of all entrained hydrocarbons, treating chemicals (e.g., glycol, amine), free water, and solids. Although most desiccants are designed to withstand some carryover, significant or persistent carryover of entrained material will cause premature reduction of bed capacity and/or mechanical damage to the desiccant material.

<r



9.27


7 A properly sized impingement separator followed by a coalescing filter separator should be installed upstream of any solid bed dehydration system. If the feed gas is at its retrograde dewpoint (pressure is above cricondentherm pressure) it is advisable to heat the feed gas slightly (-5°C [9"F]) to preclude retrograde condensation.

Some of the most common bed contaminants are listed below:

1 . Hydrocarbons - Heavy hydrocarbons such as crude oil or lube oil fractions are adsorbed by the binder in macropores which are much larger than the active sites for water. These high boiling point fractions can undergo a number of reactions during regeneration including cracking, polymerization, or coking. Each of these can leave behind a high MW non- volatile "coke" which blocks or slows the diffusion of water molecules to the active sites. This results in a loss of dynamic equilibrium and premature breakthrough. Light hydrocarbons, such as NGLs, can also be adsorbed in the macropores. Again this slows diffusion, hence the adsorption rate, and results in premature breakthrough. The difference is that light hydrocarbons are driven off the bed during regeneration and do not leave a non-volatile residue. Light hydrocarbon contamination can occur due to entrainment or retrograde condensation.

2. Glycols - Glycols behave similar to heavy oils in adsorption bed contamination. They are adsorbed in the macropores and decompose during regeneration. Again this can form a heavy non-volatile coke on the sieve or in some cases can result in the "cementing" of sieve particles together to form "chunks". This encourages gas channeling which in turn leads to premature breakthrough.

3. Amines - Like hydrocarbons and glycol, amines can also contribute to coking. In addition, ammonia is formed during regeneration. Ammonia can attack the binder and weaken the physical structure of the sieve. Two or three water "wash trays" at the top of the amine contactor are recommended to minimize the canyover of amine compounds.

4. Salt - Salt usually enters a desiccant bed dissolved in entrained water. Unfortunately, it does not leave when the water is vaporized and is not removed from the bed during regeneration. Thus, the solid accumulates and blocks pores, macropores - and in extreme cases - binds the beads together to form "chunks". Once sufficient salt has accumulated to reduce adsorbent capacity below the minimum level required to maintain cycle times, it is usually necessary to replace the adsorbent. Unfortunately, there is no dependable way to recover desiccant capacity from such a unit. Adsorbers subject to such contamination are those treating gas from salt water bearing formations, and those treating LPG, propylene, etc. from "salt dome" storage caverns.

5. Oxygen - If there is any oxygen in the system, or in the regeneration gas, it will react with H2S and some other sulfur compounds on the surface of the sieves and deposit elemental sulfur. In extreme cases this will not only block macro and micropores but also the space between the sieve particles, resulting in one large "lump" which may have to be removed with pick and shovel.

Complications resulting from oxygen in a hydrocarbon system are not limited to the production of sulfur. Reaction with the hydrocarbons present, especially during the high temperature portion of the regeneration cycle, can result in heavy coke lay down and fouling of the sieves. Since oxygen can enter a system by a number of routes, it's a good idea to request an oxygen determination during any routine stream analysis. If only trace amounts are detected early, and the source is discovered and cut off, it should be possible to prevent severe damage to the sieves.

/7

@John M. Campbell & C o m p y


SPECIAL ISSUESlNEW DEVELOPMENTS

r" 6. HzS/Sulfur Compounds - H2S is adsorbed on 4A and 5A sieves. In fact molecular sieves are sometimes used for H2S removal from natural gas. This application is discussed briefly later in this text. When H2S and COZ are present in the feed gas, special sieves which minimize the formation of COS should be specified. Another issue which should be considered is the concentration of H2S in the regeneration gas. Any H2S adsorbed on the desiccant comes off the bed during regeneration in a "spike" lasting 5-15 minutes. In other words, all of the H2S adsorbed on the bed during the adsorption cycle is removed in this short 5-15 minute interval. This can increase the concentration of H2S in the regeneration gas to several hundred or in some cases several thousand ppm. Depending on the disposition of regeneration gas this may require temporary flaring or subsequent H2S removal on the regen gas. Unfortunately, all sulfur compounds are not designed for easy removal from the sieves. Heavy mercaptans and other large molecule, high boiling, sulfur compounds, are not efficiently removed during routine molecular sieve regeneration cycles. As a result, they tend to build in concentrations a bed ages and produce a capacity reduction much as do the previously mentioned heavy oils.

7. Methanol - Methanol is frequently used for hydrate inhibition in the production and gathering systems. Methanol has a vapor pressure higher than water, so significant quantities of methanol can be present in the vapor phase in the feed gas. Methanol is adsorbed on 4A sieve and can reduce the capacity of desiccant to adsorb water. If methanol is present in the feed gas, the mol sieve supplier should be notified so that additional adsorption capacity can be included in the design. Some installations have successfully used 3A sieve to preclude the adsorption of methanol.

SPECIAL ISSUES/NEW DEVELOPMENTS

f"

Adsorbents have been used to dehydrate, sweeten and remove NGLs from natural gas streams for over 50 years. Early units employed silica gel and/or alumina. Molecular sieves were introduced to the gas processing industry in the 1960's. Several improvements in the quality and versatility of these desiccants have resulted in new and innovative applications. A few of these are discussed below:

1. Compounds Beds - Compound beds are desiccant beds which use more than one desiccant type or desiccant size. The purpose is to increase the useful capacity of the bed by increasing the equilibrium capacity or shortening the MTZ, or both. The most common example of a compound bed is the use of 8-12 mesh desiccant at the bottom of the bed and 4-8 mesh at the top. The equilibrium capacity of the two desiccants is the same, but the rate of adsorption in the smaller (8-12 mesh) desiccant is faster - hence a smaller MTZ. Since the 8-12 mesh desiccant is at the bottom of the bed, the useful capacity of the bed is increased. This can result in longer cycle times and/or shorter bed depths. The 4-8 mesh desiccant at the top of the bed minimizes pressure drop and Al' forces on the bed supports. Another compound bed application involves the use of alumina at the top of the bed and molecular sieves at the bottom. Alumina has a higher static equilibrium capacity for water as well as a lower heat of adsorption. This results in a higher useful capacity and lower regeneration heating requirements. Alumina will often not dry the gas to sufficiently low outlet dewpoints but the molecular sieve at the bottom of the bed will remove the residual water not picked up on the alumina.

2. H2S - As previously mentioned, H2S is adsorbed on 4A, 5A, 1OX and 13X molecular sieves. Molecular sieves are sometimes used for H2S removal. This application is gener-



9.29


ally limited to those cases where the concentration of H2S in the feed gas is less than 200-300 ppm. The advantage of using molecular sieves is the simultaneous dehydration and sweetening of natural gas. 5A sieve is commonly used for this application; however, 4A sieve also has some useful capacity for H2S. There are several important differences between H2S and H20 adsorption. First, the equilibrium capacity for H2S is much lower than for water. In addition, the MTZ is longer for H2S. These two factors mean much lower useful capacities for H2S. The result is significantly shorter cycle times and/or longer beds required for HZS removal compared to water. Water will be adsorbed along with the H2S, but the water loading on the bed will be much less than the full usehl capacity. This is the case because the breakthrough occurs with H2S not H20.

If C02 is present in the gas the mol sieve can catalyze the reaction of H2S and COZ to form carbonyl sulfide (COS). COS is very difficult to remove from natural gas and NGL streams. In those instances where COS can form it is recommended to select a desiccant which minimizes the formation of COS.

3. Hydrocarbons - Silica gel has long been used to remove Cg+ hydrocarbons from natural gas. These hydrocarbons are adsorbed on the desiccant just like water. In general, adsorption cycle times are very short due to the lower useful capacity of the adsorbent for hydrocarbons relative to water. In fact, these units are often called "short cycle" or "quick cycle" units. Another term frequently used in "hydrocarbon recovery unit". This topic is covered in more detail in the "Refrigeration and NGL Extraction" course.

3. Adsorption vs. Glycol - In general, when the outlet water dewpoint requirement is above -40°C [-40"F], glycol (TEG) dehydration or glycol (MEG) injection is generally preferred due to lower capital and operating costs. If H2S is present in the gas, a fixed bed system may be chosen over glycol because of the coabsorption of H2S in the glycol and subsequent environmental impact at the regenerator. In adsorption systems, the regeneration cycle can be "closed" so that no H2S need be vented or incinerated. The heat of regeneration in glycol units is less than that required in molecular sieve systems. TEG systems average about 9200-1 1600 kJkg H20 [4000-5000 Btdlb HzO] while molecular sieve systems require approximately twice as much energy. Another factor which impacts the dehydration decision is the presence of aromatic hydrocarbons in the feed gas. Aromatic hydrocarbons have become a significant environmental issues in glycol units as they can be vented Erom the regenerator. Many stateidcountries now require mitigation of these emissions.

Aromatic hydrocarbons are not adsorbed on alumina and molecular sieve and are recovered as a liquid product in silica gel units - hence no environmental problem. Some companies have used dry desiccant units (in lieu of glycol) for this reason.

Another related topic which is often raised in the design of dry desiccant systems is whether to dehydrate the gas with glycol upstream of the solid bed unit. This may be difficult to justify economically, but there are a number of positive benefits to this configuration. First, the water content of the feed gas is significantly lower, hence longer adsorption cycle times and fewer regeneration cycles. This can increase the life of the desiccant and save several bed changeouts over the facility life.

A second advantage is energy efficiency. Due to the lower regeneration heating requirements for glycol, significant energy savings can be realized by removing the bulk of the water with glycol upstream of the dry desiccant unit.

7



LIQUID DRYING

LIQUID DRYING Most liquid sales contracts require the dried liquid to yield a negative result to the Cobalt

Bromide test, which is equivalent to a water content of 15 to 30 ppm.

Liquid velocity is often 1-1.5 d m i n [3-5 ftímin] with a liquid desiccant contact time of at least 3 seconds. As a general rule, though, a minimum bed depth of at least 2.5-3 m [8-10 ft] is provided to ensure good distribution and minimize channeling effects.

Adsorption cycles will normally be longer than those for gas because of low equilibrium water content in liquids; 24 hours is not uncommon.

Alumina is a satisfactory desiccant for liquid drying. If simultaneous sweetening and drying is desired, 13X molecular sieve is typically used.

Regeneration is similar to that for gas units. Liquid draining and filling time must be taken into account when determining cycle times. Also pressuring and depressuring steps must be allowed for, since the regeneration gas is usually at a different pressure than the liquid stream being dehydrated. Care must be taken in pressuring and depressuring steps to ensure that maximum velocities are not exceeded which could cause bed lifting (fluidizing) and attrition of the desiccant.



9.31

SOLID BED DEHYDRATlON

-! NOTES:

@John M. Campbell US Company


Section 10

REFRl G ERATl O N SYSTEMS TABLE OF CONTENTS

PAGE #

AMMONIA ABSORPTION SYSTE ........................................................................... 10.1 COMPRESSION REFRIGERATION ........................................................................... 10.3

.................................................. 10.4 ................................... 10.5

Calculation of Economizer Systems ......................................................................................... 10.6 Calculation of Chiller Load (Qchiller) ......................................................................................... 10.7 Specification of Liquid Separation T ................................................. 10.8 Choice of Refrigerant. ................................................. 10.9 Effect of Temperature .......................................

Calculation of a Simple System ................................................ Detennination of the Enthalpies ...................................................

Cascade Refrigeration .......................................................................... 10.1 3 Mixed Refrigerants .................................................... ........................................................... 10.14

APPLICATIONS OF REFRIGERATION .................................................................................................. 10.1 5 APPENDIX 10A - PROPERTIES OF COMMON REFRIGERANTS ............ ............................... 10.21

APPENDIX IOB .................................................................................. .................................................. 10.30

LIST OF FIGURES

FIGURE # PAGE #

1 O. 1 10.2 10.3 10.4 10.5

10.6 10.7 10.8

1 OA.l(a)P-H Diagram for Propane ........................................ 1 OA. 1 (b) P-H Diagram for Propane .........

Flow Sheet of an Ammonia Absorption System .......... ................................... 10.1 Flow Sheet of a Simple Refrigeration System ............................................................ . 10.3 Flow Sheet of a Simple Refrigeration System ............................................................................. 10.4 Schematic View of a Compression Refrigeration System Using Glycol Injection ................... 10.7 Relative Cost of Refrigeration Compared to the Cost of Cooling with Water at Ambient Conditions ...... ............................................................................................ .10.12 Mixed Refrigerant Flow ................................................................................................................ 10.15 Flow Sheet for Simple Refrigeration Plant with Stabilizer ................... .............................. 10.15 Flow Sheet and Data for a Compression Refrigeration Plant ................ .............................. 10.16

............................. 10.20

............................. 10.21

...... 10.23 10A.3(a) Pressure-Temperature Relationships of Refrigerants .................................................................. 10.24

10.9 Refrigerant Processes ............................................... ................................................... 10.17

IOA.2(a) Pressure-Enthalpy Diagram for Refrigerant 22 ........................................................................... 10.22 1 OA.2(b)Thermodynamic Properties of Refrigerant 22 .....................................................

1 OA.3(b) Pressure-Temperature Relationships of Refrigerants .................................................................. 10.25


Technical Assistance Service for the Design, 1 O.¡ Operation, and Maintenance of Gas Plants


LIST OF TABLES "7 TABLE # PAGE # 1 O. 1 10.2(a) 10.2(b)

Utilities Requirements for Ammonia Absorption System ........................................................... 10.2 Comparison of Common Refrigerants .......................................................................................... 10.10 Comparison of Common Refrigerants ..... ....._.... .............. .......... . ......... ............ ............................. 1 O. 1 1


1o.ii BP Exploration Company (Columbia) Ltd.

Section 10

f"

RE FR I G E RAT1 O N SYSTEMS refrigeration system lowers the temperature of the fluid being cooled below that possible when A using air or water at ambient conditions. A typical building air conditioner cools air to a tempera-

ture of 10-15°C [50-6O0F]. At the other end of the scale is the liquefaction of helium at -268°C [450"F]. The temperature produced depends on the goal of the exercise. If the goal is to produce marketable liquids, basic economics controls the temperature specified. If it is to meet a hydrocarbon dewpoint, that specification sets forth the required temperature.

Several basic processes will be discussed herein.

1. Absorption refrigeration 2. Compression (mechanical) refrigeration

AMMONIA ABSORPTION SYSTEM Figure 10.1 shows a flowsheet for a refrigeration system utilizing using two concentrations of

ammonia-water solutions. The basic driving force is the heat input to the generator. Ammonia vapor is stripped from the water solutions in the rectifier or stripper. This ammonia vapor is condensed and passes through a receiver, a heat exchanger (optional) and across an expansion valve into the evaporator or chiller. Here it vaporizes while cooling the fluid to be chilled.

Rectifiei

Aqua NH3 -- Waste Heat

I

Figure 10.1 Flow Sheet of an Ammonia Absorption System



10.1


Generator's steam sat. temp. "C, exit temp. Steam pressure, kPa Generator heat required kJ/m/TR Steam rate, kgíh/TR Water r uired thm cond. & absorber, m IhíTR 31°C on, 41°C Off

"3

The ammonia vapor from the evaporator is absorbed in a weak ammonia-water solution. The result is a strong solution. The ammonia is removed from this strong solution in the generator to begin its cycle all over again. Circulation of the strong solution is usually performed with vertical multi- stage pumps.

75 37.6 536 14.0

0.85

One can write selected energy balances around this system to determine loading at various points. The starting point is the evaporator. First, determine the total enthalpy required to chill the fluid to its desired temperature. This is the heat load in the evaporator.

95 81.6 700 18.5

The heat absorbed by the ammonia per unit mass is governed by evaporator pressure or temperature. The AH is the enthalpy of a saturated ammonia vapor at evaporator conditions minus the enthalpy of the entering ammonia, which should be a saturated or subcooled liquid.

100 105 110 115 120 125 130 101.3 120.6 142.7 168.6 198.7 232.4 266.8 759 800 843 891 943 987 1028 20.2 21.4 22.6 24.1 25.7 27.1 28.5

A series of such balances, and an overall balance, enable one to determine sizes and energy loads of each component part. Table 10.1 summarizes the typical loadings for the ammonia system in Figure 1 O . 1 .('O. '1

Evap. temp. "F I 50 Steam pressure, psia I 14.1

Steam sat. temp., "F (or waste heat exit 210 temp.) Generator heat required BtulmidTR 300 Steam rate, Ib.íhr.lTR 18.9 Water rate thru cond. & absorber, gpm/TR 85°F on 105°F off 3.6

TABLE 10.1 Utilities Requirements for Ammonia Absorption System

40 30 I 20 I 10 I O -10 -20 -30 -40 -50 19.8 24.9 I 30.9 I 41.8 I 53.2 67.0 83.2 103.1 134.6 173.3

225 240 255 270 285 300 315 330 350 370

325 347 373 400 430 466 511 571 645 754 20.2 21.8 23.6 25.8 28.1 30.9 34.1 39.1 44.6 53.2

3.7 3.8 4.0 4.3 4.5 5.0 5.5 6.1 7.1 8.8

Single Stage

Evap. temp. "C 1 1 0 1 5 1 0 1 -5 Steam pressure, kPa I 97.2 I 132.3 I 164.7 I 200.5

Two Stage Generator'ssteamsat. temp. 'F,exittemp. 175 180 190 195 205 210 Steam pressure, psia 6.7 7.5 9.3 10.4 12.3 14.1 Generator heat required BtulmidTR 550 577 605 637 670 711 Steam rate, Ib./hr./TR 33.2 34.9 37.0 39.1 41.3 43.9 Water required thru cond. & absorber, g p f l R 87°F on 105°F off 4.0 4.2 4.3 4.5 4.9 5.3

Steam sat. temp., "C (or waste heat exit temp.) Generator heat required kJ/m/TR Steam rate, kg/h/TR Water rate thru cond. & absor-

220 230 240 250 265 17.1 20.7 24.9 29.8 38.1 753 799 850 905 970 46.7 49.9 53.6 57.5 62.3

5.8 6.4 7.2 8.3 10.2

~

114 361 9.8

Singl

121 385 10.5

ber, m3/h/TR 29°C on, 4 1 T off I 0.82 I 0.84 I 0.86 I 0.89 Two

80 47.4 586 15.3

85 57.9 624 16.3

90 67.9 665 17.5

0.92 0.97 I 1.01

-10 I -15 I -20 I -25 I -30 I -35 I -40 I -45 258.1 I 327.5 I 404.8 I 495.8 I 601.0 I 734.1 I 928.0 I 1166.3 stage

0.95 I 1.00 I 1.07 I 1.17 I 1.28 I 1.41 I 1.61 I 1.99

1.09 1 1.22 1.33 I 1.45 1.62 I 1.84 1 2.09 1 2.35



COMPRESSION REFRIGERATION

Expansion 41 A Valve

The heat for the generator may be obtained from any one of four sources: (1) low pressure steam, (2) fired heater, (3) a hot process stream and (4) waste heat. Since heat loads are large, the availability of ( 3 ) or (4) increases the economic attractiveness of this system.

- -

Water cooling is shown for the condenser and absorber but air cooling may be used. As with all refrigeration systems, the higher the cooling temperatures the greater will be the net energy load.

t Condenser 1L

I Accumulator

Evaporator temperatures down to about -50°C [-58"F] are practical. Units have been designed with capacities to about 35 MW [lo O00 tons of refrigeration (TR)]. The unit is simple and has few moving parts to maintain. Ammonia solutions are not difficult to handle metallurgically.

A major psychological problem is ammonia smell. In a confined space this can be a nuisance. However, the pungent odor is a safety item that immediately confirms leaks.

Ammonia systems are generally competitive with compression systems in initial cost. Operat- ing cost comparisons are dependent on the source of heat and the cost of cooling. If the economics are competitive, they are a viable alternative to compression systems for producing refrigeration.

COMPRESSION REFRIGERATION Figure 10.2 shows the simplest compression refrigeration system. Saturated liquid at Point A

expands across a valve (isenthalpically). On expansion some vaporization occurs. The mixture of refrigerant vapor and liquid enters the chiller at 3-6°C [5-10°F] lower than temperature to which the process stream is to be cooled. The liquid vaporizes. Leaving at Point C is a saturated vapor at the P and T of the chiller. This vapor is compressed and then enters the condenser as a superheated vapor. /c'

I

The refrigerant must leave the condenser as a saturated liquid or slightly subcooled. Nothing It merely serves as a reservoir for refrigerant as levels vary in the happens in the accumulator.

chiller(s) and condenser.

Figure 10.3 is a more complex cycle composed of the same type of equipment. Assuming two stages of compression, a second expansion valve and a separator are added. This system will require



10.3


a Expansion Expansion

1 st Compression 2nd ComDression

L__

Condenser

less power per unit of heat load in the chiller. The reason? Part of the circulation rate is only compressed through one stage and the refrigerant entering the chiller contains less vapor. The vapor passing through the chiller does virtually no cooling, even though it contributes to compression cost.

The separator could be called an economizer, a word that applies to any device - usually a heat exchanger - that reduces net utility consumption. A separator-choke combination can be provided between each stage of separation, regardless of the number of stages.

One also can add a heat exchanger in the suction line to the compressor to exchange heat between the vapor from the chiller and the liquid to the choke preceding the chiller. This exchanger also could be called an economizer.

7

Calculation of a Simple System There are several discrete steps in the sizing of the system shown in Figure 10.2. These are

summarized below: Determination of refrigerant circulation rate -

The balance at right is around the chiller and expansion valve. At Point A the refrigerant is a saturated liquid (or very close to it). At Point C it is a saturated vapor. Qhi,,; - Qckller is the heat load determined by specifications on 4-

the stream being cooled.

If one writes an energy balance around the system, QcklIer + mAhA = m c k . But mA = mc = m, so

I C B - A

m = QchiiieAhc - h ~ ) (10.1)

SI Units FPS Units

Btu/lbm hc = saturated vapor enthalpy hA = saturated liquid enthalpy m = circulation rate lbm/hr

Where: Qchiller = chiller heat load



w


I" 2. Determination of Compressor Power -

This is done by any appropriate method. One will calculate theoretical (isentropic) work and use an efficiency to find actual work. The circulation rate from Step (1) is used.

D

3. Determinution of Condenser Heat Load (Qcond) -

There are two ways to do this. Knowing Qckller and W, you can write an overall balance as shown in Figure - 10.2 to find Qcond.

If you are performing the calculation manually and wish an independent check of the previous work, write the balance shown at right. - The Qcond found from Equation 10.2 will be negative. This merely signifies that heat is leaving the system. The value found from the overall balance will not check the condenser balance exactly because only part of the compressor inefficiency shows up in the refrigeration system. For practical purposes, the difference is trivial.

Determination of the Enthalpies f-

The calculation requires one to find the enthalpy per unit mass at Points A, B and C. This can be done from a computer routine. Convenient tables and figures are available for all of the common commercial refrigerants. Appendix B at the end of this volume contains data on substances used as refrigerants. Appendix 10A at the end of this chapter contains pressure-enthalpy (P-H) figures for propane and Refrigerant 22 (a Freon) as well as vapor pressure and physical property information on all common refrigerants.

The P-H diagram is very convenient for solving the energy balance for a simple system.

H

P

H

The left-hand figure is a representation of the P-H diagrams in Appendix 10A. The refrigerant is all liquid to the left of the saturated liquid curve; it is two-phase inside the saturation curve and all vapor to the right of the saturated vapor curve. The lines of constant temperature are horizontal between the saturated vapor and liquid curves and then rise almost vertically in the liquid section.

@John M. Campbell & Company ~


10.5


-7 The calculation process starts by choosing the temperature of Point A. Will water, air or some

other stream be used for condensation of the refrigerant? What temperature can we realistically achieve in the condenser? That is Point A. It is on the liquid saturation curve, since it leaves the condenser as a liquid.

What is the temperature at Points B and C? Normally, it will be 3-6°C less than the minimum desired temperature for the fluid being cooled. This approach fixes the location of Point C. It is on the saturated vapor curve, since it is in equilibrium with the liquid in the chiller (evaporator).

The expansion across the choke from Point A is an isenthalpic process; a vertical line on a P-H diagram. Draw a vertical line from A to B, the pressure of Point C, and then go horizontally to C. One can read the Ah required for Equation 10.1.

The theoretical compression is isentropic. Starting at Point C, parallel the constant entropy lines until you intersect the pressure line of Point A. This is theoretical Point D.

Wihecw = (m) (hD - hC) Equation 10.2 for condenser heat load is found from the Ah between Points D and A.

For a commercially pure refrigerant, use of a P-H diagram or a corresponding table is as reliable as any method.

H

The economizer system shown in Figure 10.3 can be calculated in the same manner.

The P-H diagram is shown at left. Point E is at a pressure that allows the compression ratio to be the same in each stage of compression. Expansion from A to E is isenthalpic. The vapor formed goes to the second stage of compression as a saturated vapor. The saturated liquid leaving the separator is expanded isenthalpically to pressure B. Notice that the Ah available from B to C for this system is larger than for the simple system.

Calculation of Economizer Systems Determination of circulation rate, work and Qcond for the system in Figure 10.3 follows the

same pattern as for the simplest system.

lm The first step is the same in principle. Now, however, the enthalpy of the refrigerant entering the choke is determined by compressor interstage pressure and not by condenser pressure. With this change in hA, Equation 10.1 may be solved for "m1," the circulation rate to the chiller (and the amount of gas to be compressed through Stage 1 of the compressor).

What is the circulation rate through the condenser? It is ''m2,'' where m2 = ml + mv. If you write the balance shown at left,

rl +-pa--mz



111'

r


m2h = mlhL+mvhv

Since m2 = ml + mv, one can take m2 = 1.0 and then define x and y as the relative amounts of liquid and vapor, respectively, from the separator. Then,

h = xhL+( l -x)hv or x = hv-h hv - hL

(10.3)

The three enthalpies can be obtained, which enables one to solve for "x," and then "y." Knowing these relative quantities and ml, m2 and mv may be calculated.

In the compressor work calculation, one finds the work in the first stage (for flow rate "ml") and adds it to second stage work (for flow rate "m2") to find total work. The resultant temperature of the gases after they mix in the tee between stages may be found by a balance around that tee. In most cases, the temperature effect here is negligible.

The condenser heat load is found as before.

If one places a heat exchanger to exchange heat as shown, an extra balance is needed. The cold vapor will subcool the saturated liquid from A and there will be less vaporization upon expansion across the valve. However, the gas will go to the compressor at a higher temperature and lower pressure. Is the investment worthwhile? Only a calculation will tell. One might add the heat exchanger shown to the system in Figure 10.3.

Calculation of Chiller Load ( Qchiller)

Feed Stream

Residue I Gas I

I I

Stabilization I Refrigerant I or Storage I I

I Auxiliary Glycol Injection Equipment I

. I - - I ;y-. .. :.::::... I A. Gas to Gas Heat Exchanger B. Chiller

I ; / I C. Separator I / ! I I j D :+I D. Glycol Flash Drum I ! E. Glycol Regenerator I : L .......... ; I ' .....r... . I _ _ _ _ I

F. Glycol Pump

I

Figure 10.4 Schematic View of a Compression Refrigeration System Using Glycol Injection


Technical Assistance Service for the Design, Operation, and Maintenance of Gas Plants 10.7

m


The first step in the design is to fix the temperature (T3) in the low temperature separator (LTS). The pressure in the LTS must be high enough above the specified sales pressure to allow for pressure drop in the gas-gas exchanger and lines.

The minimum temperature coming to the gas-gas exchanger is fixed by the economics of pre- cooling the feed stream. The maximum sales gas temperature is usually fixed by contract and is seldom allowed to exceed 50°C [122”F]. Consequently, the heat load between PI, TI and P3, T3 is fixed by these considerations. The problem revolves around the distribution of this load between the gas-gas exchanger and the refrigerated chiller. For this calculation, it is convenient to assume a 34 kPa [5 psi] drop in each heat exchanger. The following general procedure is suggested if the gas-gas exchanger and chiller are to be sized as part of the exercise.

-?

1.

2. 3.

4.

5 .

Determine the amount and composition of vapor and liquid at TI, PI and T3, P3. (Don’t forget the water and glycol.) Calculate the total AH between points 1 and 3 (HI and H3). Calculate the cooling capacity for the sales gas in the gas-to-gas exchanger (H6 - H4), T6 should be fixed at (TI - 5°C) or the contractual maximum temperature, whichever is lower. Calculate the chiller duty, Qchi1ler, by difference

Qchiiier = (HI - H3) - (H4 - H6)

Perform a series of flashes on the feed stream at temperatures and pressures between points l and 3. Plot AH versus T.

Using T4, H4 and T6, H6 plot AH versus T for the residue gas.

AH

This graph is used to determine the temperatures necessary for sizing heat exchangers. It may be generated by hand using the procedure outline previously or developed by computer simulation.

The cold liquid from the LTS is also available for cooling service since usually it must be heated before entering the fractionation system. Although not shown in Figure 10.4, it may be used for cooling the feed or for any other cooling function within the system. If this liquid is not heated, a cold-feed stabilizer might be specified.

Specification of Liquid Separation Temperature The desired temperature of the gas-liquid stream leaving the chiller must be determined by the 7

strategy governing the system. If hydrocarbon dewpoint control is primary and liquid recovery




secondary, this temperature should be about 3-5°C [5-9"F] below the temperature required to achieve the specified dewpoint. This required temperature may be significantly colder than the specified dewpoint if the gas is processed at high pressure.

r

If liquid recovery is the primary function of the unit, what products are salable? The basic strategy is to condense the least amount of nonsaleable components (usually methane or ethane) compatible with the economics. Anything condensed, like methane, that must be revaporized and maybe recompressed, adds to the operating costs without contributing to liquid revenue.

For a given set of specifications, one should investigate a series of LTS separator pressures and temperatures. The pressure for maximum liquid recovery is between 3.0-4.0 MPa [435-580 psia], if propanes plus are the salable product. As the pressure increases, condensation of methane increases. However, 3.0-4.0 MPa [435-580 psia] may not be the optimum pressure economically because of system logistics. The optimum pressure must minimize total system cost, not merely that of the refrigeration system alone. As a general rule, separation is carried out at necessary sales gas pressure to eliminate recompression.

For the usual pressures chosen, what is a reasonable temperature? As noted before, this depends on the products desired. If the liquid product is to be stable at atmospheric pressure and sold as "crude oil," a common optimum separation temperature is O-5°C [32-4O0F]. When propane is the lightest salable liquid, the temperature may be -40 to -1 8°C [-40"F to O"F]. This temperature depends on the recovery desired and whether absorption or adsorption is combined with refrigeration. Below about -40°C [-40°F] you get into the cryogenic range for ethane recovery.

Choice of temperature (and pressure) is a critical specification. Do not choose arbitrarily! Calculate the economics for a series of conditions and choose the optimum one. f-

Choice of Refrigerant

The ideal refrigerant is nontoxic, noncorrosive, has PVT and physical properties compatible with the system needs, and has a high latent heat of vaporization. Any material could be used as a refrigerant. The practical choice reduces to one which has desirable physical properties and will vaporize and condense at reasonable pressures, at the temperature levels desired. The usual choice is propane, ammonia, R-12 or R-22 at chiller temperatures above about 40°C. At cryogenic conditions, ethylene and methane might be used. In general, the lower practical limit of any refrigerant is its atmospheric pressure boiling point. It is desirable to carry some positive pressure on the chiller to obtain better efficiency in the compressor, reduce equipment size and avoid air induction into the sys tern.

Table 10.2 furnishes an excellent comparison of the common refrigerants. Consider, for example, a unit where the evaporator (chiller) operates at -18°C [OOF] and the condenser at 35°C [95'F]. Ammonia requires the lowest mass circulation rate, but propylene and propane have slightly lower horsepower requirements. This table also illustrates the effect of condensing temperatures on horsepower and circulation needs. Raising the condensing temperature 17°C [3 1"F] increases horsepower requirements about 60% for propane and 43% for ammonia. There are two morals - (1) use the lowest temperature possible for condensing to minimize cost; and (2) if higher condensing temperatures are required, ammonia might be preferable to propane. Ammonia is seldom chosen because of emotional reactions to its odor. However, it is easy to handle in ordinary steel equipment containing no copper and brass and is really less dangerous than propane because of its pungent odor. No dangerous accu- mulation can build up unnoticed.

t-


Technical Assistance Service for the Design, 10.9 Operation, and Malntenance of Gas Plants

Ir


o m * 2 m m o F ? ?

0 0 0 -

8 % S W ..?a! 0 0 0 - 0 0 0 0

w r - N W

O 0 0 0 ; 9 % q % i n e l .

o i n m -

O O O N h i m O q 9

w - m g - ? . Y

0 0 0 c

E N N W ? ? ?

0 0 0 -

i n o o w m m o d m o o m w o 0 0 0

00 W

2 c! E

Lr N

E

m .. F - a a e ;f 0

e a e

a

N o

>

a 00 00

- ..

m .- B

4" 2 E

E d

.- o

a b.

y: 8 s N in

.E

I- o 24 -8

5

B 8

m m r - r - W T ? .

o i n m o * Q I N l .

O 0 0 0 Q ' ? o \ o ! - m w w m m m d

m * m N 0 0 0 4

m o o N W O N 0 N d m N 0 0 0 N

w o m y w m - ? ? o !

0 0 0 -

- - o c i n m r - O O O N .ai%.

m o - r - ? ? ? ' ?

e w r - y m o w ? - ? o !

0 0 0 -

m o - w - m 0 0 - N h i o ! 9 ?

m w m z N N N N O 0 0

m w m e y?:: o o m o m v i m N..9 O O O N

w - o m

* - o 0

N m m o e m 3 0 0 0 N

w m i n O m " r 4 m m h ( 0 0 0 4

N o -

w w r - ? ? ? 9 - 7 i n - e w i n r 4

m r - N

m m m ? h i ?

e * - - - r 4 -.- - 0 - y?-+ C n r - m m B

v i v i d

10.1 o BP Exploration Company (Columbia) Ltd.


r

I

t---

m m y ? " m *

a

r- q' I I

l r 4 a C

o .-

m o m - N o * r r i m m r

N t " o r - m h l m o

r - m a b b t t c ? "

____I

O W C - t 9 6 r - m w m v l

$'0 C

n i I I I I

10.11 Technical Assistance Service for the Design, Operation, and Maintenance of Gas Plants

I


Propane is by far the most popular refrigerant in the gas processing applications. It is readily available (often manufactured on-site), inexpensive and has a "good" vapor pressure curve. It is flammable but this is not a significant problem if proper consideration is given to the design and operation of the facility.

Freons are widely used as commercial refrigerants. They are non-toxic and nonflammable. CFC (Chlorofluorocarbon) refrigerants like R-1 1 and R-12 are being phased out due to environmental problems. HCFC refrigerants (Chlorofluorocarbons containing at least one hydrogen) are currently considered environmentally acceptable. R-22 is the most popular although other HCFC's are being developed as replacements for R-1 1 and R-12. Freons are expensive and system losses can represent a significant operating cost. They are also difficult to ship to remote locations in large quantities. Cer- tain freons will form hydrates so it is necessary to keep the system dry.

Compressor choice is linked to refrigerant choice as well as other considerations. Where weight and size are particularly important, a centrifugal or a screw compressor may be used. A reciprocating compressor is an excellent alternative for accessible land locations. The choice will depend on total power requirements.

Regardless of the choice, the controls must accommodate frequent, and sudden, load changes. As gas stream flow rate andor composition changes, so will refrigerant circulation rate. Thus, some form of speed control is often desirable. In many cases, electric motors make an ideal driver when compressor specifications are compatible with motor characteristics.

Temperature, "C

Temperature, "F

Figure 10.5 Relative Cost of Refrigeration Compared to the Cost of Cooling with Water at Ambient Conditions

Effect of Temperature on Cost Figure 10.5 shows the approximate

relative effect of temperature on compression refrigeration cost. It emphasizes the previous statement that one should use the highest temperature compatible with the goals of the installation. I keep emphasiz- ing this because too many persons buy a "standard" unit without much thought about the temperature level needed. This may be a waste of money.

The inset to Figure 10.5 shows the refrigerant often used at various temperature levels. The temperature levels are approximate. Actually, the refrigerant used at a given level depends on economics, which will vary in different circumstances.



COMPRESSION REFRlGERATiON

- -

Cascade Refrigeration When refrigeration must be provided at very low temperatures < -4OOC [4O0F], cascade re-

frigeration systems are sometimes used. Cascade systems employ more than one refrigerant and provide refrigeration at several levels. A propane cascade system is shown below.

Ethane System

I I Propane System

In this system, refrigeration is provided at five levels

Databook

Q, 103 kw 7°C [44"F]

-4OOC [-40°F]

O 2 4 6 8 10 12 14 16 18 20 100 I I I , , , , , , ,~ -20°C [-4"F] - 30 - 20 - 10

- 0 o --lo O -

- -30 5 - 4 0 - 4 0 - 4 0 I- - -70 --8o - -90

-61°C [-78"F] -84°C [-120"F] - -20 :

The propane at -40°C is used to condense the

ess is ultimately rejected to the cooling water at the propane condenser. A hypothetical cooling curve for the -120 - ethane refrigerant. All of the heat picked up in the proc-

process fluid has been developed to show the amounts -140 O 10 20 30 40 50 60 70 and levels of refrigeration. Q, MMBtu/hr



10.13

U


Cascade refrigeration systems are not common in NGL recovery. They are used in nitrogen rejection units (NRU) and helium recovery units (HRU). Low level refrigeration is typically provided using expansion or mixed refrigerants.

Mixed Refrigerants An alternative to cascade refrigeration is to use a mixed refrigerant. Mixed refrigerants are a

mixture of two or more components. The light components lower the evaporation temperature and the heavier components allow condensation at ambient temperatures. The evaporation process takes place over a temperature range rather than at a constant temperature as with pure component refrigerants. This is illustrated in the P-H diagram below for a mixed refrigerant.

H

The mixed refrigerant is blended so that its evaporation curve matches the cooling curve for the process fluid. Heat transfer occurs in a countercurrent exchanger, probably a shell and tube or an aluminum plate-fin, rather than a kettle-type chiller.

Mixed refrigerants have the advantage of better thermal efficiency, because refrigeration is always being provided at the warmest possible temperature.

The amount of equipment is also reduced compared to a cascade system. Disadvantages include a more complex design and the tendency for the heavier components to concentrate in the chiller unless the refrigerant is totally vaporized. The use of mixed refrigerants is very common in low temperature gas processing today and is standard practice in large LNG plants.

One comparison between a mixed refrigerant plant and expander plants indicates a mixed refrigerant plant to have lower capital and operating costs and slightly lower This process is shown in Figure 10.6. This application was for a low volume (7-15 MMscfd) propane recovery facility operating at medium pressure (510 psia). The capital and operating costs were approximately 15% lower while the liquid recovery was reduced by less than 3%.


10.14 ~~


APPLICATIONS OF REFRIGERATION

flux Condenser

Reflux Accumulator

Refrigerant Separato

Deethanizer Feed Flash Drum

* Deethanizer

L 1 I I +-j I I

I I Deethanizer Reboiler 'í I (485 MBtdhr) I O0 I

- - - - -J (150 NGL MBtu/hr) Cooler

To Storage

Figure 10.6 Mixed Refrigerant Flow

APPLICATIONS OF REFRIGERATION There are many ways in which refrigeration is applied.

Figure 10.7 shows a flowsheet for a very simple refrigeration plant with a stabilizer used for a field installation. Note that a salt bath heater is used and to simplify the system there is no reflux on the stabilizer.

Product To Storage n Coolina t

I Tower-

Gas-Gas

LLC

Glycol-Oil-Gas Separator

II It II - )

e GESTO. Supply

Stabilizer

Salt 0 Bath Heater

Figure 10.7 Flow Sheet for Simple Refrigeration Plant with Stabilizer



10.15

rn


Figure 10.8 shows one arrangement for recovering liquid from oil treater and stock tank vapors.

Instrumentation

Still Overhead Deethanizer Chiller Reflux Chiller

Lean Oil Rich Oil Presaturator Dernethanizer

Chiller Side Chiller

Inlet Gas Chiller

te-heat oil heat- Ib steam boilers (three 90 O00 Ibhr)

365 x 0.01 = $199 OWiyear

Gas turbme Steam turbine incremental expense

h p year for gas turbime

Figure 10.8 Flow Sheet and Data for a Compression Refrigeration



APPLICATIONS OF REFRIGERATION

More common type applications are shown in Figure 10.9.(4.7' Type C and B will have higher recoveries with C providing the best separation of NGL product from residue gas. As the gas becomes richer, the cold separator pressures higher, and the temperatures colder, the recycle rate for Type A plants becomes higher.

Recompressor

I I I

' Product

Gas-Gas Gas-Feed Chiller Cold Exchanger Exchanger Separator

Recompressor Type B

Stabilizer

Gas-Gas Exchanger

Product

Recompressor Type c

Residue

Inlet

1 1 Gas-Gas Gas-Feed Gas Chiller Cold

Exchanger Exchanger Separator

Stabilizer

Product

Gas-Feed Cold Exchanger Gas Chiller SeDarator

itizer

Gas-Gas Exchanger 111 Product

Figure 10.9 Refrigerant Processes




10.1

10.2

10.3

10.4

10.5 10.6

10.7

10.8

REFERENCES

Briley, G. C., Hydr. Proc. (May 1976), p. 173.

Linhardt, H. D., LNG/Ciyogenics (Feb. 1973), p. 7.

Bleakley, W. B., Oil & Gas J. (April 3, 1967), p. 236.

Herrin, J. P., hid. (June 20, 1966), p. 181.

Davis, C. D., /bid. (Mar. 25, 1968), p. 132. MacKenzie D. H. and S. T. Donnelly, "Mixed Refrigerants Proven Efficient in Natural Gas Liquid Recovery Process," Oil & GasJ. (March 1985), p.116-120.

Russell, T. H., "Straight Refrigeration Still Offers Processing Flexibility," Oil & Gm J. (Jan. 24, 1977).

Hollek, R. G., An Minkkinen, and Joseph Larue, Proceedings of the 1996 Laurance Reid Gas Conditioning Confer- ence, "The IFPEX- 1 Process for Natural Gas DehydrationíHydrate Inhibition, The North American Experience," Nor- inan,OK(1996),p.1-14.



APPENDIX 1OA

APPENDIX 10A

PROPERTIES OF COMMON REFRIGERANTS

f-


Technical Assistance Service for the Design, 10.1 9 Operation, and Maintenance of Gas Plants




~

10.21




T ~ ' ' '



Technical Assistance Service for the Design, Operation, and Maintenance of Gas Plants 10.23

1


-40 -30 -20 -10 O 10 20 30 40 50 60 70 80 90 100

Temperature, "C

Figure lOA.3(a)Pressure-Temperature Relationships of Refrigerants




137.38 23.7

197.8 4.38

0.920

1.11

Chemical Formula

120.93 -29.8 112

4.1 1

0.983 0.61 1 0.544 1.14

Molecular wt

Critical Pressure. MPa

85.48 40.8 96 4.94

1.402 0.623 0.531 1.18

Specific heats kJ/kg."C of liquid @ 30°C Cp of vapor @ 16T, 101.3 kPa Cv of vapor @ 16T, 101.3 kPa Ratio Cp/C, = K (30"C, 101.3 kPa) Ratio of Specific heats liquid, 46°C

187.39 47.6

214.1 3.41

0.912

1.12

Saturated liquid@ 35°C @ 41°C Vapor @ 101.3 kPaand -1°C and 4°C

170.93

and 10°C Thermal Conductivity(')

99.29

Saturated liquid@ 35°C @ 41°C Vapor @ 101.3 kPa and -1°C and 4°C and 10°C

3.6 145.7

3.27

0.996 0.653 0.607 1 .O9

1.59

Same

TABLE 10A.l

Comparative Data of Refrigerants

-33.3 122.8

4.35

1.255 0.715 0.632 1.13

I .77

Same

Refrigerant Number íARI Desienation)

0.3420 0.3272 0.0108 0.0109 0.01 11

F-11 I F-12

0.2150 0.2100

CCL3F I CC12F2

odorless when mixed

0.3893 0.3723 0.0101 0.0103 0.0105

1.55

Same

0.2463 0.2395 0.0118 0.01 19 0.0121

0.371 0.362 0.292 0.028 0.029 0.029 0.03 1 0.029 0.032

F-113

0.2253 0.5845 0.2207 0.5472 0.0120 0.0097 0.0122 0.0098 0.0124 0.0100

0.357 0.3 19 0.345 0.312 0.037 0.023 0.038 0.024 0.039 0.025


F-114 I F-500 CzCIzF4 I 73.8% CClzFz

0.271 0.262 0.035 0.036



Boiling point at I atm (F)

TABLE 10A.2

Comparative Data of Refrigerants

74.7

f-

117.6 1 38.4

F-22

-28.0

495.0 1 474.0

Chemical Name Trichloroinono- fluoromethane

Molecular wt 137.38 Gas Constant, R (ft-lbilb-R) 11.25

_ _ ~

631.0 0.335

Critical temperature (F) 388.0 635.0

Specific heat of liquid Specific heat of vapor, C,

60°F at I atin Specific heat oí' vapor, C ,

60°F at I atm

0.218 0.238

$

I

1.12

I .47

1.51

* 0.84 2.66

I I .58

54.54

0.156 0.171

0.145 0.151 1 .O9 1.13

I .59 I .77

1.65 2.10

1.35 5.96 27.96 15.22 60.94 50.29 167.85

43.46 59.82

* ((Q -50°F

@J 40°F

Net refrigerating effect (Btuilb)

Ratio Cp'C, = K (86°F at 1 atin) Ratio of Specific Heats

Liquid, 105°F

67.56 Cycle efficiency (% Carnot cycle)

Solubility of water in refrigerant 40-105°F

Neeligible

1 . 1 1

2.04 Vapor, C,. 40°F sat. press.

Liquid heat (A), 1 psi at 105°F Saturation pressure (psia)

Theoretical horsepower per ton, 0.676

Coefficient of performance 40-IO5"F

1.61

40-105°F(4.71/hp per ton) Cost compared with R-I 1 I 1 .o0

87.5 Negligible Miscible

* Data not available or not applicable

84.9 82.0 Negligible Negligible Miscible Miscible

F-12

Miscibility with oil Toxic concentration (% by vol.) 3dor

Warning properties Explosive range (% by vol.) Safety group, U.L. safety group, ASA B9.1 roxic decoinposition products Viscosity (centipoises)

Saturated liquid 95°F 105°F

40°F 50°F

Vapor at 1 atin 30°F

rhermal conductivity, k(') Saturated liquid (@ 95"

Vapor at 1 atm 30°F 105°F

40°F 50°F

-iquid circulated, 40-105°F

Theoretical displacement, (Ibiminiton)

Dichloradi- fluoromethane

CCIzFz

120.93 12.78 -2 I .62 -252 233.6 597.0 0.220

0.146

0.130 1.14

1.55

I .84

7.12 23.85 51.67 141.25

49.13

83.2 Negligible Miscible

Above 20% Same

I -

Miscible Above 10%

Ethereal, odorless when mixed with air

None None 5 1 Yes

0.3893 0.3723 0.0101 0.0103 0.0105

0.0596 0.0581 0.0045 0.0046 0.0046 2.96

16.1

None None 6 1 Yes

0.2463 0.2395 0.01 18 0.01 19 0.0121

0.481 0.469 0.0047 0.0049 0.0051 4.07

3.14

0.736

6.39

1.57

* Same

None None 4.5 1 Yes

0.5845 0.5472 0.0097 0.0098 0.0100

Monochlorodi- fluoromethane

CHCIFz

86.48 17.87 -41.4 -256 204.8 716.0 0.235

0.149

0.127 1.18

2.14

2.04

11.74 38.79 83.72 227.65

66.44

81.8 Negligible

Limited * Above 20% Above 20??

Same Same

None None None None 6 5A 1 1 Yes Yes

0.3420 0.2150 0.3272 0.2100 0.0108 0.0109 0.01 11

a

b

*

Same

None None 5A 1 Yes

0.2253 0.2207 0.0120 0.0122 0.0124

0.573 0.553 0.0060 0.0061 0.0063 3.02

1.98

0.75

6.29

2.77

O J o h M. Campbeii & Compaoy

Azeotrope of Dichlorodi-

Tnchlorotri- Dichlorotetra- fluoromethane fluoroethane ¡ fluoroethane 1 and

Difluoroethane CClzF-CCIFz 73.8% CClzFz

187.39 170.93 99.29 15.57

417.4 1 294.3 I 221.1

0.512 0.0435 0.0500 0.0037 0.0039 0.0040 3.66 4.62 3.35

* 7 0.722 0.747

2.15 2.97 2.00

( I ) k in Btu-ft/hr-f?-"F


10.27

m


APPENDIX 10B -7

There are two sets of figures, one in SI Units and one in English units. Where lines are dashed the correlation is less exact than where solid lines are shown.

These figures are useful for estimations of enthalpy necessary for planning, specification and operations. Since enthalpy is composition sensitive they are not recommended for detailed design calculations.



APPENDIX 1OB


~

10.29


O O r-

O 8

O B O O Ln

O o O' o



APPENDIX 1OB

I


e! a ii w

h m 5 o 2 a ii

r

w


10.31

APPENDIX 10B

#-

A("

O

3 O O W

O 3

O O Lo

5: o e 1 ii

3 m

O O

Io O O Lo - O

O 2

O O m - O O

N O




O in W

O O W

O in v>

O O in

5; d

8 d

O m

O O d

O O O O 2

8 O

E 0 O O O O a IC

~ ~ ~~~~


10.34 ~~

BP Exploratlon Company (Coiumbla) Ltd.

APPENDIX 1OB



10.35


h m 4 o r

e!

i i S c13


APPENDIX 106



10.37

APPENDIX 1OB

a Wr>i ‘ k h w 3 0 i i



10.39

U


O v) W

O O (o

O In In

O O v) m

I

~~


r

APPENDIX 1OB

8 r. 5; (o

8 (o

8 m

L

@John M. Campbell & Company @John M. Campbell & Company

g! a ii o)


10.41

r"

Section 11

VALVE EXPANSION: JT PLANTS TABLE OF CONTENTS

PAGE # PRESSURE DROP ACROSS VALVE ....................................................................................................... 11.5

Liquid Condensation Across a Valve .................................................................................................. 11.6

LIST OF FIGURES FIGURE # PAGE # 1 1 . I 1 1.2 1 1.3

Process Flow Diagram - LTS or J/T Plant for NGL Extraction .............................................. 11.2 Two Types of LTS Systems Used in Production ........................................................................ 1 1.3 Process Flow Diagram - Refrigerated JT Plant ........................................................................ 1 1.4

@John M. Campbeii Bc Company


11.1


NOTES:


11.N BP Exploration Company (Columbia) Ltd.

n

Section 1 1

VALVE EXPANSION: JT PLANTS Processes which use the cooling effect of the expansion of a gas across a valve or choke are

sometimes called LTS (Low Temperature Separation) or LTX (Low Temperature Extraction) units. A more common name for this type of process is a J-T plant, where the J-T refers to Joule-Thomson. This is because the thermodynamic principle that explains the expansion of gas across a valve is called a "Joule-Thomson expansion," and is named for the scientists who first explained it.

Figure 11.1 presents a process flow schematic for a typical J-T plant. The full well stream fluid enters a separator where any condensed liquids are removed. The gas then flows to a heat exchanger where the incoming gas is cooled by the processed gas stream. A pressure drop is then taken across an expansion valve and the gas temperature is further reduced. The condensed liquids are then separated from the gas stream in a low temperature separator. The cold gas is routed through the inlet gadgas heat exchanger to cool the incoming gas, and then sent on to sales. Since the gas is cooled in the process to an extent that it passes into the hydrate formation region, ethylene glycol is injected upstream of the low temperature separator (or upstream of the gadgas heat exchanger, depending on temperature and pressure levels) in sufficient quantities to depress the hydrate formation temperature below that of the low temperature separator's temperature.

r'

J-T plants are simple and easily operated facilities. However, they have the limitation that the flowing wellhead pressure must be at least 2000-3000 kPa [300-500 psia] above the sales pressure for the system to reach low enough temperatures to meet normal dewpoint requirements. When the reservoir depletes to the point that this excess pressure is not available, the process ceases to function as a dew-point control method unless front-end compression (or residue compression) is installed to maintain the inlet pressure or mechanical refi-igeration is added to assist in cooling the gas. These units are most commonly used to process high pressure, non-associated gas with low flows (less than 10 MMsc fd) .

Figure 1 1.2 shows two other versions of what are commonly called LTS or LTX systems.

Figure 11.3 provides an example of a refrigerated JT plant. This type of facility is configured like a turboexpander plant, but the expander is not installed. The mechanical refrigeration is required for rich gas streams and to obtain high recoveries. This plant had a free pressure drop and maximized propane plus recovery as shown below:

Propane Butanes 1 O0 Pentane 1 O0



11.1

lm ,


W U W I n 8 5 - O I-

I

, - - - - - - - - - - - - - - - - - - - - I

I I I - - - - ?



I- I I

I I I I I I I I I fl

I I I 1 L.L.C. o I &T.C.

I

b Sales Gas

Separator

I I I I

I

I I

I - - _ _ _ _ _ _ _ I 4

I . . 2 I

- - - - - - . . 0.

4 I I .

I Condensate

L - - - - - - - - _ _ _ _ _ _ ---- - - - - - - - - - Stream > . . Water I _I.

Off

(Alternate Line)

High Pressure Water to Wellstrearn Waste

Figure 11.2 Two Types of LTS Systems Used in Production



I


0John M. Campbell & Compauy


i l ' '

PRESSUREDROPACROSSVALVE

PRESSURE DROP ACROSS VALVE Pressure drop across a valve (choke) is an isenthalpic process, as noted previously. If no liquid

forms, the following equation applies:

(11.1)

The symbol "p" is known as the Joule-Thomson coefficient. It is positive or negative, depending on the relative size of the two terms in the numerator.

a,

3 E - 3

Temperat u re Pressure

Curve A below shows a case where the instantaneous slope is greater than the average slope. Therefore, the gas will cool on expansion. The curve C gas is just the opposite and will heat on expansion. Curve B is for an ideal gas, which will not change temperature on expansion.

Many gases exhibit a characteristic wherein the slope of the V-T curve changes sign. The temperature at which the slope changes sign (p = O) is known as the inversion temperature. The right-hand plot above shows inversion temperature versus pressure. The shape shown is general for all actual gases. Outside the curve, the gas represented would heat upon expansion. Inside, it cools on expansion.

Because of the location of the curve, hydrogen heats on expansion at normal pressures, whereas most light hydrocarbons cool. At very high pressures, of the order of 60 MPa [8700 psia], many naturally occurring hydrocarbon gases heat on expansion.

Curves which show the temperature drop expected for a given pressure drop across a choke are only applicable if no liquid forms on such expansion.


Technical Assistance Service for the Design, Operation, and Malntenance of Gas Plants

11.5


Liquid Condensation Across a Valve Across a valve the First Law of Thermodynamics

reduces to hl = h2. No work is possible and the process is almost adiabatic. The amount of heat transfer across a valve body is poor, and the gas is in it for only a short time. The calculation is inherently trial-and-error.

PI, Tt pa T2 *

1. Calculate the total enthalpy of the feed stream at PI and TI. If it is a two-phase stream, the total enthalpy is found by adding that of the liquid and vapor phases.

2. Assume the unknown temperature T2.

3. Make a flash calculation at P2 and T2 to find relative amount and analysis of each phase. 4. Find the total enthalpy at Point Two from the above flash and the assumed T2. 5. If h2 = hl, you assumed the right temperature. If not, repeats Steps 2-5 until the h’s are

equal, within the desired limits of accuracy. Since the above is very tedious, a reasonably good answer usually can be obtained by assuming two different temperatures and plotting them on the following type of figure.

A straight line between the h2 - hl, found for two assumed temperatures, are connected by a straight line. The intersection at h2 - hl = O gives approximate true T2.

For the systems shown in Figure 11.2, the inlet gas stream is cooled by the exit separated gas before going to the choke. All one knows are the inlet conditions and composition and the sales gas limitations. These fix the pressure drop across the heat exchanger and choke in series. The drop across the former should not exceed 70 kPa [ 1 O psi]. The following type of procedure is needed.

(h2 -

T2 Temperature

1.

2. 3. 4.

5 . 6. 7.

8.

Assume temperature of gas downstream from gadgas exchanger. Run flash calculation at this temperature and inlet pressure, minus 70 kPa [lo psia]. Determine enthalpy of total stream at this point from composition in Step 2. Use the previous procedure for a choke to find the temperature in the low temperature separator. Run flash at separator conditions. Find the enthalpy of the vapor leaving the separator.

The AH of the sales gas must equal the AH of the inlet gas across the gadgas exchanger. If this is not found to be true, Steps 1-6 must be repeated. Once Step 7 is satisfied, the heat exchanger may be found by conventional heat transfer principles.

Note: A flash calculation is needed on the inlet gas to the gadgas exchanger if it is two-phase and composition and relative quantity of each phase is not known.

The above procedure illustrates the general conditions that must be satisfied for all systems where expansion across a valve is involved.

Expansion across a valve may be the proper choice over an expander but the temperature drop 7 is less and no useful work is produced.



Section 12

6"

r

CRYOGENIC GAS PROCESSING TABLE OF CONTENTS

PAGE # INTRODUCTION: CRYOGENIC GAS PLANTS ........................................ 12.1 ELEMENTS OF TURBOEXPANDER PLANTS ...................................................................................... 12.2

............................................................ 12.3 ............................................................... 12.9 ............................................................ 12.9 ............................................................ 12.1 1

............................................................ 12.13

..................................................................... 12.14

Gas Dehydration .......... .......... ~ ,......... .. ...................................................

. . . . . . . . . . . . .... . . . . . . . . , . . .. .

... ....................................

Recompressor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Piping and Vessels ....................................................

PRINCIPLES OF CRYOGENIC PROCESSING ...................................................................... Effects of Temperature and Pressure on Hydrocarbon Recovery ...................................................... 12.1 5 Latent Heat of Vaporization ........................................................................................ Theory of Gas Expansion .............................................. Heat Balance ........................................................................................................................

DESIGN CALCULATIONS ................... .................................. ............................................... 12.24 Expander Power ........ ............ .......................... .............. ......................... 12.24 Estimation of Expander Efficiency ,._..........._...... .......... ......................... 12.26 Other Performance Criteria ......... ...... 12.27 Expander Plant Design ......................................................................................................................... 12.29

THE TURBOEXPANDER ....................... Turboexpander Control ................................................. 12.3 1 Booster Compressor Co ................................................................. 12.34 Booster Compressor Surge Control ..................................................................................................... 12.34 Seals ... . . . . .. .. .. .. . . . . . . . . .. .. . . . . . . .. .. .. . . .. . . .... .. .................................................... 12.35 Thrust Control ............................................# ......................................................................................... 12.37 Lubrication System ............................................................................................................................... 12.38 Expander Compressor Safety Devic ....................... ....................................................... 12.40

EXPANDEWCOMPRESSOR OPERATION ............ ......._...... ........................................ ........................... 12.4 1 Startup Operations .................................................................................................................. Normal Shutdown ...... ......... .......... . ........... , ..................... , .................................................. Normal Operation ................................................................................................................................. 12.43

PRINCIPLES OF DEMETHANIZER OPERATION .................................................................................. 12.43 DEMETHANIZER AS FRACTIONATOR ................................................................................................ 12.44

Basic Demethanizer .............................................................................................................................. 12.45 Refluxed Demethanizer ........................................................................................................................ 12.48 Demethanizer Absorber ........................................................................................................................ 12.49

......... . ....... , .............................. . ...................................

................................. . ..........................................




Demethanizer Control ........ ................................................................................ 12.51 Ethane Rejection ...................... ................................................................................. 12.52

CRYOGENIC PLANT OPERATIONS ....................................................................................................... 12.54 Operating Objectives ....................................................................................................................... 12.54

................................... 12.57 Startup Operations .... ........................................................................................................ 12.57 Cold Expander Startu .....,............. ~ .................................. ................................... 12.61 Normal Operations .................................................................... .................................... 12.61 J-T Valve Operation (Expander Down) ............................................................................................... 12.63 Shutdown ............................................................................................................................................... 12.63

CRYOGENIC PLANT TROUBLESHOOTING ......................................................................................... 12.64 Freeze-Ups ................. ..........................................,................................................................12.65 Vapor-Locked Demethanizer Reboilers ....... ....................................................................... 12.67 Deinethanizer Temperature Control ........ ..... ........................................ 12.67 Turboexpander-Brake Compressor ................................ ........................................ 12-68 Snowballing Temperatures .. .... ... ........................................................................................ 12.68

. . . . , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . , . . . . . . . . . . . . . . . . . . . . . . .



TABLEOFCONTENTS

FIGURE # 12.1 12.2

12.3 12.4 12.5 12.6 12.7 12.8 12.9 12.10

12.11 12.12 12.13 12.14 12.15 12.16 12.17 12.18 12.19 12.20 12.21 12.22 12.23

12.24

12.25

12.26 12.27 12.28 12.29 12.30

LIST OF FIGURES PAGE #

Functional Block-Flow Diagram of a Cryogenic Plant ... Block-Flow Diagram of a Dehydration System Consisting of TEG Followed by Mole Sieve Adsorption ............................................................... Basic Plate Exchanger Design ...................... Plate-and-Fin Heat Exchanger Construction Corrugated Plate-Fins .......................................................................... .................... 12.7 Plate Core with Headers Attached .................................... Basic Components of a Brazed Aluminum Heat Exchan P&ID of Typical Expander Feed Separator Showing Instrumentation Details ......................... 12. I O Early Turboexpander Plant with Stabilizer Feed Separator and Feed Pump ............................. 12.12 Placing the Stabilizer Feed Separator Atop the Stabilizer Column Eliminates the

........................... 12.12 .................................................. 12.13

Methane and Ethane that Condense at Various Tempe nd Pressure .. 12.15 Pressure-Temperature Phase Envelope for an Expand .................. 12.16 Pressure-Enthalpy Diagrams for J-T Valve and Turboexpander Processes ............................... 12.20

Heat Balance in a Cryogenic Plant with 9% of the Inlet Flow Condensing to Heat Balance in a Cryogenic Plant with 12% of the Inlet Flow Condensi Efficiency of Expansion Turbines as a Function of Specific Speed ..........

............................................... 12.2

.................. 12.4 ................................................................. 12.6

................................................. 12.7

................................................. 12.8

Stabilizer Feed Pump ........................................................... Turboexpander Plant Utilizing Propane Refrigeration.

.........

Pressure-Temperature Phase Envelope for an Expander Process .......................

Expander Control System ....... ....................................................

er Compressor.. ..... 12.35 ................... 12.45

................................................................... 12.46 Gas is Chilled and Partially Condensed

..................................... 12.49

Basic Demethanizer with Process He and Side Reboilers .................................

Warm, High Pressure Inlet Gas By-P for Use as Absorption Oil in a Packed Absorption Column ................................ Bottoms Temperature Control and By-pass Butterfly Valves ..... ................... 12.51 Typical Process Flow Diagram - Cryogenic Expander Plant ( Typical Process Flow Diagram - Cryogenic Expander Plant (Two Expander Method) ........ 12.53 The Economic Benefit of “Pushing” Bottoms Product Purity Specification ........ 12.56

.......................................................... 12.66 Approximate Solid C02 Formation Conditions ..........

LIST OF TABLES TABLE # PAGE # 12.1 Labyrinth Seals in Expander-Compressor ................................................................................... 12.36



12.111


NOTES:

@John M. Cpmpbeii & Company

12.h BP Exploration Company (Columbia) Ltd.

Section 12

CRYOGENIC GAS PROCESSING INTRODUCTION: CRYOGENIC GAS PLANTS

Webster defines the word, ctyogenics, as being "the science that deals with the production of very low temperatures and their effects on the properties of matter." The cryogenic gas plant is built around a process that produces very low temperatures in order to affect the properties of natural gas, namely to cause certain of the gas's components to condense to liquid. The gas components which condense form a liquid hydrocarbon mixture known as Natural Gas Liquids or NGL. The NGL mixture contains a number of valuable hydrocarbons that can subsequently be separated into individual products, such as propane, gasoline and petrochemical feedstocks.

The gas processing industry has at its disposal a number of processes that are capable of removing NGL hydrocarbons from a natural gas feed stream. Each of these processes will have application when it can be shown to be the best process for a given situation.

The cryogenic process has application as the most economical means for recovering a high percentage of all hydrocarbons heavier than methane. Specifically, the cryogenic process aims to recover ethane and heavier hydrocarbons, such as propane, butane and gasoline components. Different variations of this process are capable of removing more than 85% of the ethane and essentially all of the heavier hydrocarbons found in produced natural gas. By contrast, other processes may be more appropriate when the goal is to recover just propane and heavier components. The main advantage of the cryogenic plant, in terms of recovered product, is therefore its ability to recover ethane or high propane recoveries.

f"

The heart of the cryogenic gas plant is a machine called a turboexpander or expander for short. The expander is a turbine that uses high pressure gas to turn a wheel on a shaft. The expander employs certain principles of thermodynamics to produce extremely cold temperatures, about -84 to - 109°C [-120 to -165'FI for typical cryogenic plants. As the gas passes through the turbine, its pressure is greatly reduced, causing a significant temperature drop. As the gas "expands" across the turbine blades it also does work in turning the shaft. When the gas does this work its internal energy is reduced, causing the resulting temperature to go even lower. The low temperatures produced are cold enough to condense NGL hydrocarbons, including ethane.

The turbine shaft transmits the work to a "load". The load in most cases is a compressor attached to the other end of the shaft, but it could also be a pump or generator. The usual attachment is a single-stage centrifugal compressor called a booster compressor or brake compressor. Its job is to impart a pressure increase to the low pressure gas, after the liquid NGL has been condensed and removed. This gas, called residue gas, is almost all methane (95% or more) and is sent to additional compressors for pipeline transportation or perhaps reinjection into the producing formation.

rc.



12.1


'7 The remainder of the cryogenic plant consists of a fractionation tower that stabilizes the recovered NGL by removing any dissolved methane, and several heat exchangers which make the process much more efficient.

The basic cryogenic plant is shown in the functional block-flow diagram, Figure 12.1. The gas that comes to the plant inlet is wet (saturated with water vapor) and rich (contains NGL hydrocarbons). It must first be dehydrated to remove practically all of the water vapor it contains, or ice (hydrate) will quickly plug up the cryogenic plant. This produces a gas feed to the coldest parts of the cryogenic plant that is dry and rich. As the gas flows through the "cold" plant the NGL is condensed and eventually pumped out as a liquid product. The residue gas is dry and lean; it goes through recompression where its pressure is increased sufficiently to get it into a gas transmission line.

Some cryogenic plants require additional, "external" refrigeration to condense NGL. This is usually done when the feed gas is exceptionally rich. The dashed block in Figure 12.1 represents a mechanical refrigeration system that is used for this purpose.

SALES GAS

RESIDUE GAS

NGL PRODUCT TO PIPELINE *

NGL PRODUCT PUMPS

Figure 12.1 Functional Block-Flow Diagram of a Cryogenic Plant

ELEMENTS OF TURBOEXPANDER PLANTS While there are many possible variations in the design of cryogenic NGL recovery plants, they

are all made up of just a few essential types of equipment. These major pieces of equipment may differ from plant to plant in terms of number employed, size and overall design, but their functions are always the same. The equipment types described below are therefore the basic building blocks of the turboexpander plant.

Gas Dehydration The produced gas that flows through the processing plant carries water vapor as part of the

overall mixture. The amount of water vapor present depends on both the pressure and temperature of @John M. Campbell & Company

12.2 ~~

BP Exploratlon Company (Columbia) Ltd.

ELEMENTS OF TURBOEXPANDER PLANTS

the gas, but we can be assured that for any given combination of pressure and temperature, the gas will be carrying as much water as it possibly can. We say that the gas is saturated with water vapor at its conditions of temperature and pressure. If this water vapor is allowed to remain in the gas as we cool it down in the cryogenicplant, it will either freeze to solid ice or form hydrates with other gas components, quickly plugging the piping and equipment through which it passes. This water vapor must therefore be removed through the process of dehydration prior to admitting the gas to the turboexpander plant.

Dehydration processes do not remove all of the water vapor held by the gas (this would be neither practical nor cheap). Instead, we use dehydration processes to remove the water to concentrations that are sufficiently low to avoid freezing and hydrate problems. Simply put, we dehydrate the gas to a water dewpoint temperature that is several degrees lower than the lowest temperature expected anywhere in the cryogenicplant. For instance, if the lowest temperature expected in the process is -90°C [-130"F], the gas will probably be dehydrated to a dewpoint at least as low as -96°C [-i41°F], and lower, if possible.

r

The extremely low water dewpoints required for the typical turboexpander process determine the type of dehydration process to be employed. The only real choice is an adsorption process using a solid desiccant such as mole sieve or activated alumina, or a combination of the two. The plant's total dehydration system may consist of mole sieve adsorbers alone or they may be preceded by a TEG absorber for the (overall less expensive) removal of the bulk of the water. Figure 12.2 is a block-flow diagram such a system.

An essential and critical piece of equipment in the adsorption dehydration unit are the mole sieve after filters or dust filters. Even though a mole sieve unit may be operated perfectly, the solid adsorbent will begin to crack and grind to dust after many regeneration cycles. If allowed to enter the cryogenicplant, this dust can plug heat exchanger passages and erode the vanes of the turboexpander's impeller. Operators must pay close attention to the condition and operation of the unit's dust filters.

r

Filter installations typically consist of two full-flow filters in parallel; one operates and one is in standby, so that the gas is always 100% filtered even when one filter is removed from service for cleaning and changing elements. Filter elements are typically of a cartridge-element type design that can be changed quickly. A filter is operated until its pressure drop reaches a predetermined level, at which time flow is switched through the spare filter and the old elements are changed. Operators must never by-pass these filters, even if it means shutting down the cryogenicplant. The price for by-passing the dust filters may include a turboexpander overhaul, a new aluminum plate-fm heat exchanger and considerably greater production losses due to the substantially longer downtime caused by these events.

The importance of adequate and reliable dehydration cannot be overstated. Poor dehydration is probably the number one cause of operational problems in the turboexpander plant.

Heat Exchangers In passing through the turboexpander plant, the inlet gas may have its temperature reduced by

some 146°C [263"F]. This assumes the gas enters the unit as warm as 38°C [IOO'F] and is cooled to temperatures as low as -108°C [-163"F] in the coldest regions of the unit. These temperatures indicate that large amounts of heat are removed in the cryogenicplant. In order to operate efficiently, cryogenic plants make extensive use of heat exchange equipment.

"P The primary heat exchanger in any cryogenicplant is the inlet gasíresidue gas heat exchanger. In this exchanger the warm inlet gas gives up its heat to the cold residue gas which in turn gets warmer. The overall design of this important piece of equipment may vary greatly from plant to plant,



12.3

I m


I I




and may exist as one, two or more "stages" for heat exchange. This is the case when multiple exchangers are used to exchange heat in several intermediate temperature ranges; the combined effect is that of one large exchanger with a number of inlets and outlets. Regardless of the design, however, it is common for up to 85% of the inlet gas cooling to take place in the unit's gadgas exchangers. (The remaining 15% takes place in the turboexpander(s).)

r'

Other cryogenicplant heat exchanger applications include mechanical refrigeration chillers, demethanizer bottom and side reboilers, and demethanizer reflux condensers.

Cold plant heat exchanger types are generally either the aluminum plate-fin or shell-and-tube designs. The choice depends on a number of factors but the materials of construction required for cryogenic applications place a major limitation on that choice. Carbon steel has a temperature limit of -29°C [-20°F] and even specially treated low temperature carbon steel can only provide service down to 4 6 ° C [-5O0F]. The operating temperature range of common, low temperature nickel alloy steels may be pushed as low as -59°C to -101OC [-75'F to -150"F], and other, less common, low carbon nickel alloys may take the range lower. At typical cryogenic NGL recovery temperatures the choice is usually between one of the stainless steels or aluminum. One advantage of using a series of heat exchangers operating in intermediate temperature ranges is the ability (sometimes) to specify less expensive materials in the "warmer" exchangers.

It is possible to make an aluminum shell-and-tube heat exchanger for cryogenic applications and some notable examples of such exchangers do exist: there is a vertically-installed gadgas exchanger in Qatar that stands as tall as its demethanizer tower! Stainless steel shell-and-tube heat exchangers represent the most expensive alternative. The focus of this segment is on brazed aluminum plate-fin exchangers. I"

Brazed Aluminum Plate-Fin Heat Exchangers

The most popular choice by far for cryogenic applications is the brazed aluminum plate-fin exchanger. The core construction of these exchangers is similar to a deck of cards standing on edge; that is, a stack of vertical plates. The plates are sealed along their edges by brazing to end pieces. The core is usually suspended by its top and is allowed to hang freely, thereby permitting unrestricted thermal expansion and contraction. Flow through the exchanger is generally vertical (horizontal orien- tations are sometimes preferred), with the hot end at the top and the cold end at the bottom. (See Figure 12.3) Flow patterns are shown in Figure 12.4; they may be either cross-flow (perpendicular), counter-flow, a combination of cross/counter flow or complex. Any given stream may be designed to make multiple passes.

The space between the plates is filled with accordion-like "fins" of various designs the give an edge-on appearance similar to a cross section of corrugated cardboard. (Figure 12.5) The fins effectively increase the overall heat exchange area and help to insure the turbulent flow that is necessary for efficient heat transfer.

Collection and distribution headers are welded on for each stream that passes through, as shown in Figure 12.6. The completed exchanger is either covered with insulation or housed in a "cold box" filled with loose insulation such as perlite. The method used depends on the size and the operating temperature. Cold boxes are usually purged with an inert gas like nitrogen to create a positive pressure that excludes moist outside air and prevents ice formation inside. This practice also provides a convenient place to check for hydrocarbon gases, the presence of which indicates the existence of leaks.

T-



12.5


I

I

Figure 12.3 Basic Plate Exchanger Design

l A a- - - -,I Headers

Separator .- J A

Plate

Crossflow-Type Core

Two-Stream Counterflow Type

B +

Cross-Counterflow Type

I/YA

Complex: Two Reversing Streams in a Five-Stream Exchanger

Ref.: Lowe, Roderick E., Chem. Engr., (Aug. 17,1987), p. 131

Flgure 12.4 Plate-and-Fin Heat Exchanger Construction Types



ELEMENTSOFTURBOEXPANDERPLANTS

r"

r"'

Lanced Fins

Figure 12.5 Corrugated Plate-Fins Courtesy: Stewart -Warner

I

Stewart -Warner I

Figure 12.6 Plate Core with Headers Attached

Process streams may be either liquid or gas and may involve phase changes such as the vaporization of a liquid or the condensation of a vapor; plate-fin exchangers (or sections thereof) may therefore be used as chillers, reboilers or condensers. These exchangers cover the full potential for transfer- ring both latent and sensible heats.

Brazed aluminum plate-fin exchangers have higher efficiencies for their size and weight than do shell-and-tube types. The cost per square foot of heat exchange surface area is lower than that for carbon steel shell-and-tube exchangers, which, which because of lower heat exchange coefficients, require comparatively more surface area to do the same job. The low temperature properties of alumi-

P



12.7


-I num are such that the metal gets stronger as the temperature goes lower and high pressure (up to 96 barg [ 1400 psig]), low temperature applications are not a problem.

Plate-fin exchangers are often constructed in complex designs that provide simultaneous heat exchange between as many as 8 or 10 different streams. (Figure 12.7 shows a three-stream exchanger.) The result of doing so is the creation of a complex network of simple exchangers, in parallel andor in series, that would be impossible to duplicate with shell-and-tube exchangers. This practice yields much higher heat exchange efficiencies in the process unit overall and has the practical benefits of reducing both capital costs (fewer exchangers) and operating costs (heating and cooling utilities). Complex exchangers represent an interesting design problem for engineers and can often confuse even the most experienced operators who try to follow process lines through the plant, but they are worth it.

Figure 12.7 Basic Components of a Brazed Aluminum Heat Exchanger

@John M. Campbcil US Company



r"

r"

There are a few disadvantages of brazed aluminum plate-fin exchangers that operators must be aware of:

1.

2.

3.

Separators

Because of their generally narrow passages, these exchangers can plug up with solids easily. Operators must ensure that mole sieve dust filters are well-maintained. Hydrates and solid C02 can also easily plug the exchanger so they are often installed with methanol injection points on their inlet ends. Hydrates and solid CO2 may be melted but an exchanger plugged with solids ends up on the trash heap. Differential pressure (AP) indicators and alarms are an operational necessity. The brazing material used in these exchangers is a solder-like material that has a fairly low melting point. The upper safe operating temperature limit for these exchangers in hydrocarbon recovery processes is 65°C [ 15O"FI. Operating procedures covering startup, pres- surization, normal operation, deriming and dryout must all address this issue and be consistent with the manufacturer's recommendations. Be aware that the mole sieve dehydration unit has the potential for sending hot gas to the cryogenicplant in at least two ways: leaking valves can pass hot regeneration gas and freshly regenerated towers can switch into drying service before cooling sufficiently. These exchangers are less resistant to corrosive fluids than are shell-and-tube exchangers and may be highly susceptible to attack by gaseous mercury in the inlet gas stream. Mer- cury attack is a topic of increasing concern and research. Liquid mercury can severely attack aluminum (and brazing materials) in the presence of free water but these are conditions that almost never occur in the cores of operating cold plant exchangers. Instead, mercury failures have occurred in high magnesium content pipe welds and alloy headers where liquid mercury had been trapped for some time (years) at temperatures above -39°C [-38"F], the freezing point of mercury. Inlet gas is sometimes treated for mercury removal; the mercury is usually chemically trapped-out in an upstream mercury removal unit (MRU) consisting of an adsorber bed of sulfur-impregnated activated carbon. UOP offers a Type X or Y molecular sieve material with silver and polysulfide to provide up to 99% mercury removal. Other successful approaches have been to apply metal overlays in critical exchanger areas, to conduct thermal treatment of welds and alloy parts and to apply protective coatings.

Turboexpander plants require one or more separators. Hydrocarbon liquids will condense through the various plant sections as the feed stream gets colder. Separators collect these liquids and then dump them to the demethanizer tower at various feed locations which correspond to the expected composition of each liquid stream. Perhaps the most important function of such a separator is to prevent liquids from entering the turboexpander inlet, where they can do damage. Expander feed separators are equipped with mist eliminator and high liquid level alarms and shutdowns to prevent such an occurrence. Figure 12.8 is a section of a P&ID drawing that shows a typical turboexpander feed separator and its instrumentation.

Expander Compressor The expander is a single-impeller turbine similar to a steam turbine. It is a free-wheeling

machine that reduces the pressure and temperature of a gas stream and converts the gas's pressure energy into useful work. The work is extracted by a direct-coupled centrifugal compressor that recom- presses the gas stream after condensed NGL liquids have been extracted. (The expander and compressor wheels occupy opposite ends of a common shaft.) This combination of expander and compressor



~

12.9


q , !L

5 N

n

" u



ELEMENTSOFTURBOEXPANDERPLANTS

in a single machine is known by a variety of names, with Expander/Compressor, ExpanderBrake-Corn- pressor and Expander43ooster-Compressor being the most common. r"

The expander/compressor is the heart of cryogenic plant. It is discussed in great detail in a later segment.

ider/Compressor

I -T Valve

The J-T valve is a required part of the turboexpander plant, even though it is rarely used. It is a rather ordinary control valve used to drop the pressure and expand the volume of a high pressure, NGL-rich hydrocarbon gas. The combination of reduced pressure and increased volume results in temperatures low enough to cause NGL condensation. This cooling phenomenon is known as the Joule-Thomson or (J-T) effect, named for the physicists who first explained and applied it.

t-

The combination of gas expansion and work extraction that takes place in an expander results in much lower temperatures and hence greater liquid condensation than can be achieved with a J-T valve (the J-T valve does no work). All expander installations include a J-T valve mounted in a parallel to the expander, however, as J-T valve operation is necessary for:

1. Plant startup operations,

2. Handling gas flows in excess of expander capacity, and

3. For continued operation during those times when the expander is down for maintenance.

Demethanizer

Early demethanizers were actually just top-feed NGL stabilizers used to cook-off unwanted light ends. A stabilizer produces a vapor overhead product and a liquid bottom product that has had its content of unwanted light hydrocarbons reduced; we call this liquid a stabilized NGL. A stabilizer was used in conjunction with a stabilizer feed separator and stabilizer feed pump. Later modifications eliminated the need for the feed pump by placing the separator (the so-called "separation bell") directly on top of the stabilizer column, resulting in the familiar silhouette of today's demethanizer towers. (Compare Figures 12.9 and 12.10) A trend since the early 1980's has been to install trays inside the separation bell and to add reflux to the top of the tower for the purpose of improving ethane recovery. Demethanizer towers and internals are usually made of stainless steel.

I/"

BJobn M. Campbell & Company


12.11


OVERHEAD VAPORTO RECYCLE

DRY GAS FEED

EXCHANGER SEPARATOR

I STABILIZER FEED WW

NQL STABILIZER

PRODUCT i--F&* PUMP

Figure 12.9 Early Turboexpander Plant with Stabilizer Feed Separator and Feed Pump

RESIDUE

d

RECOMPRESSOR J-TvAL% I

DRY GAS FEED FROM MOLE SIEVE EXPANDER I

COMPRESSOR

SEPARATOR EXCHANGER

y - DEMETHANIZER

NGL PRODUCT

PRODUCT 6 PUMP OUT

Figure 12.10 Placing the Stabilizer Feed Separator Atop the Stabilizer Column Eliminates the Stabilizer Feed Pump




The stabilizer portion of the vertical demethanizer column usually contains 20 to 30 trays (usually valve type trays) or equivalent packed sections. Liquids falling from the separation bell flow downward, contacting the warmer stripping vapor that rises from the bottom of the column and reboiler. (The reboiler is a heat exchanger that provides the process heat necessary to create stripping vapor.) In this way any dissolved methane will vaporize and flow upward and out with the tower's overhead vapor. The NGL liquid continues downward to the bottom of the column and is removed as "demethanized" C2f NGL product. The residual amount of methane in this product is controlled by adjusting the heat input to the reboiler, and hence the "boil-up" rate of warm stripping vapor.

r"

There are wide variations possible in demethanizer installations, some of which are proprietary designs that claim to increase NGL recovery, reduce process heat requirements or offer the flexibility to recover or reject ethane as part of the NGL product under a variety of conditions.

The principles of demethanizer operation fall under the heading of the mass transfer operations of stabilization or fractionation (refluxed demethanizers can certainly be classified as fractionators). The subject of demethanizers and their principles of operation are examined more closely in upcoming segments of this manual.

Refrigeration System

When the feed gas stream is especially rich in ethane and heavier hydrocarbons, extra cooling is sometimes needed to maintain NGL recovery effectiveness. (This is because there will be relatively larger amounts of latent heat to be removed and expander operation will be colder and more efficient in recovering ethane if the heavier hydrocarbons are condensed earlier on.) The extra cooling is provided by flowing the feed gas through a chiller. The chiller is part of a conventional mechanical refrigeration system as described elsewhere in this manual. The chiller is always located somewhere upstream of the coldest gadgas exchanger, and can be counted on to reduce the feed stream's temperature to around -40°C [-40"F] with propane refrigerant. This is shown in Figure 12.1 1.

r)"

$J-z&ucT PRODUCT PUMP

rc' Figure 12.1 1 Turboexpander Plant Utilizing Propane Refrigeration



12.13

U


OC -29 -46 -59

-101 -195

-250

Recompressor /c7 One disadvantage of the turboexpander NGL recovery unit is that it produces a low pressure

residue gas. The pressure of this gas must be increased if it is to enter a sales gas line or be reinjected. A residue gas recompressor is therefore a necessary companion of the turboexpander.

Recompressors may be either reciprocating or centrifugal machines, depending on many factors unrelated to the operation of the cryogenicplant. Gas compression is not cheap ... it represents one of the largest costs involved in building and operating such a plant.

O F -20 -50 -75 -150 -320

4 2 5

Piping and Vessels Equipment used in cryogenic service must be constructed of materials that can withstand the

low temperatures involved. Ordinary carbon steel is not suitable because it would become brittle enough to shatter under such conditions. Regular carbon steel (0.20 to 0.35% C) retains its strength only down to -29°C [-20°F], and low-carbon steel (0.05 to 0.25% C) can be specially treated to withstand temperatures down to 4 6 ° C [-5O0F]. At lower temperatures, nickel must be added to the low-carbon mix to make a suitable metal, and ratings down to -195°C [-320°F] are possible. Nickel and chromium are the main additives used in the production of various grades of stainless steels, which have ratings down to -250°C [425"F]. Various grades of 304 stainless steel contain 18 to 20% chromium and from 8 to 10.5 or 12% nickel, while 316 stainless consists of 16 to 18% chromium, 10 to 14% nickel and 2 to 3% molybdenum. Both include lesser quantities of other elements: 2% manganese and 1 % silicon, but less than O. 1 % carbon.

Material selection for cryogenic applications is based on factors other than temperature rating alone; the overall cost, availability, ease of fabrication and ease of welding must be considered. This is why, when all is said and done, 304 stainless steel is such a popular choice, even though it may have a higher original cost than some other materials.

Vessels and piping must be made of the proper metal to withstand the temperatures to which they are exposed. This is a particular concern when managing change in the process plant. When new piping or fittings are installed, care must be taken to insure that the new parts can withstand the low temperatures and are also compatible with the metals to which they attach. Compatibility is especially critical when new joints are being welded.

his

The following table is a short list of some low temperature materials:

Acceptable Metal Carbon Steel Low Carbon Steel, Charpy Tested Low-Carbon, 2.5% nickel steel Low-Carbon, 3.5% nickel steel Low-Carbon, 9% nickel steel 304 or 316 stainless steel


12.14 ~ ~


PRINCIPLES OF CRYOGENIC PROCESSING

r PRINCIPLES OF CRYOGENIC PROCESSING

Effects of Temperature and Pressure on Hydrocarbon Recovery When we chill a gas to condense NGL, we hope to extract or recover the maximum amount of

valuable NGL. However, a substantial amount of methane will also condense to liquid and become part of the mixture. The methane is undesirable for two reasons: first, the methane has zero value as an NGL component and second, methane elevates the vapor pressure of the NGL mix and may cause difficulties in liquids transportation. Therefore, almost all of the methane is removed from the NGL in the demethanizer tower.

There are costs associated with operating a demethanizer; these include the capital cost of the equipment and its operating cost, which includes several factors. Since methane must be rejected from the NGL, it stands to reason that the cost of demethanizing will be less if the amount of methane condensed is less. Thus, one criteria in the design and operation of the turboexpander plant is to minimize the amount of methane that liquefies while still maximizing the recovery of ethane and heavier hydrocarbons. The quantity of methane that can be removed in the demethanizer depends upon its design. It will have been designed for some maximum case, where the maximum amount of methane condensed and rejected corresponds to the maximum recovery of ethane.

Turboexpander plants can usually recover 98+% of the propane and 100% of the butane and heavier hydrocarbons from the feed gas stream. Ethane recoveries vary according to the design of the plant and its operational objectives, but one thing is certain: as ethane recovery targets go higher, so does the difficulty in preventing excessive methane condensation. This fact is illustrated by Figure 12.12.

SI UNITS -40

O 20 40 60 80 I00

%Methane or Ethane That Condenses

-20

-40

-60

-80

LL e g -100 e im

-

-120

-140

-160

-180

ENGLISH UNITS

O 20 40 60 80 100

%Methane or Ethane Thai Condenses

Figure 12.12 Methane and Ethane that Condense at Various Temperature and Pressure



12.15

CRYOGENiC GAS PROCESSING

7 The quantity of methane which condenses along with the NGL depends upon the pressures and temperatures at which the system operates. The greatest amount of methane condensation will occur when the pressure is high and the temperature is low. As shown in Figure 12.12, less methane will condense, even at lower temperatures, if the pressure is lowered. At the same time, the recovery of ethane is practically unaffected. (Compare the illustrated example with the results at -102°C and 1380 kPa [or -140°F and 200 psia].)

The disadvantage in operating a cryogenic plant at lower pressures is the increased cost of recompressing the residue gas. At some point, the additional cost of compression will offset any benefits realized at the demethanizer by operating at lower pressures. The optimum operating point is therefore that combination of temperature and pressure where ethane recovery is maximized and the combined cost of compression and demethanizer operation is at a minimum.

Operators should therefore not arbitrarily change the operating pressures in the turboexpander plant. They should instead follow the guidelines established by the plant engineer, who is usually the person responsible for unit optimization. Having said this, the reader must realize that the job of optimization is a complicated one and, if done correctly, will involve plant performance tests as well as detailed calculations and simulations. More will be said on this topic in the section on demethanizer operation.

latent Heat of Vaporization The turboexpander process condenses hydrocarbon liquid from the feed gas stream by suffi-

ciently lowering the temperature of that stream. The thermometers and other temperature indicators around the unit indicate that heat is successfully being removed from the gas, but these tell only part of the story, since they indicate only sensible heat. In the process of condensing a hydrocarbon vapor to ,--#

Critical Temp.

Temperature (F)

I Courtesy: GPSA Databook

Figure 12.1 3 Pressure-Temperature Phase Enve- lope for an Expander Process

I

liquid, the vapor’s latent heat of vaporization must also be removed. Figure 12.13 helps to illustrate this point.

The figure shows, among other things, the P-T phase envelope for the plant inlet gas. Point #1 on the diagram corresponds to the temperature and pressure of this gas feed. It is obvious that the gas contains superheat at these conditions. As the gas passes through the plant, its temperature is reduced at essentially constant pressure, as shown by the dashed line. As the temperature is reduced, the heat that is removed is sensible heat, until the gas’s dewpoint temperature is reached.

Continued cooling of the feed gas after reaching its dewpoint, results in liquid condensation because the two-phase region of the phase envelope has been entered. This means that more and more of the gas will condense to liquid as the temperature drops. Now both sensible heat and latent heat are being removed as the remaining vapor continues to be cooled and increasingly condensed. The condensed liquids are further cooled as they pass

through the remaining heat exchangers, so the -?




process removes sensible heat from these liquids, too. Cooling in the plant's heat exchangers continues until point #2 is reached. The dashed-line phase envelope is that of the turboexpander feed separator vapor. (We will study the rest of this diagram later.)

r'

As operators know, we cannot directly gauge the removal of latent heat because it cannot be directly measured with an instrument like a thermometer. The level of hydrocarbon recovery is none- theless affected by our ability or inability to remove the required amount of latent heat. Operators must therefore have some way of determining whether or not latent heat removal is adequate. Avail- able methods of doing so are indirect and qualitative (rather than direct and quantitative, as is the measurement provided by a thermometer's temperature reading), but it does give the operator a general sense of what is happening in so far as latent heat removal is concerned. This situation is best illustrated with an example.

We begin with the assumption that the average latent heat of vaporization for a mixture of hydrocarbon liquids passing through the cryogenicplant is about 465 kJkg [200 Btdlb]. This is the latent heat that must be removed from each kg [lb] of hydrocarbon vapor to cause condensation of a kg [lb] of liquid. On the other hand, the average amount of heat removal required to cool the feed gas stream by 1 degree Celsius [Fahrenheit] as it passes through the plant is about 2.72 H k g [0.65 Btdlb] (this is the average heat capacity of the gas mixture). We first remove a certain amount of sensible heat to cool the hydrocarbon vapor to its dewpoint temperature. Then, we continue cooling to remove the hydrocarbon's latent heat of vaporization to make it condense to a liquid. Consider the following example, which is a rather crude estimation:'

r" In Example 12.1, 16% of the total heat removed from the gas is latent heat. In cryogenic plants, the quantity of liquid that condenses depends upon the amount of latent heat removed. If the total heat removal is less, less liquid will condense. For instance, if the heat removal was only 99% of that in the example, the amount of liquid condensing would be 14.8% of the feed gas stream. This would correspond to a reduction in the amount of liquid recovered of 1.3%. (0.2/15 = 1.3%) A plant capable of making 10,000 kg/hr [22,000 l b h ] of NGL would only make 9,870 kg/hr [21,714 lbíhr] in such an instance.

As a practical matter, there are two things operators can observe to gain clues about their plant's ability to remove the required amount of heat, provided the feed gas composition is constant:

1. The ratio of liquid produced per volume of gas processed - This ratio can be expressed in a number of ways, depending upon the preferred units of measurement. The most common units used in the English system are gallons of liquid per 1000 scf of gas, or simply "gallons per thousand", GPM. (The G in GPM stands for gallons and the M is the Roman numeral for one thousand. Not to be confused with "gallons per minute".) In the SI system, the most common units are m3/1000 sm3, that is, cubic meters of liquid per 1000 standard cubic meters of gas.

Operators can calculate this ratio easily by plugging in numbers from daily production reports (if not already calculated by reporting software). This ratio should be checked against recoveries calculated for both design and actual feed gas compositions. The reason follows:

1 The calculation shown here is only approximately correct as condensation occurs continuously from the inlet of the first exchanger on through the expander. As these liquids condense, the composition of the remaining vapor continually changes, so the vapor's heat capacity and the latent heats of vaporization of the condensing liquids are also changing. Additionally, the composition of the condensed liquid continuously changes, as does the liquid mixture's heat capacity. Detailed calculations for this type oí problem do not rely on using "average" values for heat capacities and heats of vaporization, but are instead calculated anew for each incremental change in composition. Such calculations are very tedious without the aid of a computer.

f-



12.17


Example 12.1 A "typical' cryogenic plant feed gas stream enters the plant at 35°C [95"F] and will be cooled to a temperature of -101"C[-150"F]. In so doing, 15% of the feed gas stream (by weight) will condense to liquid. What portion of the heat removed is sensible heat? Latent heat?

SI Solution Heat removal required to cool by 1 "C (avg. feed gas heat capacity): Inlet gas temperature: 35°C Chilled gas temperature: -101°C

Weight of gas (unit weight):

2.72 k J k g

Total temperature reduction: 35°C - (-101°C) = 136°C 1 kg

Sensible heat removed (kJ): (2.72 kJ/kg*"C)( 1 kg)( 136"C)= 370 kJ Average hydrocarbon latent heat of vaporization: Weight of condensed liquid (15% of gas): Required latent heat removal (kJ): Total heat removal (sensible + latent heats):

FPS Solution Heat removal required to cool by 1 "F (avg. feed gas heat capacity): Inlet gas temperature: 95°F

465 k l k g (0.15)(1 kg) = 0.15 kg (0.15 kg)(465 kJ/kg) = 70 kJ 370 kJ + 70 kJ = 440 kJ

% of total that is latent heat: 70/440 x 100 = 16%

0.65 Btu/lb

Chilled gas temperature: -1 50°F Total temperature reduction: 95°F - (-150°F) = 245°F Weight of gas (unit weight):

Average hydrocarbon latent heat of vaporization: Weight of condensed liquid (1 5% of gas): Required latent heat removal (Btu): Total heat removal (sensible + latent heats): %O of total that is latent heat:

1 Ib

200 Bidlb (0.15)(1 Ib) = 0.15 Ib (0.15 Ib)(200 Btu/lb) = 30 Btu 159 Btu + 30 Btu = 189 Btu 30/189 x 100 = 16%

Sensible heat removed (Btu): (0.65 Btdlb-OF)( 1 lb)(245"F) = 159 Btu

Liquid recovery is not dependent upon operating temperatures and pressures alone, but also upon feed gas composition. If the feed gas gets "richer", that is, contains proportionally higher amounts of NGL hydrocarbons and lesser amounts of methane, NGL recovery levels will tend to go higher, even at warmer temperatures. And a gas that gets "leaner" will have to be chilled to lower than normal temperatures to meet the recovery targets.

2. Turboexpander outlet temperature - The turboexpander outlet temperature provides a more direct indication of plant performance when the feed gas composition and plant oper- atingpressures are fairly constant. An increasing temperature, for instance, may be due to warmer inlet gas or a reduction in heat exchanger efficiencies.

The plant equipment responsible for heat removal are its heat exchangers, the turboexpander and the mechanical refrigeration system (if used). In order to maximize liquid production, all of this equipment must be operating up to par. One characteristic of the turboexpander plant is its reliance on heat exchangers; most of the heat removal occurs in its various heat exchangers. If one heat exchanger in a series fails to perform as it should, the other exchangers and the expander may all be adversely affected (expander outlet temperature will go higher if warmer gas comes to its inlet), and the overall heat removal and liquid recovery would be less.

Again, all other things being equal, feed gas composition will have a visible effect on expander outlet temperature, with richer feeds tending to cause a temperature increase and




r leaner feeds causing a temperature reduction. Operators must therefore inspect all available data in order to correctly interpret a change in expander outlet temperature.

Because of their reliance on heat exchangers, a typical cryogenic plant (that does not use mechanical refrigeration) will not possess any kind of "excess cooling capacity". This means that the system is self-limiting according to the capabilities of its heat exchangers. It is extremely important that the plant's exchangers operate properly; the slightest bit of fouling, due to dirt, mole sieve dust, freeze-ups or other problems, can result in reduced NGL recovery.

Turboexpander plants with mechanical refrigeration (chillers) do offer some flexibility in this regard. The system's response to changing heat loads will be "automatic" as long as the refrigeration system is operated to provide a constant chiller shell pressure. As the heat load on the chiller fluctuates, it will cause greater or lesser refrigerant vaporization (and pressures) in the chiller shell. System controls will cause the refrigerant compressor to load or unload in an effort to maintain chiller shell pressure. In this way the chiller will act as a "swing" exchanger and minor fluctuations in the feed gas composition or temperature can be adjusted for. All the other exchangers in the plant must still work as they should, however, to prevent unnecessary refrigeration loads and wasted horsepower.

Theory of Gas Expansion We have seen how turboexpander plants remove both sensible and latent heat from a feed gas

stream to condense hydrocarbon liquids. Any gas will possess both sensible heat and its latent heat of vaporization. Heat is a form of energy, so a gas contains heat energy. A gas may contain other forms of energy as well, such as pressure energy and velocity. The total energy, of all kinds, possessed by a gas is called its "internal energy". Another name for this internal energy is enthalpy.

r*" In gas processing, the enthalpy of a gas, or a mixture of gases, is primarily dependent on the

pressure and temperature of the gas. (The velocity (or kinetic) energy of a gas changes little as the gas moves through the plant, so it is considered to be negligible and therefore can be ignored in the discussion which follows.) Different gases will have different enthalpies at the same temperature and pressure (because the different gases have differing latent heats of vaporization and specific heats), so the enthalpy of a gas mixture is also, therefore, composition-dependent.

One of the laws of classical physics states that energy can be neither created nor destroyed. So if the internal energy of a gas is increased or reduced, either the gas's pressure or temperature or both must also change. The relationships among a typical gas's enthalpy, pressure and temperature are illustrated in Figures 12.14. The vertical axis on the charts represent pressure and the horizontal axes represent enthalpy. Note that the units of enthalpy used in these charts are Kilowatts per lo6 m3 .(of gas flow) per day (SI units) or horsepower per MMscfd (FPS units). Lines of constant temperature on these charts curve downward from upper left to lower right.

Another set of lines, called the "compression or expansion lines", curve upward from lower left to upper right. These lines describe how a gas's internal energy increases when compressed or how it decreases when allowed to expand. (For those who are familiar with the term, these are lines of constant entropy.) We can use these charts to understand (and compare) what happens when a gas expands through both a J-T valve and a turboexpander. This is demonstrated by the following example.

A result corresponding to that in the example can also be described with a P-T phase envelope. Consider Figure 12.15, which is a repeat of Figure 12.13. From point #2 on the dewpoint curve of the expander feed gas, there are two lines, #3 and #4, which compare the results of a turboexpander expansion with those of a valve expansion. On the P-T diagram, lines of constant enthalpy would run

íc'



12.19


7000

8000

5000

4500

3500

3Ooo

2500

2000

1500

1W

5000

4500

4Ooo

3500

3Ooo

2500

1500

O 250 500 750 lo00 1250

1000

800

800

700

800

500

3 450

gw

350 B u 300

250

200

150

A loo0

w)o

800

700

800

500

450

400

350

300

250

200

150 O 10 20 30 40 50

Enthalpy, kW per I O ' m3/d Enthalpy, HP per lo6 tt"/d I Figure 12.14 Pressure-Enthalpy Diagrams for J-T Valve and Turboexpander Processes

Example 12.2: Gas passing through a cryogenic plant has been cooled to -40" in the gadgas exchangers. Its pressure is 7000 kPa [lo00 psia]. (These conditions are represented by point "A" on the respective curves.) If the gas is allowed to drop in

pressure to 2500 kPa [350 psia], by simply passing through a valve, it will end up at point "C" on the charts. This is because the pressure reduction valve does no work on the gas; meaning that the internal energy of the gas is unchanged. This is exactly what happens in a J-T valve; the expansion is called a constant-enthalpy or isenthalpic expansion. The gas ends up colder because of the pressure-temperature relationship, but its internal energy is unchanged. This is an illustration of the J-T Effect. The new conditions of the gas at point "C" are 2500 kPa [350 psia] and -77°C [-107"F]. Significant hydrocarbon condensation will occur at these conditions, but not as much as would condense at colder temperatures.

Suppose instead that the gas at point "A" flows across the blades of a turboexpander. The expander is connected to a compressor. It takes energy to turn the compressor and this energy is provided by the expanding gas. We say that the gas does work when it expands in the turboexpander. As a result, the gas ends up at the conditions corresponding to point I'B" on the charts. In this case, the expansion follows the "compressioníexpansion lines" (these are lines of constant entropy corresponding to an isentropic expansion). Now at the same pressure, 2500 H a [350 psia], the gas is much colder, -99°C [-146"F], because the gas did work; that is, its enthalpy was reduced. The reader can compare the enthalpies that correspond to conditions at points "B" and "C." These

' figures illustrate the basic difference between a J-T valve and an expander.


12.20 ~ _ _ _ _ _ _ ~ ~



Critical T ~ D .

Temperature (i) Courtesy: GPSA Databook

Figure 12.1 5 Pressure-Temperature Phase Envelope for an Expander Process

parallel to line #2-4; this is, therefore, the line describing the valve expansion. Lines of constant entropy, if plotted on this diagram, would run parallel with line # 2-3; this is, therefore, the line that describes the turboexpander expansion. Compare the positions of the endpoints of these two lines. It is obvious that the turboexpander expansion ends up cutting across more of the envelope's quality lines and the amount of liquid condensed is greater.

r""

Another characteristic of J-T valve expansion versus expander/compressor operation is evident in Figure 12.15. Because the expander has its companion booster compressor, it can expand the gas to a lower pressure than the J-T valve can. The reason has nothing to do with the design of the J-T valve, but rather with the operation of the recompressor. The expanders booster compressor raises the residue gas pressure to the required recompressor suction pressure. In order for the J-T valve to provide the same recompressor suction pressure, it's operating pressure drop must be less.

The compressiodexpansion curves illustrated in Figures 12.14 show the maximum amount of energy that can be removed from the gas when its pressure is reduced from point "A" to point "B". The same amount of energy would be required to raise the pressure of the gas from point "B" back to point "A". We might expect the gas discharged from the brake compressor would be at the same pressure as that at the expander inlet and this is true, but only in the theoretical case. In the real world, some of the gas's energy is lost to friction and heat transfer as it passes through the machine. Real expanders are 75 to 85% efficient in converting the gas's internal energy to useful work, and the compressor is about 65 to 80% efficient in working on and raising the pressure of the residue gas, so its discharge pressure will only be a fraction of the original expander inlet gas pressure.

If gas flowed through an expander that was not connected to a shaft load like the brake compressor, the gas expansion would be just a J-T expansion and it would get no colder than if it had gone through a J-T valve. When attached to a compressor and the compressor is loaded fully, the maximum amount of energy is removed from the gas and the temperature will be at its lowest. If, however, the

f-



12.21


compressor is partially unloaded, the temperature at the expander outlet will not be as cold, and as the compressor is unloaded even more, the expander outlet temperature will rise to nearly that of the J-T valve. Therefore, in order to get the lowest possible temperatures, the brake compressor must be fully loaded.

Figures 12.14 indicate the theoretical temperature of gas leaving an expander at a known pressure, when the temperature of the gas entering the expander is also known. In actual operation, the temperature of the expander outlet will be slightly higher than that shown on the graphs because the stream leaving the expander is two-phase, gas and condensed liquid. The latent heat of vaporization given up by the condensing liquid goes into raising the temperature of the total stream above that predicted on the graphs. However, a rough estimation of expander outlet temperature can be made using the following relationship: each 1 percent of gas that liquefies in the expander will raise the temperature above that shown on the graphs by about 1.7"C [3"F].

~~ ~

Example 12.3: 5% of the gas condenses to liquid when its pressure is lowered as shown in going from point "A" to point "B" in Figures 12.14. Determine the actual temperature of the gas leaving the expander.

Percent of gas that condenses: 5%

Percent of gas that condenses: Percent of gas that condenses:

5% 3 x 5 = 15'F

Heat Balance

We have said elsewhere in this manual that only about 15% of the total gas cooling takes place in the expander. However, it is the expander that drives the cooling for the whole unit while the heat exchangers just tend to "multiply the effect" of that cooling. The importance of the expander can be demonstrated from a "heat balance" point of view.

The principle of heat balance is quite similar to the principle of material balance, in that "what comes in must equal what goes out". In the case of the turboexpander plant, we can look at the process as a whole (i.e., draw a boundary), such that only one stream enters, the feed stream. Two outlet streams, then, cross the plant boundary, the residue gas and the extracted liquid NGL. This is shown in Figure 12.16. (Heat that transfers within the plant boundary need not be considered because its algebraic sum is zero.)

The figure shows a comparison of the relative heat contents of these three streams. Since energy can't be created nor destroyed, the heat energy content of the feed stream must be equal to the total heat energy contents of the two streams leaving the plant. (Note that we have arbitrarily chosen the stream heat contents to be equal to percentages of the total heat content that gets transferred in the process. 100% of this heat comes in with the feed and 0% of this heat leaves in the NGL. Therefore, the residue gas must carry away an amount of heat energy equal to that transferred from the inlet stream.)

/c7




The liquid NGL product entered the plant as a vapor. The net effect of processing the gas was to change the phase of the NGLs from gas to liquid. The net heat energy removed from the inlet stream in the plant is the latent heat of vaporization of the hydrocarbons in the liquid product.

The expander/compressor is an energy transfer device. Heat energy removed from the gas in the expander transfers to the gas in the compressor in the form of increased pressure and temperature (less efficiency losses). The degree of liquid condensation is directly proportional to the percentage of energy transferred in the expander/compressor.

In the heat balance shown in Figure 12.16, the expander removed 9% of the heat energy contained in the inlet gas and the result was that 9 volume % of the feed gas liquefied. I f the expander had removed only 8% of the heat energy, the quantity of liquid product would reduce to about 8 volume YO and the quantity of residue gas would increase to 92 volume 'YO. Thus, the quantity of extracted product varies directly with the amount of energy transferred in the expandedcompressor.

GAS TO BOOSTER COMPRESSOR 91 volume %

Heat Content = 91 %

GAS OUT OF BOOSTER COMPRESSOR 91 volume %

Heat Content = 100%

INLET GAS 100 voiume %

Heat Content = 100% EXPANDER Heat Energy Transferred = 9%

9 vdume % Heat Content = 0%

Figure 12.16 Heat Balance in a Cryogenic Plant with 9% of the Inlet Flow Condensing to NGL

The previous heat balance is based on a volume of liquid product equal to 9% of the volume of inlet gas. Suppose we want to increase the volume of product to 12% of the inlet gas. In this case, the energy transferred in the expander/compressor will have to increase from 9 to 12% of the heat energy contained in the inlet gas. However, the amount of energy transfer which can occur in the expander/compressor is limited to that shown in the Figures 12.14. If the maximum energy transfer is IO%, some other means, such as refigeration, must be provided to remove the additional heat. A heat balance for this situation is shown in Figure 12.17.

We know that latent heat must be removed from the feed gas mixture in order to lique6 the NGL hydrocarbons. Most of the heat removed by the expander is latent heat that is removed to condense the NGL. Therefore, if we wish to recover even more NGL, we either have to somehow remove more latent heat in the expander or do it through refigeration.



12.23


INLET GAS loo Volume %

Heat Content = 100% - GAS TO BOOSTER COMPRESSOR 88 volume %

Heat Content = 88%

I GAS OUT OF BOOSTER COMPRESSOR 88 volume %

Heat Content = ge% - 1 e - L - ’ L - ‘- .;

EXPANDER Heat Energy Transferred = 10%

REFRIQERATION Heat Energy Remowd = 2%

crpganfk

12volume% Heat Content = 0%

Figure 12.17 Heat Balance in a Cryogenic Plant with 12% of the Inlet Flow Condensing to NGL

DESIGN CALCULATIONS While the above discussion provides some simple guidelines concerning turboexpander plant

operations, the following information provides the methods for more rigorous plant designs. Even though most, if not all, expander plants are designed using computer simulators, this discussion provides the background to understanding plant designs. Information which is important to know is the turboexpander power, outlet temperature, efficiency, and impeller diameter.

Expander Power For an expander, the power or work may be calculated by,

Wact = (E) (m) (h2 - hl)theor

Where: E = efficiency (isentropic), expressed as a fraction m = mass flow rate h2 = outlet enthalpy hl = inlet enthalpy

(12.1)

The first step is to find theoretical work by using an equation of state, table, or graph of thermodynamic properties. The value of h2 is found by trial-and-error by assuming a temperature so that S I = s2. This involves assumption of the theoretical T2, running a flash and proceeding until the entropy check is obtained.

The true outlet temperature will be higher than the above theoretical temperature because actual work output is less than theoretical work output.

(1 2.2)

so, Wac, = (h2 - h1)act (1 2.3)

The efficiency may be estimated from Figure 12.1 8, but 80% is a good planning number.

The procedure for refrigeration is the same as the outlined above. Once actual work has been found, one knows actual Ah. Since hl can be calculated from inlet composition, P1 and TI, h2 can be



DESIGN CALCULATIONS

Figure 12.18 Efficiency of Expansion Turbines as a Function of Specific Speed

calculated. By trial-and-error one finds T2 that corresponds to P2 and the h2 for the two-phase mixture leaving the expander.

As part of the process calculation, one will know (or specify) PI, P2, TI, the inlet volumetric flow rate (qi), and composition. The general process is trial-and-error because enthalpy and entropy are explicit in T and P. The process is as follows:

P, T, Q

1. From PI and TI, calculate hl and S I .

2. Assume a value of T2.

3. Run a flash calculation at the assumed T2 and known P2 to establish if liquid is formed.

4. Calculate h2 and s2. (If the outlet is two-phase, these will be total stream values.)

5. If s2 from (4) equals S I , you have assumed the right temperature. If not, repeat Steps 2-4 until s2 = S I .

6. Once s2 = S I , h2 - hl = Ah for the isentropic process.

7. Calculate Ah (actual): Ahact = (E)(Ahthe,,)

8. Calculate actual power output using standard techniques discussed earlier for compressors.

This procedure is straight forward if one has a value for E, the isentropic efficiency. This will


be discussed in a later section.


12.25


1 - l

Since the above calculation process begins with the guess of a temperature, what represents a good first guess? Equation 12.4 may provide a good guessing value, even though it is based on ideal gases and assumes no liquid formation.

Where: T = P = m = E =

absolute temperature in consistent units absolute pressure in consistent units

isentropic efficiency (k- l)/k

(12.4)

If one has an enthalpy diagram of the fluid involved that includes P, T, and s, one can follow a constant entropy line to do the same thing.

Estimation of Expander Efficiency

The best way to estimate efficiency is from actual performance data.

It is difficult to correlate efficiency data because many factors affect actual performance. As with rotating compressors, normal manufacturing tolerances can affect performance measurably. Ero- sion can alter the shape of a wheel and thus efficiency. The presence of liquids likewise may have a dramatic effect. For all of these reasons, there may be a significant error in the estimated efficiency.

In correlating efficiency data, one may use the basic similarity parameters governing all tur- .? bomachinery. It appears that correlation of E versus specific speed CN,) is realistic. Figure 12.18 shows such a correlation based on data from several sources, including Reference 12.2.

Specific speed is one of the criteria for determining performance. It was discussed previously in the centrifugal pump section. The basic equation for expanders is:

SI Units Where: N =

Ah =

9 2 = A =

shaft speed isentropic Ah turbine exhaust volume conversion factor

(12.5)

FPS Units

rpm Btu/lb ft3/sec

778

It is apparent from Figure 12.18 that specific speed should be above 70 to achieve maximum efficiency. For a radial inflow turbine, the optimum E is for a specific speed between 70 and 100. However, this is not always possible because of the limitation in sizes of expanders available and other parameters which affect performance.

Figure 12.18 is for expanders where the amount of liquid formed is minimal. How much does liquid affect efficiency? This depends on where, and how, it was formed. I have seen data on steam turbines where predicted efficiency was about double actual efficiency when too much liquid was formed in an axial turbine.

-?


12.26 BP Expioratlon Company (Columbia) Ltd.

DESIGN CALCULATIONS

r" In the early planning stages, flow rates and gas compositions are rather inexact. Potential errors in both indirectly affect predicted efficiency. Some use a planning efficiency about 10% lower than that predicted by a correlation like Figure 12.18. I prefer a simulation calculation that recognizes the uncertainty and uses several values of E, to see what effect it has on those parts of the system depending on turbine performance. Obviously, E affects outlet temperature, gas-liquid ratio and the composition of both fluids which, in turn, affects all downstream equipment. I believe that picking one "magic" number is unrealistic.

Other Performance Criteria In addition to specific speed there are other parameters affecting expander performance. Spe-

cific diameter is one of these, defined by the equation

(d) [(A)(Ah)]o.25 4 =

SI Units Where: d = turbine diameter

Ah = isentropic Ah q 2 = turbine outlet volume A = conversion factor

m Híkg m3/s 31

(12.6)

FPS Units ft

Btullb ft3/sec

778

The d, shown above is a dimensionless number and would be the same when using both metric and FPS units. This was done to facilitate use of existing relationships between d, and N,. For natural gas processing the specific diameter should be approximately 1.2 to 1.4 to achieve maximum efficiency. This is a useful number for estimating impeller size and shaft speed. t-

Another parameter is the relative Mach number, which is simply the ratio of velocity at turbine discharge conditions to the speed of sound at the same conditions. Sometimes this ratio is replaced by the ratio (P2/P,).

The fourth parameter is a Reynolds number defined by the equation

u 2 d P

Re = - (12.7)

Where: U2 = tip speed d = wheel diameter p = kinematic viscosity

Any consistent set of units that makes Re dimensionless may be used.

The problem of the machinery specialist is to choose a machine that fits the system data supplied. Of these parameters, specific speed is the basic parameter of mutual concern between the various specialists involved.

!-

@.John M. Campbeii & Company


12.27


from 700 psia and 4 0 ° F to 225 psia. Expander specific speed is 80. Feed gas rate is 100 MMscfld. Gas composition is shown below.

Component

C2 c3

iC4 nC4 iC5

nC5

mol%

89.56 5.54 2.39 0,3 1 0.59 0.14 0.11

II C k I 0.31 /I

composition to expander. (Used PR EOS)

700 psia -40°F

b

* 250 psia

I Feed I Lil Comp. E

0.64

0.41

89.56

5.54

2.39

0.31

0.59

O. 14

0.1 1

lb-moi/hr I mol 70.28 0.09

45.02 0.52

9834.1 1 42.22

608.32 14.94

262.43 18.30

34.04 3.91

64.78 8.64

15.37 2.55

12.08 2.1 1 34.04 6.7 1

10980.47 99.99

iid I V lb-moi/hr

0.45

2.60

211.38

74.80

91.62

19.58

43.26

12.77

10.56 33.59

mol 0.67

0.40

91.82

5.09

1.63

0.14

0.21

0.02

0.01 0.00

500.66 1 100.00

lb-moiíhr 69.82

42.42

9622.73

533.52

170.81

14.46

2 1.53

2.61

1.51 0.45

10479.86


12.28 ~

BP Exploration Company (Columbia) Ud.

DESIGN CALCULATIONS

r

r"'

f"

Ixample 12.4 (Cont'd.):

Step 2: Expand using an isentropic efficiency of 83%. (Use PR EOS) Tout = -122.3"F Ahisen = -628.0 Btdlb-mol MW = 17.58 p = 1.36 lbm/ft3 (outlet density) Vapor fraction = 0.9451

Step 3: Calculate shaft work

Step 4: Estimate shaft speed and wheel diameter. Calculate exhaust flow:

10 480 lb-mol)[ 17.58 lbm )[ 1 ft3 )( 1 hr ) = 37.63 ft3,sec q = ( hr Ib-mol 1.36 lbm 3600 sec

= 25.72 Btdlbm 628 Btu 1 lb-mol = [ lb-mol )[ 17.58 Ibm)

Ns = 80

= 28 O00 rpm (NJ(778 AhlserJ0.75 - - (80) [ (778) (3 5. 72)]0.75 (s)0.5 (3 7.63)0.5

N =

@ N , = 80 , D, = 1.3

Expander Plant Design

The basic design of an expander plant is straight forward and proceeds as follows:

1. Identie:

a. Inlet gas condition: composition, temperature, pressure, and rate

b. Processing objectives and recovery requirements: ethane, propane, dewpoint control only

c. Product specifications: C1 in C2, C02 in C2, Cz in C3, etc.

d. Residue gas pipeline pressure

2. Estimate residue gas analysis by material balance. For deep C2 recovery, assume C3 to be 5-10% of C2 in residue gas.

3. Assume a demethanizer pressure and calculate the dewpoint temperature of the residue gas. This will be the demethanizer overhead (low pressure cold separator) temperature. For deep NGL recovery, overall expansion ratio (P,/P2) will typically be 3.0 to 3.5.




4.

5.

6.

7.

8.

9.

1 o.

11.

-I Assume a demethanizer feed temperature (expander outlet) that is 1-2°F colder than the demethanizer overhead. Perform material and energy balance calculations around expander using:

a. Outlet temperature from step 4

b. Outlet pressure from step 3 (add 2-3 psi for AP in piping)

C. Inlet pressure from step 1 (subtract 10-20 psi for AP in exchangers) d. Isentropic efficiency of about 80%

Perform material and energy balance calculations (tray to tray) around demethanizer using: a. Liquid feed from expander discharge and expander inlet scrubber b. Product specification to determine bottoms rate

Based on results from step 6, check product recoveries and make new assumption about demethanizer pressure if required. Assume 1 0-20°F approach on gadgas exchangers, demethanizer reboiler and demethanizer side heaters. Perform a heat balance around the inlet and residue streams.

a. If inlet refrigeration duty is equal to refrigeration available, go to step 9.

Qiniet Qresidue gas + Qside rebiter + Qbottom reboiler

b. If inlet refrigeration duty is greater than refrigeration available

Qiniet > Qresidue gas -t Qside reboiler + Qbottom reboiler

i) add external refrigeration ii) lower DeCl pressure

iii) lower recovery levels and repeat steps 3-7 c. If inlet refrigeration duty is less than refrigeration available

Qiniet < Qresidue gas + Qside reboiler + Qbottom reboiler

i) raise DeCl pressure

ii) lower recovery levels and repeat steps 3-7

Using horsepower from expander calculation and accounting for a 5-10 psi AP across gadgas exchangers, perform a single stage booster compressor calculation. With pressure out calculated in 9, determine recompressor horsepower to get residue gas up to pipeline pressure. Size recompressor aftercooler using 25-30°F approach for air fan cooler.

THE TURBOEXPANDER The gas expander had its beginnings as a modified form of steam turbine, which is a common

machine used to drive pumps, generators, and other rotating equipment. However, development of a high efficiency gas expander has occurred over the past 35 years. The expanders available today can recover up to 85% of the energy given up by gas as its pressure is lowered. This energy is transmitted to the rotating device such as a pump, compressor, generator, etc. In typical cryogenic plants, a compressor is attached to the expander shaft. The compressor also has a single impeller or wheel. It rotates at the same speed as the expander.

-!

@.John M. Campbell & Company


THETURBOEXPANDER

f- High pressure gas enters the expander and is directed at the outer tip of the expander impeller blades, causing it to rotate. The gas flows to the center of the impeller and exits the expander at a lower pressure. The rotating speed of the expander can be in excess of 50 O00 rpm, depending upon the volume of gas entering the unit, and the pressure drop the gas takes in flowing through the unit (inlet pressure minus discharge pressure). A high flow rate and high pressure drop result in a high expander speed and obviously a high power output.

E ixpan idei r/Co m pressor

The power developed at the rotating shaft of the expander is used to drive a single impeller compressor attached to the other end of the shaft. Low pressure gas enters the center of the impeller and discharge gas is withdrawn from the tip of the impeller blades at a pressure about 1.1 to 1.4 times

Ir' that of suction pressure.

Tu r boexpander Control

The expander is usually controlled to maintain a constant pressure on the inlet or outlet side of the plant. This is often done through a cascade flow controller as shown in Figure 12.19. The flow controller receives a remote setpoint from a pressure controller. The resulting flow controller signal is transmitted to a set of adjustable guide vanes on the inlet side of the expander. If the flow controller signals to increase the flow through the expander; the guide vanes open slightly allowing more gas through the wheel. Conversely, when a flow reduction is required, the guide vanes close restricting the throughput.

In the event the flow of gas to the expander is more than the expander can handle, a bypass valve will open and allow the excess gas through it. The bypass valve is often referred to as a J-T (Joule Thomson) valve. This is typically done by use of a split range flow controller.

Output from the flow controller is transmitted to both the expander guide vanes and J-T valve. The range of the guide vane controller is 4-12 ma or 20-60 kPa [3-9 psi] for a pneumatic controller. The J-T valve operates on an input signal from 12-20 ma or 60-100 kPa [9-15 psi]. It does not start to open until the guide vanes are fully opened. At this point the expander is fully loaded. An example of this control scheme is shown in Figure 12.19.

Most turboexpander plants are designed to operate at very low temperatures, -101°C [-15O0F]. These temperatures are required to achieve high NGL extraction levels, especially ethane. In these plants the expander outlet stream flows to a low temperature separator which is usually installed at the top of a demethanizer column.

r @John M. Campbell & Company


12.31


Cross-Section of Expander-Compressor

Outlet and Liquid

inlet Gas

@ Remote Setpoint from Plant - - - _ _ _ _ _ _ _ _ _ Pressure Controller

I I

‘ - - - - I - - - -

I 9-15 psi I I

I h Discharge I

Gas I J-T Valve \ \ \

Suction Gas

Expander Compressor

Figure 12.19 Expander Control System

@John M. Campbell US Company


THETURBOEXPANDER

f- The condensed liquids drop down on the top tray of the demethanizer where they are stabilized to meet an NGL product specification. The demethanizer bottom product is a mixture of Cz, C3, C4's and C5+.

In hydrocarbon dewpoint control plants, these extremely low temperatures are not required. The expander outlet temperature is usually O to -20°C [32" to -4"FI. Expansion ratios are typically about 1.5. A typical hydrocarbon dewpoint control plant using a turboexpander is shown in Figure 12.20.

r"

TO PIPELINE OR COMPRESSOR

J-T VALVE + r

EXPANDER INLET

EXCHANGER SE PARATOR

EXPANDER 1 COMPRESSOR

COLD SEPARATOR

I I

Figure 12.20 Expander in a Dewpoint Control Application

As gas flows through the expander, its temperature is lowered and some of the stream condenses. The liquid which forms has no detrimental effect on the expander. If the gas stream entering the expander contains solid particles of dirt or debris or contains moisture or carbon dioxide which will freeze at the low temperatures in the unit, serious damage to the machine may result. At the high operating speed of the machine, the presence of solid materials - debris or ice - will quickly sandblast the wheel and casing. The expander inlet separator is a very important piece of equipment. It must be sized properly to remove these components from the gas.

Guide Vanes and Expander Impeller Vanes are in the open position.

A screen is normally installed on the inlet gas line to the expander to remove solid particles from the gas stream. Moisture is removed from the gas in the dehydrators in the front end of the plant.

r @Job M. Campbell & Company


12.33


Anti-Surge Controller

-1 Booster Compressor Control

consists of a recycle line and quick-opening control valve that is positioned by an anti-

The booster compressor uses the shaR work generated by the expander to boost the gas into the final recompressor. It is a single-stage centrifugal compressor with an open backward bladed impeller.

I I

Open Backward-Bladed Impeller

job is to recycle discharge gas back to the suction side when a surge condition is ap- proached.

The compression ratio across the booster compressor wheel depends on the tip speed and inlet rate to the compressor, but typical values range from 1,l to 1.4.

In most installations, the suction pressure to the booster compressor is essentially constant since it is controlled either by the turboexpander guide vanes or by the final recompressor. Oc- casionally the booster compressor discharge pressure will be controlled, but this is unlikely.

The primary reason the suction pressure is constant is because it is also equal to the expander outlet pressure (less AP through piping and exchangers). The expander outlet pressure sets the expander outlet temperature which is the control point for the NGL extraction levels.

If the expander outlet pressure was allowed to vary, the expander outlet temperature would change as well - making it very difficult to stabilize the NGL extraction.

Since the suction pressure to the booster compressor is usually fixed, the discharge pressure varies with expander power and gas rate. However, the gas flow rate to the booster compressor is proportional to the gas rate to the expander. The flow will not be exactly the same, but will be approximately the same. Thus, if the gas volume to the expander decreases by 20%, the volume to the booster compressor will also decrease by 20%. Since the shaft work available from the expander will also have dropped by 20%, the pressure rise (head) across the booster compressor will remain essentially constant.

"',



THETURBOEXPANDER

machine's compression ratio. Both measurements are required by the anti-surge controller so that it can sense an approaching surge condition while still permitting normal operation over a wider range of suction gas rates.

r

Surge in a single-stage booster compressor usually occurs at about 65% of the design flow rate for design suction and discharge pressures (or design compression ratio). If low-flow recycle is required, the total flow through the compressor is the suction gas flow rate plus the recycle rate. The net discharge flow will still be the same as the suction gas flow rate.

Gas pressures in a typical expander-compressor in a deep NGL recovery plant are shown in Figure 12.21. The highest gas pressure is that at the inlet to the expander. The expander and compressor sections are connected with a common housing so that if leakage should occur, high pressure inlet gas to the expander could leak into the compressor. The compressor gas is residue gas from the demethanizer, so any leakage which occurs will result in some gas simply by-passing the expander and cold separator. Thus, one function of the seal system is to prevent high pressure expander inlet gas from leaking into the compressor.

A seal is a device for preventing gas or liquid under high pressure from leaking to a point of lower pressure. A common example is a mechanical seal used on centrifugal pumps that prevents liquid under pressure inside the pump form leaking outside to the atmosphere. Mechanical seals require lubrication of the moving parts to prevent excessive wear. Mechanical seals cannot be used on the expander because oil to lubricate the seal would enter the expander and solidify at the low temperature and seriously damage the machine.

/c"

2415 kPa

Expander Lube oil Reservoir

Figure 12.21 Typical Operating Pressures and Locations of Labyrinth Seals in Expander Compressor

@John M. CampbeU lk Company


12.35


Seal No. on

Drawing

7 Labyrinth type seals are used on the expander and compressor to prevent gas under high pres-

sure from flowing to a lower pressure point. A labyrinth seal is a series of teeth which are machined in the impeller and seal rings to a clearance between the end of each tooth and the shaft of approximately 50-100 microns [0.002-0.004 in.]. The labyrinth acts as a pressure reducing device. Gas is at a higher

pressure on one side of the labyrinth than on the other side. A small amount of high pressure gas leaks across each tooth in the seal. As it passes from one tooth to another, its pressure reduces, and its volume increases. By the time it reaches the final tooth, it is at its lowest pressure and largest volume. The clearance between the teeth and shaft will allow only a small volume of gas to flow. Thus, there is a continuous flow of gas through a labyrinth seal, but it is a negligible quantity. A slight amount of wear in the labyrinth teeth will obviously increase the quantity of gas flowing through the seal.

Labyrinth Seal

Pressure on Each Side of Seal Disposition of Gas That

High Pressure Side Low Pressure Side Leaks Through Seal

Labyrinth seals are shown in Figure 12.21 and tabulated in Table 12.1.

I 2 3

TABLE 12.1 Labyrinth Seals in Expander-Compressor

Expander inlet gas Expander outlet gas Expander outlet gas Seal gas Expander outlet gas Expander outlet gas Seal gas Housing Compressor suction gas

Location of Seal Expander impeller Expander shaft Expander shaft Compressor shaft

Compressor Impeller

Between Suction and Housing discharge pressure

Compressor suction gas

5 1 Compressor discharge Between suction and Compressor suction gas discharee messure

As you can see in Figure 12.21, the lowest gas pressure is that in the housing between the expander and compressor. Gas which leaks across the labyrinth seals on the expander and compressor flows into the housing. This gas in turn enters the lube oil reservoir, which is vented to the suction side of the compressor. Thus, the pressure in the housing will be the same as that at the suction to the compressor.

If cold expander gas leaks into the housing, it can chill the lube oil to the point that it would lose its lubricating properties. To prevent this from occurring, a seal gas system is used which serves as a barrier between the expander gas and the lube oil.

Seal gas is clean dry gas at a pressure about 350 kPa [50 psi] above that of expander outlet pressure. It is injected between two labyrinth seals on the expander shaft. Since this pressure is higher than that on the back side of the expander impeller, some seal gas will leak through the labyrinth seal (number 2) and end up in the expander outlet gas. This is a small quantity of gas which has very little detrimental effect on temperatures in the expander.

Some seal gas also leaks across the labyrinth seal (number 3) into the housing between the -I expander and compressor. This gas flows to the suction side of the compressor.


12.36 BP Exploration Company (Cdumbla) Ltd.

THETURBOEXPANDER

Seal gas must come from a source at least 350 kPa [50 psi] above that of the expander outlet gas. The seal gas must be clean and dry and at approximately ambient temperature. A filter is usually included in the line to remove solid particles from it. Outlet gas from the plant is often used as a source of seal gas. Seal gas is introduced to the expander before startup, and maintained during operation. A control valve and flow indicator are included in the seal gas line for adjusting the flow rate.

Thrust Control

In any type of centrifugal device, thrust forces develop which tend to move the shaft toward one end or the other. If it were to move laterally along its axis, the impellers would touch the casing and quickly wear out. In an expander-compressor, thrust bearings on each end of the shaR prevent lateral movement. However, the thrust forces against the bearings must be controlled at a moderate level to prevent bearing failure and serious damage to the machine.

The thrust force is due to a difference in suction and discharge pressure acting on the front and rear face of an impeller. Look at the expander impeller in the following drawing. High pressure inlet gas enters at the tip of the impeller, and leaks around the labyrinth impeller seal to the rear face and exerts a force to the left. Low pressure outlet gas pressure is imposed on the front or left side of the impeller. In order to neutralize the thrust in the expander impeller, gas which leaks around the labyrinth seal on the rear face is slightly above outlet pressure. One manufacturer has holes in the impeller which allow gas under pressure behind the impeller to flow through the outlet gas line. In any event, the thrust force of the expander impeller is partially neutralized. The gas pressure behind the impeller is slightly more than that on the front face, so that a moderate thrust force is exerted toward the expander.

r"

Inlet Dischame

Outlet Gas

r"

1 - Inlet gas leaks through impeller seal to rear face of impeller and exerts thrust force to the left. Thrust is partially neutralized with holes in the impeller that release most of the pressure behind the impeller to the front face which is at outlet gas pressure. A moderate thrust force is exerted to the left.

2 - Gas at discharge pressure leaks around impeller seal to rear face of impeller and exerts thrust force to the right. Thrust control valve holds enough pressure behind the impeller to exert a thrust force equal to that on the expander end.

Thrust Control



12.37


The thrust force imposed on the front or face. The thrust force is

-I on the compressor impeller is in the opposite direction. Suction pressure is right face of the impeller and discharge pressure acts against the rear or left to the right.

The thrust control system regulates the thrust force toward the compressor end so that it is about the same as that toward the expander end. The compressor thrust force is regulated with a control valve in a line which connects the backside of the impeller to the inlet gas side. The control valve is usually preset by the manufacturer to maintain a given differential pressure between the back side and front side of the impeller. One manufacturer uses a piston actuator on the control valve in which pressure on each side of the piston is that of thrust bearing pressure on each end of the shaft. If the shaft moves slightly in one direction, thrust bearing pressure will rise on one bearing and fall on the other, which will cause the piston to move so that more or less gas behind the impeller flows through the control valve and thereby equalizes the thrust bearing pressures. The previous drawing shows such a system.

If the thrust control system should fail, and the control valve closes, a high thrust force toward the compressor end will result, which may cause the compressor thrust bearing to fail. Any type of bearing failure in a machine rotating at 25 O00 rpm or more can result in serious damage. Thus, the thrust balancing system should be carefully watched and properly maintained.

If the thrust neutralizing arrangement on the expander end should fail due to plugging of holes in the expander impeller, the thrust force developed will be far greater than the thrust control system can handle. A high thrust force toward the expander will result, which can quickly damage the expander thrust bearing. A shutdown from high thrust occurs in this situation to prevent damage to the machine.

-?I

Lubrication System

The expander-compressor has a bearing on each end of the shaft. These bearings must be continuously lubricated with clean lubricating oil of the approved type at the proper temperature. Lu- brication failure for a short-period of time may result in a bearing failure which may seriously damage the machine.

The lube oil system is shown in the drawing below. Pressures shown are typical for a plant in which the expander outlet pressure is 2070 kPa [300 psi]. Flow is as follows:

Oil from the reservoir enters one of the pumps which is able to raise its pressure several thousand kPa [several hundred psi]. A control valve in the pump discharge line releases excess pressure through a spillover line that returns to the reservoir. Oil from the pump enters a temperature control valve, which is positioned by a temperature controller that allows some of the oil to by-pass the cooler to maintain a constant temperature. The oil then flows through one of two filters to remove solid particles and enters the bearings on each end of the shaft. The oil flows out of the bearings and drops to the bottom of the housing and flows by gravity into the reservoir, and the cycle is repeated.

The control valve in the spillover line is positioned by a dzfferential pressure controller (AF' contr) which is set to hold the pressure of oil entering the bearings about 900 kPa [130 psi] higher than the pressure of oil flowing out of the bearings. This difference in pressure will assure adequate flow of lube oil through the bearings. Increasing the differential pressure will raise the flow of lube oil to the bearings, and vice versa.

/c4\



THETURBOEXPANDER

+--

I I 2880 kPz [415 psi]

Pressure Tank

A stand-by pump is provided which turns on automatically in the event pressure in the system drops due to failure of the primary pump. The lube oil pumps are normally driven by electric motors. If a power failure occurs, the motors will obviously stop, and the expander-compressor will shut down from low oil pressure.

In order to provide lubrication to the bearings during the run-down period, a pressure tank filled with oil is provided which will let oil flow from the tank to the bearings when the oil pumps stop. The pressure tank has a bladder which is inflated to a pressure slightly above that of the oil reservoir. If both lube oil pumps shut down, oil pressure will fall and oil will flow out of the tank to the bearings.

The oil reservoir operates at the suction pressure of the compressor. A vent line connects the reservoir to the compressor suction line. Seal gas or compressor discharge gas which leaks through labyrinth seals into the housing around the bearings, flows into the oil reservoir and exits in the vent line to the suction of the compressor.

A heating coil is installed in the oil reservoir for use in cold weather, or in the unlikely possibility of cold expander gas leaking into the reservoir.


Technical Assistance Service for the Design, nnnmtinn and Maintnnant!! nf Cae Plante

12.39


The heat shield shown in the previous drawing is a plastic plate behind the expander impeller which insulates the metal housing from the low temperature gas passing through the expander. With- out the insulating shield, the temperature at the expander bearing could reduce to the point that lubricating oil would solidify and the bearing would quickly fail.

The temperature of lube oil entering the bearings should be about 38°C [lOO°F]. The oil not only lubricates the bearings, but also removes the heat of friction released when the bearing faces rub against each other. Thus, the temperature of oil draining into the reservoir should be higher than the temperature of oil entering the bearings.

If the temperature of oil draining into the reservoir is lower than the temperature of oil entering the bearings, leakage of cold expander gas into the housing is indicated. This can result from:

1. 2.

3.

Seal gas flow has stopped. Excessive wear in the expander impeller labyrinth seal, which results in more gas leaking across the seal than the holes in the impeller can handle. Pressure behind the impeller will rise above seal gas pressure, and it will flow through labyrinth seals on the expander shafi into the housing and cool the lube oil. Failure of the heat shield which will allow expander gas to leak into the housing.

If a low oil temperature is caused from stoppage of seal gas flow, the situation can be corrected by returning the seal gas flow to its normal rate. The oil temperature should rise to its normal point if this is the problem.

If seal gas flow is normal, and low oil temperature occurs, failure of the heat shield or the

7 expander impeller seal is indicated and the machine should be shut down and the cause corrected.

Maintaining proper lubrication in the bearings in the expander-compressor is essential for pro- longed life of the machine. The manufacturer’s recommendations should be followed in selecting the type of oil to use, and the oil temperature should be carefully controlled by adjusting the amount of cooling in the cooler or using the heater in the reservoir. Oil filters should be closely watched and elements changed when the pressure drop across a filter indicates it is plugged with dirt.

It is a good idea to withdraw a sample of the oil from the bottom of the reservoir at periodic intervals and visually check it for the presence of metallic particles, which indicate bearing wear.

The oil should be tested monthly for its lubricating qualities, and replaced when a 10% deterioration is reached. In the absence of monthly tests, it should be replaced at three to six month intervals.

Expander Compressor Safety Devices Because the expander operates at a very high speed with close tolerances between moving

parts, the slightest irregularity can cause serious damage to the machine. The following safety devices are provided on most expander-compressor units which will shut the unit down. The unit is shut down by closing a valve in the inlet gas line to the expander.

1. Overspeed

2. High thrust load 3. Low lubricating oil pressure

4. Low temperature of oil leaving from the housing

5. High temperature of oil entering the bearings



EXPANDERICOMPRESSOR OPERATION

w (I Coder )I __* Pump

Typical Expander Shut Down

In most installations, an annunciator panel is provided which indicates the cause of shutdown. The cause must obviously be corrected before restart of the machine.

EXPANDER/COMPRESSOR OPERATION

The actual starting of the expander/compressor is usually the last step in the cryogenicplant's normal startup procedure. And stopping the machine is usually the first step in an orderly shutdown procedure. The most involved operational concerns are focused on these two (hopefully infrequent) events. Normal operation is mostly concerned with making frequent checks of the expandedcompressor's utilities (i.e., sealing, thrust control and lubrication systems), and with the machine's response to the plant's normal control functions.



12.41

I


Startup Operations -l Expander/compressor startup usually occurs after the plant has operated for some time in a J-T

valve mode known as "cool-down". The plant is therefore quite cold before the expander takes over. Typical expandedcompressor startup consists of a few major steps:

1 . Establishing seal gas flow 2. Establishing lube oil circulation 3. Commencing flows through the machine The seal gas and lube oil system must simply be operational before turning the machine, and at

all times thereafter, or the safety systems will lock out any attempt to run the unit. If at any time the critical variables in these systems go out of limits, the local control and shutdown systems will by-pass gas through the J-T valve, shut the expander inlet valve and stop the machine. (The booster compressor is commonly by-passed on shutdown through a check valve in an otherwise open by-pass line.)

d

Discharge Gas

Check Valve Normally open Block Valve

I+- Suction Gas

Compressor

The task of commencing flows through the machine is essentially a manual job. Once the expanders inlet and outlet block valves are open, the operator manipulates a manual loading station to gradually close the J-T valve while opening the expanders inlet vanes. Rotation actually begins when the inlet block valve is first opened, but the rotational speed increases with increasing flow. The speed at which this is done depends upon the expanders starting temperature; in general, the warmer the machine is to start with, the longer the operator should take to "load" the expander. Operators must follow the manufacturer's startup guidelines which should be incorporated into the unit's authorized operating procedures.

Startup of the compressor half of the machine is fairly simple. The most basic requirements are that isolation valves be open and the anti-surge controls be operational. As the expander shaft turns the compressor wheel faster and faster, there is an increase in pressure on the discharge side. Higher discharge pressure closes the check valve in the compressor by-pass piping and all of the flow is directed through the machine.

Normal Shutdown

The expander/compressor orderly shutdown consists of manually taking control of the expander flow controls, and gradually by-passing gas through the J-T valve while closing the expander inlet



l i '

PRINCIPLES OF DEMETHANIZER OPERATION

vanes. As the machine slows, the booster compressor discharge pressure will decrease and the gas flow will by-pass the compressor when its by-pass check valve comes open.

6"

After the machine has been stopped and isolated, the plant may continue operating on its J-T valves, or the entire unit may be shut down. Always refer to your authorized operating procedures for details that cannot be covered by a generic procedure such as presented here.

Normal Operation

The expander/compressor must respond to the plant's normal control functions of pressure and flow control. The inlet vanes on the expander should work as expected without unnecessary by-pass flow through the J-T valve. The booster compressor's anti-surge system should successfully keep the machine out of surge conditions. As explained earlier, the expander/compressor system is essentially self-limiting where operating pressures and flows are concerned and should not be of great concern to the operator so long as upstream units and downstream facilities like pipelines and recompressors are also operating normally.

Multiple-machine installations, in both parallel and series configurations, may require more operator attention and involvement, the degree of which is a function of the chosen control system and its ability to do the intended job. These are site-specific concerns that must be learned for the individual plant.

Part of normal operation involves being on the lookout for abnormal conditions, the most common of which are related to freeze-ups caused by high COZ levels or inadequate gas dehydration.

F"

PRINCIPLES OF DEMETHANIZER OPERATION

Demethanizer towers come in a variety of configurations. They may contain trays or packing, they may have one or more side and bottom reboilers, they may have provisions for a recti@ing (reflux) section or not, and so on. This section aims to explain the most common of such differences among demethanizers; but, before amplifying these differences, it is prudent to first discuss the things that all demethanizers have in common.

All demethanizers are muss transfer devices. That is, they allow us to selectively transfer hydrocarbon molecules between their vapor and liquid phases for the purpose of separating them (or groups of them), one from another. Specifically, demethanizers separate methane from the ethane and heavier components of a hydrocarbon mixture by selectively placing almost all of the methane in the vapor phase while placing as much of the ethane and almost all of the heavier hydrocarbons into the liquid phase.

Such a separation as we intend to make requires a specialized piece of equipment known generically as a distillation column or fractionator. Fractionation is one of the most basic and common processes to be found in the hydrocarbon and chemical process industries. All sorts of feed mixtures lend themselves to separation by fractionation. Regardless of their different uses and the wide variety of specialized names, all share the same basic principles of operation.

We begin this discussion of demethanizers, therefore, by examining the basic principles of fractionation. This scrutiny of basic principles will supply us with the knowledge we need to understand how demethanizers, in all their variations, are intended to work.

P




DEMETHANIZER AS FRACTIONATOR "7 The liquids which condense as the feed gas flows through the cryogenic plant's heat exchang-

ers and turboexpander are a mixture of all the hydrocarbons present in the feed, including methane. As the gas passes through the plant, condensation is more or less continuous as the gas becomes colder and colder. The heavier components, especially the butanes and heavier, condense more easily than the lighter components, but still, a significant amount of the lightest component, methane, ends up in the liquid mixture.

There are at least two reasons for methane condensation:

1. Because of the usually high methane content of the feed stream, there is a considerable equilibrium driving force to cause methane to exist in the condensed liquid.

2. The other condensed hydrocarbons act like "absorption oil" and physically absorb methane molecules.

Since NGL specifications place maximum limits on the amount of methane remaining in the product, it must be removed to the specification limit. NGL specifications may vary, but the maximum amount of methane permitted in the NGL product is usually from 1.5 to 3.0 liquid volume percent of the ethane.

The result is a rather large methane content in the NGL.

Additionally, from the point of view of the gas stream, while more and more of the heavier hydrocarbons leave the vapor state as they condense, there will still be significant amounts of propane and ethane (especially ethane) that are contained in the vapor, even though the temperatures get very cold. If these components are allowed to leave in the gas stream they must be considered to be "lost production". Since these NGL components are more valuable as part of the liquid product, steps may be taken to improve upon their recovery.

Compositional specifications for the overhead demethanizer product (the residue gas) will likely vary from plant to plant, with a minimum heating value specification (in Btu's per standard cubic foot of gas) being the most common. Typical minimum residue gas Btu specifications (FPS Units) are in the range of 950 to 1000 Btdscf [35 to 37 mJ/std m3] (gross basis'), with 950 Btdscf [35 MJ/std m3] being common for high (88%) ethane recovery plants. The amount of ethane remaining in the residue gas is the thing that will have the greatest effect on the residue gas heating value, but optimal operation for a given plant design should produce specification residue gas. Therefore, most plants are simply operated to maximize ethane recovery according to their design basis.

So if we are trying to separate NGL's from the remainder of the gas stream, it's fairly obvious that we cannot rely on condensation alone to do the job; we simply do not get a good enough split between the methane and the heavier hydrocarbons. The demethanizer tower sharpens this separation. The job of the demethanizer is twofold:

1. Remove the unwanted methane from the NGL liquid, and 2. Improve the recovery of ethane and propane. The DeCl is essentially a top feed stabilizer with a bottom and side reboiler. It is seldom

refluxed. Operating pressure usually ranges from 175-350 psia. Bottom product specification in deep 2 From the GPSA Data Book: Heating Value (Heat of Combustion) -The amount of heat developed by the complete combustion of a unit

quantity of a material. For natural gas, the heating value is expressed as the gross or higher value in Btu per cubic foot of gas at designated conditions. The gross or higher heating value (normally referred to in the United States) is that measured in a calorimeter when the water produced during the combustion process is condensed. The heat of condensation of the water is included in the total measured heat. The net or lower heating value (normally referred to in Europe) is that obtained when the water produced by the combustion process is not condensed. This is the usual circumstance in equipment burning fuel gas. The net value is the maximum portion of the heating value that can be utilized in usual equipment. The difference between the gross and net heating value is the heat that could be recovered if the water produced is condensed.



DEMETHANIZER AS FRACTIONATOR

f- ethane recovery is normally expressed as a Cl/C2 ratio. Values typically range from 0.01 to 0.03. In the U.S., the C02 content in the ethane product must be less than 100 ppm. Occasionally this specification is met collaterally with the Cl/C:! specification. In most cases the C02 specification cannot be met without incurring high ethane product losses. In these cases, downstream treating of the NGL product is required for C02 removal. Treating is typically done in a liquid-liquid contactor using a DEA solution.

Basic Demethanizer The basic demethanizer works as an NGL stabilizer. The focus of stabilizer operation is to

remove an unwanted light component (methane) from a liquid product (NGL), through the action of a stripping vapor. In doing so, a significant amount of the desired liquid product (NGL) may be lost in the stabilizer overhead vapor (residue gas), depending on the overall design of the stabilizer.

The basic demethanizer differs from a fractionator in a number of ways: first is the obvious lack of reflux facilities; second, several feed streams may enter at several different locations, whereas the fractionator usually has but one feed; and third, there may be several bottom and side reboilers on the demethanizer, whereas the fractionator usually has but a single reboiler at its bottom. Either trays or packing may be used for contacting liquid and vapor in the demethanizer.

r-

Figure 12.22 is an example of a basic demethanizer in a cryogenic plant. The greater of the two liquid feed streams is the one that comes in from the expander feed separator. Liquids that condense in the gadgas exchanger fall to the bottom of the separator and then get dumped out on level control to the demethanizer tower. This stream will be significantly "two-phase" since the liquid flashes partially to vapor as it takes a pressure drop in going across the separator's level control valve to the demethanizer, which operates at a much lower pressure. This feed stream enters the demethanizer at a point several trays down from the top (or on top of the next-to-highest packed section), where its composition and temperature match those of the liquid fraction expected inside the tower at that point.

Figure 12.22 Basic Demethanizer with Bottoms Reboiler and Side Reboiler



12.45


-7 The feed from the turboexpander outlet enters the top of the demethanizer, at the separation

bell. This is the coldest feed stream. It is also two-phase, gas and liquid. The temperature of this feed stream matches (as well as determines) the tower's top temperature. The incoming liquid collects in the bottom of the separation bell and simply overflows into a liquid distributor that deposits the liquid on the top tray (or over the top surfaces of the uppermost packed section). The gas portion of the feed stream combines with the vapor rising from the stabilizer section below and exits from the top as overhead vapor. The separation bell commonly contains a mesh-pad mist eliminator (not shown) to guard against liquid carry-over.

The demethanizer in Figure 12.22 has a bottoms reboiler that uses a heating medium (usually hot oil or hot glycol/water solution) to supply the process heat needed to create stripping vapor. This reboiler is usually of the thermosyphon design; liquid from the bottom tray of the tower is drawn off and passes upward through the reboiler where it vaporizes partially and then returns to the tower. When back inside the tower, the vapor portion acts as stripping vapor and flows upward through trays or packing. The liquid portion falls to the bottom where it dumps out on level control as bottoms (NGL) product.

The figure also shows a side reboiler, attached about 1/3 of the way up, which uses a portion of the warm inlet gas to add more heat to the tower. The use of side reboilers reduces the heat input necessary at the bottoms reboiler. Side reboilers are also of the thermosyphon type, taking in liquid from a draw-off tray and returning vapor and liquid to a point in the tower just below the draw-off tray. The side reboiler operates at a temperature that closely matches the tower's temperature profile at the point where the warmed vapor and liquid return to the tower. The heat supplied by the side reboiler is contained in the warm feed gas that passes through the tubes. This stream returns (significantly colder) to the feed stream at a point downstream of the gadgas exchanger.

"1 The demethanizer pictured in Figure 12.23 uses the heat contained in the feed stream to pro-

vide all of the required process heat at the demethanizer. The warm feed flows first through a bottoms reboiler and then on through the side reboiler before rejoining the feed stream.

I

Figure 12.23 Basic Demethanizer with Process Heat Supplied Entirely by Feed Stream in Bottom and Side Reboilers "7

0 J o h M. Campbell & Compauy

12.46 ____ _ _ _ _ _ _ _ _ _ _ _ _____ ~



In Example 12.5 the CO2 specification of 100 ppm is not met in the DeCi. Some type of downstream processing would be required to remove the CO;? from the NGL product.

The side reboiler duty in Example 12.5 was arbitrarily set at 1.5 MMBtdhr. No optimization was performed to determine the optimum duty. As the side reboiler duty is increased more C2 and C3

Example 12.5: Using the information from the previous example, estimate the bottom product rate, temperature profile and reboiler duty for a demethanizer. Bottom product specification is 3% C1 in C2. Tower pressure is 225 psia.

Feed I

-is. EXPANDER

Residue Gas to

High Pressure Cold Liquid Feed

Bottom Reboiler

DEMETHAN E IZER NGL Product

C /G Exchanger

Separator

1.5 MMBtu/hr

I Residue Gas mol YO

0.68 0.38

94.98 3.64 0.3 1 0.01 0.01 0.00 0.00

Ib-moi/hr 70.27 39.69

9827.14 376.08

3 1.65 0.66 0.53 0.01 0.00

mol YO o .o0 0.84 1.10

36.61 36.38

5.26 10.13 2.42 1.90

c6' 0.00 0.00 5.37 Total 100.00 10346.03 100.00

'oduct Percent lb-molhr Recovery bPd

0.00 5.34 6.97

232.24 38.2 1,343.82 230.78 87.9 1,375.87

33.38 98.0 236.22 64.26 99.2 438.31 15.36 99.9 121.58 12.07 100.0 94.61 34.04 100.0 339.70

634.44 3,950.09

-122.0 -16.4 -1 14.7 -94.6

4 -73.8 9 25.8 5 -19.2 10 55.5



12.47


"7 are lost at the top of the DeCl tower due to higher top temperatures. As the side reboiler duty is decreased, C2 and C3 retention goes up, but the overall process is less thermally efficient because a low temperature heat sink (side reboiler) is sacrificed in favor of a high temperature sink (bottom reboiler). Other factors which affect the selection of the side reboiler duty and location include vapor traffic in the bottom of the column and COZ freezing.

COZ freezing is an issue which has a significant effect on the design and operation of an expander plant. When COZ concentrations in the vapor andor liquid phase are relatively high at low process temperature, solid C02 may form in the system. This solid COZ behaves much like "ice" in the system, restricting flow and plugging equipment. COZ freezing can be predicted by most process simulators using an EOS. Figure 12.30 provides a quick reference for checking for the presence of solid c02.

The most critical locations for COZ freezing are downstream of the expander and in the top section of the DeCl tower. For example, in the previous example, the concentration of COZ in the two-phase stream leaving the expander is 0.4%. The concentration of C02 leaving stage 4 of the DeCl is 1.2% in the liquid phase and 1.4% in the vapor. While COZ freezing does not occur in this system you can see how the COZ concentrates in the coldest areas of the DeC,.

Refluxed Demethanizer

The basic demethanizer is a top-feed stabilizer. It uses no external reflux facilities such as found on typical fractionators. Reflux in a top-feed stabilizer is provided by the liquid portion of the feed that enters on the top tray. Design engineers can add more trays to the stabilizer to make the available reflux more effective, but still, significant quantities of ethane (and even propane) may slip by and not be recovered. 7

When the amounts of propane and ethane that would be lost from the top of the basic demethanizer are great enough to make worthwhile the extra effort required to recover them, external reflux may be added to the demethanizer. When external reflux is added to a stabilizer, the result is a fractionator, since fractionators have both a stripping section and a recti@ing section. There may be a few features that keep a refluxed demethanizer from looking like a "real" fractionator (multiple feed streams, multiple bottom and side reboilers), but its operation is very much like that of a true fractionator. A refluxed demethanizer is shown in Figure 12.24.

In a "typicalt' fractionator, a portion of the overhead vapor is condensed for reflux liquid in overhead condensers and that liquid is pumped back onto the top tray of the tower as external reflux. The reflux liquid is the same or nearly the same composition as the top product, whose purity may be 90% or greater. The job of external reflux is to cool the rising vapor, thereby condensing heavier components in the vapor and forcing them back to the bottom of the tower. We might say that extemal reflux "washes" the heavy components back to the bottom of the tower where they belong. In the demethanizer, ethane is the "heavy end" that we would like to wash back to the bottom as part of the NGL product.

Whereas the "typical" fractionator (i.e., a depropanizer or debutanizer) uses atmospheric methods (aerial finífan condenser, water-cooled condenser) to condense its overhead vapor for reflux, an expensive refrigeration system of some kind would be needed to condense demethanizer overhead. In fact, some designs have gone as far as to install an additional turboexpander on the demethanizer overhead for this purpose. The most common method used to create demethanizer reflux, however, is to return a portion of residue gas taken downstream of the recompressor and cool it in cryogenicplant heat exchangers until it is nearly as cold as the demethanizer overhead. The result is a vapor no colder

7




I I

Figure 12.24 A Small Recycle Stream of High Pressure Residue Gas is Chilled and Partially Condensed to Cre- ate Reflux for a Dernethanizer Column

than the top of the demethanizer, but at a higher pressure, so condensation will occur. No reflux pump or accumulator is necessary.

T The flow of reflux to the demethanizer can be directly controlled as was shown in Figure

12.24. The answer to the question, "How much reflux is enough?" can be determined in perhaps several ways. Operators can inspect lab reports showing either residue gas heating value or composition. If either the heating value or the ethane content of the residue gas appears to be too high, the operator can increase the flow of reflux to the demethanizer. Too low a reading and the reverse action is taken. In any event, operators must take care not to flow excessive amounts of reflux; doing so only places an unnecessary load on the tower and reboilers, and may result in overall reduced recoveries.

Another problem with excessive reflux is this: practically everything recycled for reflux will exit the top of the demethanizer tower and add to the flow going to the brake compressor, increasing the load on the machine and possibly backing up pressure on the demethanizer. In addition, the greater the flow of gas and liquid through the recycle circuit, the greater will be the horsepower requirements for the recompressors. It appears that it can be quite easy to wipe out the advantages of increased ethane recovery when reflux is excessive.

Again, it only makes sense to add external reflux to demethanizers when the additional revenue from the increased NGL recovery is greater than the cost of installing and operating the reflux loop.

Demethanizer Absorber Another of the enhancements that can be made to the basic top-feed demethanizer is the addi-

tion of an absorber column (sometimes called a "rectifier") in the demethanizers overhead vapor stream. The purpose of the absorber is to maximize the recovev of ethane and heavier components

f-



12.49


from the inlet gas stream and is comparable in that respect to refluxed demethanizers. Recovery performance is reported to be comparable as well. The main advantage over other processes is that warm gas is taken upstream of the gas to gas exchanger, thereby permitting greater total plant throughput than other plants with similarly-sized turboexpanders. The basic design was first patented by Ortloff Engineers, Ltd. The basic process is shown in Figure 12.25.

"i

Figure 12.25 Warm, High Pressure Inlet Gas By-Passes the Turboexpander to be Totally Condensed for Use as Absorption Oil in a Packed Absorption Column

By-passing gas around an expander usually decreases overall recovery. But in this case, the high pressure gas is subjected to the very low temperatures of the cold residue gas stream which causes total liquefaction. This stream is then reduced in pressure across an expansion valve, resulting in partial vaporization, auto-refrigeration and even colder temperatures (the coldest in the plant, actually). The cold liquid is then fed to the top of the absorber column where it functions as reflux or absorption oil (with a molecular weight of about 18).

The expander outlet stream joins with the overhead vapor from the demethanizer column in the bottom of the absorber vessel. The combined vapor stream travels upward through the packing and is contacted with the absorption "oil''. The downward flowing liquid absorbs ethane and heavier components from the rising vapor. Liquids that collect in the absorber bottom must be pumped to the top of the demethanizer column as tower feed. (A popular modification is to mount the absorber atop the demethanizer column to utilize gravity flow, thereby eliminating the need for the pump.)

The efficiency of the absorber is temperature dependent, with colder temperatures resulting in greater recovery. The process operates under the constraints of flowing enough warm by-pass gas to provide sufficient liquid for absorption, while not flowing so much as to cause higher temperatures and reduced recoveries. In other words, the absorber cannot be operated by varying the flow the absorption oil as can be done with reflux.

,7




r

F'

Demethanizer Control The major control point on the demethanizer is its bottom temperature. The higher the tem-

perature, the greater will be the production of stripping vapor and the smaller will be the percentage of methane remaining in the bottom product. Demethanizer bottoms temperature control is usually straightforward; a temperature controller mounted somewhere on the tower or reboiler return line will regulate reboiler heat input. A temperature control valve in the reboilers hot oil supply line, for instance, will directly control the amount of heat transferred in the reboiler.

When feed gas is used to reboil the tower bottoms andíor provide heat to the side reboilers, temperature control becomes a bit more complicated. Figure 12.26 shows a demethanizer column with a bottoms temperature controller that operates a control valve to adjust the flow of warm feed gas through both the bottom and side reboilers. If, for instance, the demethanizers bottom temperature should drop below setpoint, the temperature controller will cause the control valve to open wider, increasing the flow of warm feed gas through the reboilers. This will eventually increase the tower's temperature, but, because the feed gas is a relatively poor heat medium, it may take several minutes to notice any such change.

-@ I I I I I I I

Figure 12.26 Bottoms Temperature Control and By-pass Butterfly Valves

When warm feed gas is taken through the reboilers, the remainder of the feed flows through the gadgas exchanger. The pressure drop across a heat exchanger may be 35 to 70 kPa [5 to 10 psi]. The portion of the feed gas that flows through the reboilers may pass through several exchangers, whereas the remainder passes through the single gasigas exchanger. We can expect that the total pressure drop through the reboilers will be greater than that through the gadgas exchanger. What then, keeps the gas flowing as needed through the reboilers? The answer is that the operator must make the flow go where he needs it to go. He (or she) does this with a restrictive valve in the line going to the gadgas exchanger, usually a manually-operated butterfly valve.

OJohn M. Campbell & Company


12.51


The operator throttles this valve sufficiently to make gas flow to the reboilers as needed. The key to positioning this valve is the position of the demethanizers bottom temperature control valve. If the temperature control valve is driven wide open by its temperature controller, it is probably because there is not enough warm gas flow available. In this situation, the operator should begin throttling the butterfly valve (in small increments) until the control valve is positioned about 50% open. (Remember that it will take time to see a change in bottoms temperature, so patience is called for in working to reposition the butterfly valve.)

7

On the other hand, if the temperature control valve is running with a small opening, the butterfly valve is probably too severely restricted. If this is the case, too much gas will be by-passing the gadgas exchanger and overall plant performance will suffer.

Additionally, there may be manual by-pass lines around the reboiler and side reboiler that permit the operator some flexibility in varying the relative flows through the two exchangers. In general, however, the bottom reboiler is the preferred point to add the heat and the side reboiler is the "trim" point. "Trimming" in this context means to adjust the side reboilers heat input to yield a preferred tower temperature profile. So, the bottom reboiler should not be by-passed at all but the side reboiler may be according to the situation.

Ethane Rejection Ethane is the lightest NGL product extracted and recovered in an expander plant. The recovery

of ethane is frequently marginal from an economic standpoint. In other words, extraction costs (including shrinkage) are often not covered by the net-back product price at the extraction plant. When this is true, the decision may be to reject ethane into the residue gas at the demethanizer. This is accomplished by raising the reboiler temperature and driving the C2 overhead with the C1 and COZ. Reboiler temperatures vary but typically range from 150-200°F. 7

When designing a facility for ethane rejection, two considerations must be addressed. The first is the reboiler heat medium. A high temperature heat source will be needed. This is frequently recompressor discharge gas. Piping design and equipment and materials specification should address this fact.

Secondly, high propane retention is desired during C2 rejection. This requires better separation than in a demethanization mode. The average relative volatility of C1 to C2 in the DeCl from previous example is 8.5, the average relative volatility of C2 to C3 is 4.8. The difficulty of separation is roughly proportional to the logarithm of the relative volatility. This means that 30-40% additional stages are required in the ethane rejection to achieve the same bottom specification (3% light key in the heavy key). Most demethanizers are designed to operate in an ethane rejection mode (de-ethanization) by installing these additional stages. Some demethanizers have been designed with reflux capabilities, but this is not common. Two examples of reflux schemes are shown in Figure 12.27 and 12.28.

Figure 12.27 is the minimum cost method which utilizes the lower temperature expander discharge vapors to condense a portion of the warmer "stripped" vapors for reflux. The method in Figure 12.28 utilizes a second expander to condense a reflux stream which is then pumped back to the demethanizer top tray. This method gives superior separation but it is more expensive both in capital and operating costs. Reference 12.7 provides an excellent discussion of these and other ethane rejection schemes.



f-


I RECOMPRESSOR To

I I SALES

f

INLET

Figure 12.27 Typical Process Flow Diagram - Cryogenic Expander Plant (Reflux Exchanger Method)

I

Figure 12.28 Typical Process Flow Diagram - Cryogenic Expander Plant (Two Expander Method)



12.53

m


CRYOGENIC PLANT OPERATIONS

Operating Objectives The operating objectives for the cryogenic plant can be summed-up in three words: Quality,

Profit and Production.

Quality refers to the unit's ability to make specification product. It is the prime operating objective because it makes no sense to produce a product that cannot be sold. The other side of this coin is to make products that meet their specifications without exceeding those specifications by wide margins. It makes no sense (and increases costs) to make products that are more "pure" than they need to be.

Profit is our reason for being in the business. Plants that don't operate profitably don't operate for long. Profits are realized when the revenues generated from the sale of quality products are in excess of the cost of producing those products. Operating costs are lowest when the plant runs in an "optimal'' manner.

Production of quality products should be maximized to generate greater profit. Generally speaking, the greater the production, the greater will be operating profits. We must still make quality products at the lowest possible cost, however.

Plant operation is fully optimized when we produce the greatest possible amount of products that meet their sales specifications without being overly pure, and these products are being made in the lowest cost manner. Plant optimization is the job of the plant engineer and includes performance tests and computer simulations (neither one without the other is sufficient). The plant engineer studies various possible outcomes and then presents recommendations to the operating management. Opera- tions management takes the recommendations and develops operating guidelines for the plant's operators. Operators work to stay within those guidelines while making the maximum amounts of quality products at the lowest possible cost.

r7

Each of these operating objectives and the things affecting them are discussed for the turboexpander plant in the paragraphs which follow.

Production - The volume of liquids produced in the cryogenic plant (its recovery level) depends primarily on the feed gas composition, that is, the amount of "potential" product it contains. For gas of a given composition, the level of liquids recovery depends on the temperature to which the feed gas is chilled in the exchangers and turboexpander. More liquids will condense at lower temperatures, which corresponds to removing the maximum amount of heat energy in the turboexpander. The turboexpander will extract the greatest amount of heat when:

This is achieved for a given feed flow rate by increasing the brake compressor's discharge pressure until the expander begins to slow down. When this speed reduction is evident, the operator then backs the discharge pressure down a bit so that full speed is maintained. The feed gas pressure drop across the expander is maximized. Gas pressure at the expander inlet should be kept as high as possible and the pressure at its outlet should be as low as is practical. Operators should not pinch the butterfly valve that diverts warm feed gas to the demethanizer reboilers too severely; doing so creates an excessive pressure drop

the demethanizer pressure. Maximum liquid condensation occurs when demethanizer

The brake compressor is fully loaded.

that is better taken in the expander. The expander outlet pressure is essentially the same as 7 @John M. Campbell & Company


CRYOGENIC PLANT OPERATIONS

f-

pressure is at its lowest; however, excessive methane condensation is to be avoided, since removing it will be more costly.

Operating Costs - The main operating expense in the turboexpander plant is the cost of recompressing the residue gas. Lowering the demethanizer pressure increases the cost of compression. There is a point where the additional liquids produced by lowering the demethanizers pressure aren't sufficient to pay for the increased cost of compression.

The second largest operating cost in turboexpander plants is found in those plants that must use an external heating medium to reboil the demethanizer bottoms. This cost increases as the amount of condensed methane increases, since more heat must be added to vaporize the excess methane.

Additional operating costs are attached to those plants which use refluxed demethanizers. These costs include the capital cost of installing the additional equipment (which cost is amortized over a number of years and treated as an operating expense) plus the increased cost of the process heat required to revaporize the reflux liquid. Remember, everything that goes in as reflux will be revaporized and will come out the top of the tower; and it takes heat to vaporize that liquid. Operators should guard against excessive reflux or the increased fuel cost may be greater than the increased revenue derived from greater product recoveries.

Additional "controllable" costs common to turboexpander plants include those expenses attrib-

costs attributable to "over purification" of the NGL product which means removing more

preventable downtime and maintenance, including the opportunity cost of "lost" NGLs that

the cost of consumables such as methanol whose usage increases as gas dehydration effec-

the cost of replacing damaged equipment such as aluminum plate-fin exchangers that have

Other cost categories common to operating any process include wages, insurance, administra-

utable to:

methane than necessary to meet the specification.

go down the residue gas line.

tiveness decreases.

become plugged with mole sieve dust.

tive charges and the like.

Quality - The quality of the NGL product is determined by its methane content. There are two ways of stating this specification, one refers to the amount of methane as liquid volume percent of the entire NGL product and the other is given as a percent relative to the amount of ethane in the NGL. Regardless of how it is measured, operators should strive to make specification product. Operators should also, however, keep the amount of contained methane right up to the specification limit. Doing so yields the dual benefit of reducing operating costs and increasing production:

operating costs are reduced because it takes less process heat to produce NGL of higher methane content. This is illustrated by Figure 12.29. production goes up because of both the additional volume provided by the methane that is there and because the amount of ethane lost from the top of the tower will be at a minimum when the amount of methane in the bottom is at a maximum.

The methane content of the NGL is determined by the bottom temperature of the demethanizer tower.

r" Demethanizer Control - The ability of the demethanizer to consistently produce NGL of the Most processes desired quality depends directly on the operator's ability to control the process.

@John DI. Campbeii & Company


~~

12.55


Specification Maximum I-

4- Selling Price of Spec. Prodct

Maximum Profit at Specification Limit

Operating Costs

Figure 12.29 The Economic Benefit of "Pushing" Bottoms Product Purity Specification

contain adequate amounts of instruments and controls to do the job, but only if the operator under- stands their purpose and how they should work together. 7

The process controls in the turboexpander plant should work together to promote smooth demethanizer operation. Smooth demethanizer control is essential to optimizing tower operations. The better the control, the more confident operators will be in pushing product specifications. The one prerequisite for smooth fractionator control is smooth control of the tower's feed stream. The same is true of demethanizers; the incoming feed streams must not be erratic.

The things most greatly affecting demethanizer feed flows are the turboexpander(s) and the

Turboexpanders are usually controlled with pressure controllers on their inlet separators. If the feed flow coming into the plant is erratic, plant pressures will not be stable. Two things must be done to insure stable expander operation: first, upstream units must be operated in as stable a manner as possible and second, the expander suction pressure controller must be tuned to dampen pressure surges. The feed separator itself provides a cush- ion to pressure surges and this can be taken advantage of, provided the pressure controller is not tuned too "tightly". It is better to let the separator take some pressure swings and keep the expander inlet flow stable rather than doing things the other way around. Multi- ple turboexpander installations, both series and parallel, present even greater challenges to smooth control.

various separators which dump liquids to the tower.

Liquids that condense in the cryogenicplant upstream of the turboexpander will end up in an expander feed separator. These liquids dump out on level control and feed the demethanizer at various locations. These liquid streams must flow in a smooth, stable manner or they will upset the demethanizer tower. The problem is complicated by the fact that there is usually a large pressure drop involved as the liquids pass through the separator's level control valve; this gives the valve a large ''dynamic gain", meaning that small valve



PROCEDURES

r movements can have large effects on separator level. A swinging level control valve in this application will not just cause liquid surges, but will also produces varying amounts of liquid and flashed vapor to complicate things further. The separator represents a surge capacity in the process and it should be utilized as such. The separator's liquid level controller should be tuned so that the separator level is allowed to fluctuate between upper and lower limits while level control valve travel is minimized. Operating separators as "feed surge dampers" is one of the best ways to minimize demethanizer upsets.

An additional source of potential demethanizer upsets is erratic process heat input. Tempera- ture control loops have notoriously slow response times. Response time is both a fiinction of the physical equipment and the nature of heat transfer. It is not enough to have a properly tuned temperature controller; we must also have operators who are patient enough to let the instruments do their jobs and experienced enough to know when they aren't. While the control problem can be a tough enough challenge in a reboiler that uses heat medium, it can be extremely difficult with reboilers that use warm feed gas for process heat, since fluctuations in feed gas flow, pressure and temperature will severely complicate control action.

Modern cl;irogenic plants often contain very sophisticated control schemes and equipment. Pro- grammable logic controllers (PLCs) are small, dedicated computers that have application to controlling the entire demethanizer tower. They receive a number of inputs such as feed flow, temperatures, pressures and demethanizer bottoms product (NGL) composition. Operators should receive specialized and site-specific training in the care and feeding of such systems.

Profit - Profits will be maximized when all of the above factors are successfully controlled.

PROCEDURES

Caution: The procedures outlined below are only guidelines. Always follow the procedures written specifically for your plant. Specific procedures always consist of details too intricate to be covered by a generic procedure.

Startup Operations

The selection of a procedure for turboexpander plant startup will depend upon the existing temperature and pressure conditions of the plant, the length of the shutdown and whether or not the equipment has been opened to atmosphere. The procedure given here assumes the worst: the plant has been down, depressured and opened for inspection and repair. This condition requires a "total" startup procedure consisting of the following major steps:

1. Dryout

2. CoolDown

3. ExpanderíCompressor Startup

Any time the plant has been down and opened to atmosphere, the first steps in restarting the plant involve purging the equipment and piping of air and then removing the moisture which invariably enters as water vapor in air.



12.57


Dryout takes place after the plant has been thoroughly drained and purged of air. The purpose of dryout is to remove all traces of water vapor that could cause freeze-ups upon restart. Dryout is accomplished by circulating dry gas throughout all areas of the cryogenicplant.

The preferred gas to be used for dryout is dry, lean residue gas that is back-flowed into the plant. If such residue gas cannot be used, feed gas that has gone through mole sieve dehydration can be used. Dryout gas is circulated through the plant and recompressor and is then recycled back through the mole sieve dryers for reuse. The pressure of the dryout gas should be no more than about 300 lbs. The combination of low pressure, dry gas creates a driving force for any liquid water in the plant to evaporate and enter the gas stream as water vapor. In order to minimize the dryout time, the gas flow rate should be as high as possible, so long as the mole sieve capacity is not exceeded.

Both ends of the expander/compressor are blocked-in for the duration of the dryout period and all flow will go through the wide-open J-T valve. (The expander and brake compressor should have been isolated as part of the shutdown procedure and not opened to atmosphere if it could have been be avoided. The machine will be purged and dried with seal gas as part of its startup procedure.) The dryout gas should flow not only through the normal gas flow paths, but also through liquid paths such as level control valves.

The plant is dried out one area at a time, such that there are several phases of dryout:

1. Dryout of the main gas flow path by flowing dryout gas through the gadgas exchanger, expander feed separator, J-T valve, the Demethanizer top and back through the gadgas exchanger.

2. Dryout of the gas flow paths through the demethanizers side reboiler and bottom reboiler. This is done by closing the butterfly pressure differential valve and directing warm dryout gas through the gadgas exchanger by-pass to the reboilers. This gas ends up in the expander feed separator and then goes onward through the rest of the plant as in phase #1.

3. Dryout of the lower portion of the demethanizer. Some plants have a dryout line provided at the bottom of the demethanizer which permits dryout gas circulation through the entire tower. In the absence of such a line, the demethanizer bottoms can be dried by venting the bottom of the tower while it is set up for gadgas exchanger dryout as described in phase #1. While drying the demethanizer bottom, each reboiler should be vented to remove any moisture found there.

As dryout continues in each phase, all low spot drains on equipment being dried should be opened and vented a short time to remove any trapped liquid water. Also be sure to vent any dead legs in all affected piping. Each phase of dryout should continue until a moisture analyzer sampling the residue gas reads "used" dryout gas that is as dry as that coming out of the dehydrators. Each plant's initial startup manual will contain a detailed dryout procedure.

Cool Down

When the plant has been sufficiently dried out, high pressure feed gas is introduced and the plant gradually cools down as J-T valve operation commences. Cool down continues until the plant is as cold as it can get with the J-T valve. The cool down phase continues until the expander is ready to start.

The transition from the dryout phase to the cool down phase is a smooth one. The gas flow is initially set up as in phase #1 of the dryout. The pressure controller which operates the J-T valve is placed into service and the flow of feed gas is gradually increased. The dryout recycle line from the recompressor discharge is gradually closed, increasing the flow through the sales gas line. Gradually,



PROCEDURES

as the J-T valve's pressure controller takes over and more feed gas is admitted, the inlet gas pressure upstream of the J-T valve will rise to its normal value.

f-

Initially, all of the inlet gas should flow through the gadgas exchanger. The temperature control valve that takes warm feed gas to the bottom and/or side reboilers is closed and the butterfly valve in the feed gas line to the gasígas exchanger is open fully. As the pressure climbs on the upstream side of the J-T valve, temperatures throughout the plant will start to decrease and condensed liquid will start to collect in the expander feed separator.

When the level in the separator rises to its normal point, the level controller is placed in service and the flow of liquid to the demethanizer commences. Shortly thereafter, a level will appear in the bottom of the demethanizer tower. When enough liquid has collected to fill the shell of the bottom reboiler, the flow of warm feed gas (or heating medium) can be introduced. By this time there should also be a level in the side reboiler so warm gas flow through this exchanger can commence also. When the liquid level in the bottom of the demethanizer reaches its normal height, the bottoms product pump can be started. (Although the bottoms temperature may be cooler than when operating normally, the methane content of the bottoms product should be sufficiently low at this point, since the plant will still be relatively "warm".)

The temperature in the expander separator is the guide for determining when to start the expander. When the temperature goes no lower, the maximum cooling provided by the J-T valve has been realized. At this point the expander is readied for startup, or, the plant may continue to operate as a J-T plant for some period necessary to "line out" plant operation under existing conditions. It provides a good opportunity to ensure the proper operation of all instruments, controls and auxiliary systems, such as methanol injection. r

Expander/Compressor Startup

Expander startup consists of four parts:

1. Establish the flow of seal gas.

2. Establish lube oil circulation.

3. Start gas flow through the expander.

4. Start gas flow through the brake compressor.

Prior to commencing machine startup, all the following are checked o The expander inlet and outlet valves are closed, and the brake compressor suction and

discharge are closed. The control valve in the compressor recycle (anti-surge) line should be open. Gas is flowing through the J-T valve in the expander by-pass line and also through the compressor by-pass.

All gauges, controls and safety devices on the expandedcompressor are operational. All local panel annunciator lights indicate startup readiness status.

F'

Suitable seal gas is available.

The lube oil reservoir contains sufficient oil for startup, the oil reservoir heater is on and the lube oil cooler is operational.

Establish the Flow of Seal Gas - The procedure for commencing seal gas flow follows:

1. Crack open the impeller case drain valves on both the expander and brake compressor and drain any liquid from each casing. Open a vent valve on the lube oil reservoir.



12.59

CRYOGENIC GAS PROCESSiNG ~ ~ ~~

”? 2. Slowly open a valve in the seal gas line and allow seal gas to flow out of the drain valves on the machine cases and also flow seal gas from the shaft housing through the oil reservoir to the vent.

3. After the casings, shaft housing and reservoir have been thoroughly purged, first close the drain valves on the expander and brake compressor casings, then close the lube oil tank vent valve. It is important that the pressure in the lube oil tank be lower than or equal to the pressure inside the machine casings or oil may be pushed up into the impeller casings. Should lube oil get into the impeller casings, in should be drained and the casing flushed with gasoline before starting. Otherwise, the oil will freeze and impair operation.

4. When the expander and compressor casings and oil reservoir pressures have equalized at seal gas pressure, slowly open the compressor discharge valve. This provides an outlet for the seal gas leaking past the labyrinth seals. Set the seal gas flow at its normal rate.

Establish Lube Oil Circulation - The procedure for starting the lube oil system follows: 1. The lube oil system has two pumps, each with a HAND/OFF/AUTO switch. During nor-

mal operation, one pump is operating, with its switch in the HAND position, while the other pump is in standby with its switch in the AUTO position. The standby pump will automatically start if the lube oil pressure drops during operation. Before setting the pumps up for normal operation, it is a good practice to set each pump in the HAND position and observe pump operation to be sure both pumps work satisfactorily.

2. After lube oil circulation is established, check pressures and temperatures in the lube oil system for normal operation. The temperature of the oil flowing from the shaft housing to the reservoir should be about 38°C [100”F]. If the expander is cold at the time of startup, lube oil may be cooled below its normal operating temperature. In this case, check the operation of the oil reservoir heater and lube oil cooler by-pass temperature controller. Oil circulation must continue until a normal operating temperature has been reached.

3. Crack open the drain valves on the compressor and expander cases once again to ensure that no lube oil is leaking into either end of the unit.

Start Gas Flow Through the ExpanderBtart Gas Flow Through the Compressor - The

-l

procedure for starting the expandedcompressor follows: 1. 2. 3.

4.

5. 6.

7.

8.

Veri@ that the compressor’s recycle (anti-surge) valve and discharge valve are open. Open the compressor’s suction valve. At the expander loading station, switch to the manual position and adjust for zero output to the guide vane actuator. Verify that the vanes are in the closed position. Verify that the output of the split-range pressure controller that regulates expander operation is working in upper half of its output range, corresponding to J-T valve control. Open the expander outlet valve. Open the expander inlet valve. A small amount of inlet gas will flow through the expander because the guide vanes do not provide positive shutoff of all gas flow into the unit. The expander will start to run. Slowly increase expander speed by increasing the manual loading to the guide vane actuator. Verify that as the output to the guide vanes is manually increased, the flow of gas through the 3-T valve decreases because the pressure controller’s output to the J-T valve is decreasing. When the output of the pressure controller matches the manual output to the guide vane actuator, switch the manual loading station to the auto position. All gas flow should now be going through the expander (J-T valve fully closed).

7 @John M. Campbeii & Company

~ ~ ~ _ _ _ _ _ _ ~


PROCEDURES

f- 9. Place the brake compressor in operation by slowly closing its recycle (anti-surge) valve. As the valve closes, the compressor will start to place a load on the expander. When the recycle valve is fully closed, switch to automatic anti-surge operation.

The time it should take to manually bring the expander up to normal operating speed will depend on whether or not the plant is cold at the time of startup. If the plant was warm, the time involved should be 30 to 60 minutes, to allow instruments and equipment time to adjust to the changing temperatures. If the plant is cold at the time of startup, the expander can be brought up to speed more quickly.

The time it takes to perform an overall startup of the turboexpander plant may vary from a few minutes to a few days, depending upon whether the plant is cold or warm, whether dryout is necessary, the degree of automation, and so on. A longer startup time is required for cool down if equipment in the plant has had a chance to warm up to ambient temperature. If the cool down is performed too quickly, some damage to piping or insulation from sudden metal contraction may occur. On the other hand, if startup commences a few minutes after a shutdown, and all equipment is cold, no cool down is necessary and the expander can be placed into service as soon as the seal gas and lube oil flows are established.

Cold Expander Startup Procedure A cold expander startup may occur as a restart after a short power outage, since the lube oil

pump will stop and the low oil pressure trip will shut down the machine. Special precautions must be observed when starting a cold expander. In the case of the power outage and shutdown trip, lubrication will be provided during rundown by the pressure in the lube oil system's pressure pot. After rundown, however, there will be no lubrication. r"

With no warm oil flowing through the bearings, the cold expander casing and impeller will slowly cool the shaft and bearing on the expander end of the machine. In a matter of a few minutes, the lube oil in the bearing may become so viscous that no oil will flow to the bearing when the lube oil pump is restarted. In this situation, it may even be necessary to warm the bearing by applying external heat in the form of hot water or steam, until the bearing temperature is high enough for lube oil to circulate through it, but try the following procedure first.

The procedure for a cold expander startup when the bearing temperature is 15°C [60"F] or

1. Start the flow of seal gas. Be sure the compressor discharge valve is open so that there is an outlet for seal gas.

2. Set the thermostat or temperature controller in the lube oil tank at 93°C [200"F]. Check to see that the heater in the lube oil tank has turned on.

3. After a few minutes, move the setpoint on the lube oil temperature controller back down to about 55°C [130"F] and start the lube oil pump.

4. When the expander bearing temperature rises to 38°C [100"F], lower the heater thermostat set point back to its normal setting.

5 . Start the expander according to the procedure in the preceding segment.

colder, follows:

Normal Operations Normal operation of the turboexpander plant mainly involves making routine operating checks.

Any number of process variables may be recorded or trended and the operator should monitor these to

f /"



12.61


gain early warning of impending problems. A list of some of the things that should be routinely checked in the cryogenic plant follows:

O O O

O O O O O

O O O O

O O O O O O O O

O

O

O

Inlet gas temperature, pressure and flow rate Gadgas exchanger gas temperatures in and out

Gasígas exchanger AP Expander feed separator temperature Expander feed separator liquid level Pressure controller output to Expander Guide Vanes / J-T Valve

Expander inlet screen AP

Expander A€' Expander/compressor lube oil pressure and temperature Expanderícompressor seal gas pressure and flow Expander thrust bearing pressures

Expander/compressor lube oil filter AP Brake compressor suction and discharge pressures Temperatures on shell & tube sides of the demethanizer reboilers Demethanizer bottom temperature Demethanizer bottoms level Methane content of NGL Demethanizer pressure Demethanizer top temperature

Demethanizer tower AP Demethanizer top (demister) A€'

AP and AT on all other heat exchangers Similar readings on all additional separators, expander compressors, etc., plants

in larger

The most common source of problems, by far, for the normally operating cryogenic plant are freeze ups caused by either inadequate gas dehydration or COZ solidification. Excessive differential pressures (AP) are the first warning signs of freezing or hydrate problems. Frozen solids commonly build up in heat exchangers and anywhere the feed gas can experience a pressure drop, such as across the expander inlet screen, separator and demethanizer demister pads plus liquid level and J-T control valves. These areas should be constantly monitored, especially if there are no high A€' alarms in problem spots. Prompt action (discussed under Troubleshooting) can prevent poor plant performance and shutdowns.

The second most common problem to crop up in otherwise normally operating plants (it may be your plant's biggest problem!) is that of unstable feed flows. Unstable flows can cause erratic expander operation, especially when so fully loaded that the J-T valve starts to swing open and closed. Unstable feed gas flows also upset heat transfer in the demethanizer side and bottom reboilers. There is usually little that operators can do about this problem but everything that can be done upstream should be done. Otherwise, operators may be able to keep the NGL product on spec by running with high demethanizer bottoms temperatures, but recovery efficiency will suffer.

rl



PROCEDURES

r" J-T Valve Operation (Expander Down) During those times when the expander must be shut down for routine maintenance or other

repairs, or if the expanderícompressor should trip and shutdown for some reason, the plant can continue to run in a J-T mode of operation. Operators can expect to make somewhere between 40 to 60% of normal liquid production when they go to J-T operation. This is the difference between the expanders isentropic operation and the J-T valve's isenthalpic operation the J-T valve does no work.

When the expander is shut down or trips and goes down, the gas flow automatically goes through the J-T valve in the expander by-pass piping. At the same time, the brake compressor ceases to work and gas will flow through the compressor by-pass.

Temperatures will begin to rise everywhere in the plant due to the difference in refrigeration no longer supplied by the expander. If the recompressor operates on constant suction pressure control (as is usually the case), the pressure in the demethanizer will increase to recompressor suction pressure (plus line losses) and the pressure drop across the J-T valve will decrease, further promoting a general plant warming. It may take a considerable amount of time (hours) for the plant, especially the demethanizer, to stabilize at the new temperatures.

The liquid feed coming to the demethanizer will contain less methane, ethane and propane (the feed will be heavier). The tower pressure will also be higher, as explained above. The combination of higher pressure and heavier feed will result in a lighter bottoms product; that is, the amount of methane relative to the ethane content will actually increase! As a result, demethanizer bottoms temperature set point will have to be increased. If the tower has a bottoms product on-line analyzer this result will be evident. If the analyzer is used to reset the bottoms temperature setpoint, it can be expected to raise

/- that setpoint.

Shutdown of the expander plant begins with expander/compressor shutdown. There are two

1. Depressing the STOP button on the control panel will close the expander inlet (isolation) valve. The machine subsequently runs down in a short time due to the stoppage of flowing gas. Pressure will build on the upstream side of the J-T valve until the pressure controller increases its output sufficiently to begin operating the valve. Demethanizer pressure will eventually rise to recompressor suction pressure and begin to flow through the brake compressor by-pass. This method provides a test, under full load, of the expanderícompressor shutdowns.

2. The operator can first place the expander controls into MANUAL and then gradually begin to unload the machine. As the inlet guide vanes go closed, the inlet gas pressure increases until the J-T valve starts to open. When the guide vanes have completely closed, the expander will still be turning at greatly reduced speed. The machine is stopped now by either manually closing the inlet isolation valve or pushing the console's STOP button as before. The same things happen on the compressor end as explained in the first method, only more gradually. This method is a much smoother method of shutting down the expander/compre~sor.~

Once expander/compressor rotation has stopped, lube oil circulation and seal gas flow can cease, but these may be left on if a restart is imminent, or they may be allowed to run and flow for a

approaches to expander shutdown:

3 There are situations, such as a "thrust unbalance" condition, which require shutting down the expander according to more specific procedures. Operators must be aware of these kinds of conditions and must follow the procedures found in the vendor's manuals.

r"

0 J o b DI. Campbell & Company


12.63


r7 short period as the expander casing warms up. The brake compressor discharge valve must be left open for as long as seal gas is flowing. If the machine is going to be opened or worked on, expander outlet and compressor suction and discharge valves should be closed, and trapped gas pressure vented.

Beyond this, the remainder of the plant is shut down by reducing and then halting the flow of feed gas and closing feed and residue gas isolation valves. The product pump is stopped and the demethanizer bottom is blocked in. Trapped pressure and liquids can remain inside the plant unless lines, vessels or equipment has to be opened. The only ''tricky" part of cryogenic plant shutdown occurs when the plant must be depressured.

The problem of venting the cryogenicplant is due to the low temperatures at which the plant runs. Additionally, the NGL liquids that have condensed have very low atmospheric boiling points. Ordinary flare headers, being made of carbon steel, are not designed for these temperatures. The typical solution is to include in the plant design a stainless steel "cryogenic vent stack". There are hazards associated with such a stack.

1. The vent stack vapors may not be burned as in a flare but released directly to atmosphere. The use of a vent stack is therefore infrequent and only handles full flows under emergency situations where rapid venting is less of a hazard than the alternatives.

2. If a small amount of vapor leaks from process equipment, such as PSVs, during normal operation, the cold gas may not disperse well until certain atmospheric conditions and may become a ground-level fire hazard.

3. If stack effluent should ignite during venting, any personnel in the immediate area may be burned by flame radiation or ground level flashing vapor.

4. Accumulations of vented gases, if ignited, may fuel an "unconfined vapor cloud explosion" which can create damaging pressure waves.

Low pressure areas of the cryogenicplant, being constructed of stainless steel, can vent directly to the cryogenic vent stack. Certain high pressure areas of the plant, however, may contain NGL liquids and may not be made of materials that can withstand the low temperatures resulting from venting directly to the atmosphere. It is a common practice therefore, to vent the liquids inside these high pressure vessels to the demethanizer tower. Here they can safely boil off to the cryogenic vent.

-I

Be sure to follow the authorized procedures at your plant for safely venting the cryogenic unit.

CRYOGENIC PLANT TROUBLESHOOTING The most common problem in the cryogenic plant is a freeze-up. Freeze ups may be caused by

hydrate formation or COZ solidification. The next most common problems involve either a reduction in the volume of liquid product produced or else the NGL goes off-spec (contains too much methane). The first step in troubleshooting is to be certain that a problem exists, by checking the instruments involved to be sure that they are not in error. Beyond this, troubleshooting involves checking all possible causes of the problem through the process of elimination.

Severe operating problems can be avoided if they are promptly noticed and identified, and prompt corrective action is taken. Operators need to know what to be on the lookout for, however. One of the best indicator of cryogenicplant operation is the expander feed separator. The temperature in this separator should remain fairly constant as long as the inlet gas composition and flow rate are also fairly constant. A change in separator temperature signals the development of a potential problem.

r7



CRYOGENIC PLANT TROUBLESHOOTING

r An increasing temperature in the expander feed separator can be caused by: A freeze-up in a heat exchanger. A vapor-locked demethanizer reboiler. An open J-T valve is by-passing gas around the expander. There is a problem in the expander. Excessive feed gas flow to the demethanizer reboilers.

A decreasing temperature in the expander feed separator can be caused by: Snowballing temperatures Insufficient warm feed gas flow to the demethanizer reboilers.

Each of the above problems is discussed below.

Freeze-U ps In most cases, either the dehydration unit is not sufficiently lowering the water dewpoint of the

inlet gas or COZ is solidifying. Other causes of solids or semi-solids formation include amines, glycols and compressor lube oils, although these will form upstream of the expander, whereas hydrates and COZ blockages can also form downstream of the expander. Freeze-ups can be identified by high differential pressures across affected equipment.

Hydrates - The usual first method of attacking a hydrate plug is by injecting methyl alcohol, methanol. Methanol will dissolve a hydrate provided the methanol liquid can reach the plug. As long as there is enough flow to carry the methanol to the plug it will do its job. This may be next to impossible to do in horizontal flow passages since there is no flow through a plugged passage. If the methanol can travel to the plug by gravity flow, it will also do the job. When methanol injection is unsuccessful, a derime procedure must be followed (See below).

rc'

Carbon Dioxide, COZ - Both liquid phase and solid phase CO2 can exist in various cryo- genícplant areas, depending upon the pressure, temperature and concentration of the carbon dioxide. The solid C02 is the problem phase. Figure 12.30 is a chart that has been created to help predict the possibility of solid COZ formation.

To Use the Solid COZ Formation Chart: If operating conditions are in the methane liquid region as shown by the insert graph, the dashed solid-liquid phase equilibrium line is used. For other conditions the solid isobars (lines of constant pressure) define the approximate COZ concentration limits. For example: Find conditions of 300 psia and -170°F on the insert graph; this point is in the liquid methane region. The dashed solid-liquid phase equilibrium line indicates that concentrations of at least 2.1 mol % COZ in the liquid phase would be likely to form solids. However, at the same pressure and -150°F, conditions are in the methane vapor phase, and 1.28 mol % C02 in the vapor is the minimum concentration that could lead to solids formation. It seems that liquid methane inhibits solid COZ formation.

The most likely trouble spot for solid COZ formation is in the demethanizer. If expander liquid is fed to the top tray of a demethanizer (non-refluxed) the COZ will concentrate in the top equilibrium stages. This means that the most probable condition for solid C02 formation may be several trays below the top of the tower rather than at expander outlet conditions. The COZ will plug the section of trays (or packing), causing a high AP through that section of the tower,

Methanol cannot remove a solid COZ plug. It won't melt solid COZ like it can melt ice or a (f- hydrate. Solid COZ removal must be removed by deriming (described below).



12.65


Temperature, "C -70 -75 -80 -85 -90 -95 -100 -105 -110 -150 -120

50.0

40.0

30.0

20.0

10.0 9.0 8.0 7.0 6.0

5.0

4.0

3.0 8 - O E

8 2.0 o

1 .o 0.9 0.8 0.7 0.6

0.5

0.4

0.3

0.2

0.1

Figure 12.30 Approximate Solid COZ Formation Conditions



r7

rl


Derime Procedure - Derime is the name given to the process of removing ice (or hydrate, or solid COZ) buildup on or in equipment. Rime is defined as a mass of tiny ice crystals formed from vapor. This is a good description of how solid deposition starts out in the cryogenicplant. Derime literally means "to thaw out". So, if we can't remove a solid plug with methanol, we have to warm up the plant (or lower its pressure, or both) until the solid melts.

Derime procedures commonly involve shutting the plant down and back-flowing warm residue gas through to the flare, while injecting methanol. The exact procedure depends upon the plant, so operators should follow their site specific procedures. Some plants have piping installed that permits deriming the demethanizer without going through a complete shutdown. This arrangement takes warm NGL liquid from the discharge of the bottoms product pump and introduces it onto the top tray of the tower, from which it will run down and melt whatever solids have deposited on the trays.

Vapor-Locked Demethanizer Reboilers

Liquid from a draw-off tray in the demethanizer flows under the influence of gravity into the side or bottom reboiler. The liquid head on the draw-off tray is sufficient to move the liquid through the piping and into the exchanger. The liquid enters on the bottom side of the shell and flows upward, around the tube bundle (or through a section of a plate-fin exchanger). As it fiows through, the liquid is partially vaporized by the warmer fluid on the other side. The rising vapor bubbling through the remaining liquid lifts it up through the exchanger outlet and back into the tower; this is the thermosyphon effect. The combined vapor-liquid stream returns to the tower below the draw-off tray.

If something should happen to increase the pressure drop in the piping and through the exchanger, the liquid level on the draw-off tray will back up until it overflows into lower areas of the tower. If this happens in the demethanizer tower, cold liquid will enter the bottom and cause the methane content to increase. Furthermore, there will be no heat input from the "locked" reboiler, so the tower bottom temperature falls further and methane content rises higher. An important side effect is that the warm feed gas passing the other side of the reboiler is not cooled down. This still-warm gas travels back to rejoin the rest of the feed stream but warms it up, causing reduced liquid condensation. Attempts to cram more warm feed gas through the reboilers just makes things worse.

f-

Reboilers may be locked by excess vapor production ... a gas "bubble" forms inside the exchanger, becomes trapped there and keeps new liquid from entering. Solids in the tower can also plug the reboiler inlet line to create the same effect, but a vapor lock is the most common cause. The simplest way to clear a locked reboiler is to shut off the heat. Vapor production will falter and liquid can reenter the reboiler. Slowly readmitting the warm stream will get things going again.

Sometimes an injection gas line is tied in at the base of the line returning to the tower. A small flow of injection gas will create a lifting action to get the liquid moving again. The success of injection gas is evidenced by rising temperatures in the reboiler and tower and the injection gas is then gradually shut off.. Injection gas should be warm, lean gas such as high pressure residue gas.

Demethanizer Temperature Control

Basically just a few things can go wrong here, as evidenced by temperatures that are out of control and either too warm or too cold. A temperature control valve can get stuck in either the open or closed position, the butterfly valve in the main gas stream may be too far open or too far closed, or the temperature controller itself may fail. The operator will have to perform a visual inspection and attempt to solve the problem. (The best way to ferret out a controller problem is to put the controller

r @Job M. CampbeU 8s Company


12.67


in manual and attempt manual control. If you can control manually you have eliminated all the other possible causes and the controller is the problem.)

If the problem with temperature control is that the temperature is too low, be certain that you do not have locked reboilers before blaming the temperature controller.

Turboexpander-Brake Compressor Practically every potential problem with this machine is annunciated and alarmed at the ex-

pander control panel. The best advice we can give here is to refer to the vendor's manual for the troubleshooting procedures for your particular machine.

Snowballing Temperatures Snowballing is an interesting phenomenon that gets its name from what happens to a little

snowball as it rolls down a snow-covered hill it gets bigger and Bigger and BIGGER. This particular cryogenicplant problem is like the snowball in that once started, the problem grows and grows. In the case of snowballing in the cryogenicplant, temperatures start out getting too cold and keep on getting colder and colder in a constantly ampliQing cycle. An explanation of snowballing follows:

If the expander feed separator temperature starts to decrease, the temperature of the gas leaving the expander will also decrease, which in turn lowers the temperature of the demethanizer overhead gas, which causes greater chilling in the gadgas exchanger, which further decreases the temperature in the expander feed separator, and so on. The result is more liquid production than the demethanizer can handle. Some of this liquid can carry over in the demethanizer overhead and finds its way to the gas to gas exchanger where it vaporizes and, like a good refiigerant, chills the inlet gas even further. This of course results in even more liquid production and the race is on.

-I

The net results of snowballing are colder temperatures in the plant and a decrease in NGL production due to the large amount of liquid that leaves from the top of the demethanizer and ends up "lost" in the residue gas leaving the plant.

The method used to stop snowballing is to slow down the expander (or even shut it down) to reduce chilling in the plant and halt excess liquid production. Slowing or stopping the expander will divert gas through the J-T valve and cause the plant to warm up. When things are back to normal, the expander can run again.

The cause of the problem is usually found at the demethanizer reboilers when the bottoms temperature controller does not allow enough gas flow to the reboilers. The net effect is colder gas entering the expander feed separator. If the problem persists, it may be necessary to raise the plant's pressure control setpoints.




REFERENCES 12.1 12.2 12.3 12.4 12.5 12.6 12.7

Pehnec, S. T., "Production and Processing Operations," John M. Campbell and Company (May 1995). Linhardt, H. D., LNGKryogenics (Feb. 1973), p.7. Campbell, J. M., "Gas Conditioning and Processing, Vol. 2: The Equipment Modules, 7th Ed. (1994). Engel, Carl G. and W. Rosen, Cryogenic Gas Plants, Petroleum Laming Programs, Ltd. (1983) GPSA Engineering Databook, 10th Ed., p.13-43. Lowe, R. E., Chem. Engr. (Aug. 17, 1987), p.131. Evans, "Flexibility Can Boost Profits in Cryogenic Gas Processing Plants, " oil & Gas J. (July 14, 1980), p.76-84.

@Job M. Campbell 8s Company


12.69


NOTES:



ill

O John M. Campbell and Company

CLASS PROBLEM #1 f-

Overview

Determine the molar flow (kmolih or l b m o l h ) and mass flow (kgh or I b h ) for the following gases. Also calculate the Relative Density and the Ethane-plus Liquid Content in gailons/Mcf (mm).

Gas is water saturated at inlet conditions. 1



CLASS PROBLEM #2

Based on the Gas A inlet gas analysis from Problem #1, what type of process may be employed for recovery of ethane and heavier hydrocarbons?


6"

CLASS PROBLEM #3 Overview

Based on the following investments and product recoveries when processing Gas A, which process will have the best pay out?

I ~ " ^ _ - _II

Installed Cost

Product Recovery (%)

Carbon Dioxide

Ethane

Propane

Butane plus

Plant Fuel (Mcfd)

t

Use 2001 average Mont Belvieu prices of:

_I--___

$/gallon

0.24

0.40

0.44

0.45

0.51

. - -. - i

I-- -- I---_

i I Ethane

- _ _ l _ l - - _ l ~ r-

ct 5$/gal T & F from all prices

And a gas price of $2.75/MMBtu.

Base pay out on the processing margin:

0 Value of Residue Gas (after fuel)

Cl + value of Liquids F.O.B. Mont B.

i3 Less: Value of Inlet Gas ~ ~ ~~

Technical Assistance Service for the D e s i m b M. & comPmY 13.3 Operation, and Maintenance of Gas Plants

CLASS PROBLEM #4 Phase Behavior

Print out the phase envelope for Gas A and identi@ the Cricondentherm, Cricondenbar and critical point. If the gas is separated at 2°C [35"F], what pressure will yield the maximum liquid condensation?

Oio-ii & Corn= 13.4 BP Exploration Company (Columbia) Ltd.


CLASS PROBLEM #5 t-

Water-Hydrocarbon Behavior

Determine the water content of the following gas at 13800 kPa [2000 psia] and 49°C [ 120"FI by using:

a) The McKetta and Wehe or Campbell charts (Figures 4.2 and 4.3 respectively)

b) Equation 1.3, Figures 4.4 and 4.5

c) The Wichert method (Figure 4.6)


lm I ,

O John U. Campbell and Company

CLASS PROBLEM #6 Water-Hydrocarbon Behavior

A gas stream leaves a free water knockout at 20 O00 kPa [2900 psia] and 50°C [ 122"FI. Some distance downstream, the gas enters a separator at 6900 kPa [ 1000 psia] and 10°C [SOOF]. How much water should be drained from the downstream separator?


O John Y. Campbell and Company

CLASS PROBLEM #7 Water-Hvdrocarbon Behavior

Determine the hydrate formation temperature at 6900 kPa [ 1 O00 psia] for a gas with the analysis shown in Class Problem #5.

Using the method of Katz et.al.

Using the method of Trekell-Campbell

Plot the hydrate locus using HYSYS, characterize iC5+ as Cg


13.7


CLASS PROBLEM #8 W a t er-H yd roca rbon Be havio r

2.82 x lo6 m3 (std)/d [ 100 MMscfd] of natural gas having a hydrate formation temperature of 15.6"C [60"F] cools to 4.5"C [40"F] in a buried pipeline. Minimum pipeline pressure is 6200 kPa [900 psia]. The concentration of lean DEG inhibitor is 75 wt%. What volume of inhibitor solution must be added daily if the gas enters the line at 6900 kPa [ 1 O00 psia] and saturated with H20 at 32°C [90"F] using:

a) DEG

b) Methanol (1 00% solution)


W ' '


i"

CLASS PROBLEM #9 Glvcol De hvd ra t io n

141 3 x 1 O3 m3 (std)/d [50 MMscf/d] of a 19.1 MW sp. gr. natural gas enters a TEG contactor at 30°C [%OF] at 6200 kPa [900 psia]. The gas is saturated with water. The outlet water content specification is 50 kg H20/106 m3 (std) [3.11 lb H20/MMscfJ. Calculate the following:

I . Water dewpoint specification

2. Recommended lean TEG concentration

3. The lean TEG circulation rate (circ ratio = 25 liters TEGkg H20 [3 gal TEGílb H20]

4. No. of bubble cap trays

5. Height of packing

6. Contactor diameter

7. Lean TEG concentration required if inlet temperature is 35°C [95"F]. Assume pump capacity is 1.1 m3k [5 US gpm]

8. Reboiler duty

9. Firetube area

1 O. Still column diameter


13.9


CLASS PROBLEM #1 O Solid Bed Dehydration

Refer to class problem #I in the "Glyco Dehydration" section. An adsorption dehydration system is to be installed in place of a glycol unit. The inlet gas temperature is 35°C [95"F]. All other conditions remain the same. A 2-tower system with 12 hour adsorption cycles has been selected.

Calculate the following:

1) Determine which type of desiccant should be used

2) Adsorption tower diameter (assume 1/8" mol sieve beads)

3) Adsorption tower bed height (use an effective water capacity of 10%)

4) Mass of desiccant

5) Regeneration heat load (assume 5 hours heatingí3 cooling)

6) Regeneration gas rate

7) Regeneration heater duty 3



CLASS PROBLEM #11 Refrifleration

P

Given Gas B inlet gas conditions shown in Problem No. 1 , design a straight-refhgeration plant using the following guidelines:

a) Inlet gas stream is water saturated. This is approximately 961 kg H20/106 m3 [60 lb HzO/MMscfl inlet gas.

The Gas-Gas Exchanger (E-201) has 35 kPa [5 psi] pressure drop for both shell and tube passes and the hot end approach temperature is 6°C [lO"F].

The Gas-Liquid Exchanger (E-202) has 35 kPa [5 psi] pressure drop for both shell and tubes passes and the liquid is heated to give a -29°C [-20"F] tower feed.

The Chiller (E-203 has a 35 kPa [5 psi] pressure drop for the tube side (process side) and cools the inlet gas to -34°C [-30°F].

The raw NGL (unstabilized) is fed to the top of the Stabilizer (T-501).

The Stabilizer (T-501) operates with tower pressure of 1750 kPa [255 psia], 35 kPa [ 5 psi] pressure drop, 1 O theoretical stages and an Ethane specification of 0.06 mol ratio C*/C3 in the NGL Product. The Reboiler is E-204.

The Stabilizer Overhead is recompressed and either:

b)

c)

d)

e)

f)

g)

Recycled to the inlet

Mixed with Residue Gas

Choose the most economical.

h) A Product Cooler (A-301) is used to cool the NGL to 46°C [ 120"FI. Assume a 70 kPa [ 1 O psi] pressure drop.


13.1 1

O John Y. Campbell and Company

Class Problem #I 1 (Cont'd.) Ref rige ra t ion Systems

Determine the following information for this plant:

Plant recoveries, liquid production m3/d [gal/day], and recovery efficiency as % of inlet.

Residue gas Cricondenthenn.

Inlet gas, residue gas and stabilizer overhead GHV [Btu/scfl.

LPG true vapor pressure at 38°C [lOO"F].

Duty and surface area for the following heat exchangers, given the listed overall heat transfer coefficients.

I U0 1 1 i 1 Exchanger 1 kJ/(h.m'.OC) ;

55

90

90

The inlet/outlet temperatures from the non-process side are shown below:



CLASS PROBLEM #12 Refrigeration Systems

Design the refrigeration loop based for Problem #11 (with recycle of Stabilizer overhead), on the following assumptions:

a) Propane refrigerant composition of

1

2

.-

i-Butane

n-Butane "_l 111 r ----

Total [ 100

r" b)

c)

d)

e)

f)

The Chiller duty (E-203) from Problem #6 plus 5% heat loss.

The condensing temperature if 49°C [ 1 20°F] in A-302 with a 35 kPa [5 psi] pressure drop.

Set the pressure of the Economizer (V-406) to balance the FirstlSecond Stage ratios.

The Chiller (E-203) has a 3.5 kPa [ O S psi] pressure drop, and the vapor out is at -35°F.

Assume the compression (C-102 N B ) has a 65% adiabatic efficiency.

Based on these assumptions, check the following information:

a)

b)

c)

Second stage discharge pressure. Compare with Figure 1 OA. 1 (b)

What horsepower is required for the Refrigerant Compressors (C- 102 N B ) ?

For extra practice, calculate the Refrigerant Compressor horsepower (C-102 A B ) without the Economizer (V-406).

r" F


13.13


CLASS PROBLEM #13 Cryoclenic Gas Processincl

Given Gas A at the inlet conditions as described in Problem No. 1, design an ISS or GSP plant using the following guidelines:

a) All heat exchangers have 35 kPa [5 psi] pressure drop, except tower reboilers.

b) Economic optimum ethane recovery is desired.

c) Residue gas must be delivered at 5520 IP [800 psig].

d) The NGL pipeline specification is 1.5 LV% methane in the ethane and 0.35 LV% COZ in the ethane. Product pipeline pressure is 6892 kPa [ 1 O00 psig].

e) The turboexpander (EC-701E) has an 82% (abiabatic) efficiency and the compressor (EC-701C) has a 75% (adiabatic) efficiency with 2% mechanical loss.

f) The demethanizer has 10 theoretical trays (ISS) or 12 theoretical trays (GSP).

The “handles” to adjust the model are listed below.

a) Cold Separator (V-402) temperature.

b) Product Heater (E-202) process outlet temperature.

c) Demethanizer (T-501) pressure.

d)

e)

Side Reboiler (E-203) duty and balance with bottom Reboiler (E-204) duty.

Split between Expander and Reflux Condenser (GSP).

To confirm the plant design, check the following items:

a)

b)

c) Demethanizer (T-501) temperature profile.

COZ freezing in top of tower.

Heat curves for heat exchangers to maintain minimum (weighted) approach of 2OC [3.6OF]

Once these conditions are met, then determine the following information:

a)

b)

c) Residue gas cricondentherm.

d)

Summarize the checks listed above.

Plant recoveries, liquid production (m3/d [gal/day]), and recovery efficiencies as % of inlet.

Inlet gas and residue gas heating content, GHV in Btu/scf.

Residue compression, expander and compressor power requirements.


tratamiento de gas natural

Documents

gas production

maintenance of gas plants

com section

campbell compnny

com email

overview of oil

written permission of

com pleteness