cover revamps - digitalrefining · this revamp was performed by altairstrickland. the main scope...

r e v a m p sp t q

2009

Supplement to PTQ

cover and spine copy.indd 1 13/8/09 14:18:28

Improving perfection

Improved performanceStability over broader operating ranges - especially for high gas rates as applicable in hydrocrackers

Easy to install Installation can be done in as little as 4 hours per tray. Fewer parts to handle and wedge fastening instead of bolting, resulting in shorter downtime during shut-downs

Compressed design Both trays and mixers enable loading of more catalysts

Massive references More than 450 reactor internals installed in 200 hydrotreaters

Enhanced reactor internals from Topsøe

W W W. T O P S O E . C O M

www.ptqenquiry.com for further information

haldor.indd 1 13/8/09 15:55:53

http://www.eptq.com/jump.aspx?a=Haldor&p=revamps09

©2009. The entire content of this publication is protected by copyright full details of which are available from the publishers. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means – electronic, mechanical, photocopying, recording or otherwise – without the prior permission of the copyright owner.The opinions and views expressed by the authors in this publication are not necessarily those of the editor or publisher and while every care has been taken in the preparation of all material included in Petroleum Technology Quarterly the publisher cannot be held responsible for any statements, opinions or views or for any inaccuracies.

3 Enhancing heat transfer in Texas Towers PeterEllerbyandPeterDrögemüllerCal Gavin Ltd EdgarVazquez-RamirezandGrahamPolleyUniversity of Guanajuato SebastianErlenkämperBP

11 Stepwise simulation of vacuum transfer line hydraulics HarryHaFluor Canada MatthewReisdorfFluor Enterprises AbdullaHarji Fluor Canada

19 Revamping hydrogen and sulphur plants to meet future challenges AdrienneBlume,PatrickChristensen,BrettGoldhammer andThomasYeung Hydrocarbon Publishing Company

27 A novel process to reduce aromatics and benzene in reformates MikhailLevinbuk,AMeling,VZuberandALebedevMoscow Oil Refinery VKhavkin All-Union ScientificResearch Institute of Petroleum Processing

35 Hydroprocessing upgrades to meet changing fuels requirements JayParekhandHarjeetVirdiChevron Lummus Global

40 CFD study of a VDU feed inlet device and wash bed DebangsuRay Indian Oil Corporation AjayAroraandAnnePhanikumarSulzer Chemtech

44 Application of CFD in NOx reduction

ShahebazMullaandGRamanTechnip USA

LiftofregeneratorheadandcyclonesattheMarathonTexasCityrefinery.ThisrevampwasperformedbyAltairStrickland.Themainscopefortheregeneratorwastheadditionoftwocyclones.Theturnaroundrequired132000man-hoursonAltairStrickland’spartandwasdonewitha0.019weldrejectionrate. Photo: AltairStrickland

ptqYLRETRAUQYGOLONHCET MUELORTEP

Editor Chris Cunningham [email protected]

Production EditorRachel [email protected]

Graphics EditorMohammed [email protected]

Editorial tel +44 844 5888 773fax +44 844 5888 667

Advertising Sales ManagerPaul [email protected] Advertising SalesBob [email protected]

Advertising Sales Officetel +44 844 5888 771 fax +44 844 5888 662

PublisherNic [email protected]

CirculationJacki [email protected]

Crambeth Allen Publishing LtdHopesay, Craven Arms SY7 8HD, UKtel +44 844 5888 776fax +44 844 5888 667

r e v a m p sp t q

contents/ed com copy 4.indt 1 19/8/09 11:45:07

© 2008 Criterion Catalysts & Technologies L.P. (cri815_0508)

www.criterioncatalysts.com

Travel with us, and you’re always on the road to innovation.

T H E C L E A N F U E L S T E A M

cri815_A_roadInnovation_ptq.indd 1 5/29/08 12:07:17 PM


criterion 1 9/3/09 10:13:45

http://www.eptq.com/jump.aspx?a=Criterion&p=revamps09

Enhancing heat transfer in Texas Towers

T exas Towers are vertical shell and tube heat exchangers, used as heat recovery units and

placed around reactors in petro-chemical plants. Since the stream being heated forms the feed to the reactor, from which the hot stream flows, these units are also referred to as feed/effluent exchangers.

In this article, we discuss the design of feed/effluent exchangers, how exchanger design affects operating costs, and how effective retrofit can lead to both significant energy savings and improvement in throughput. A section of flow sheet for a typical unit, together with operating temperatures, is shown in Figure 1.

A feedstock that is liquid in ambient conditions is to be vapourised, heated to a high temperature and then reacted with a gas in the presence of a catalyst. The reaction is exothermic. The reaction does not take place in the absence of a catalyst. It is therefore common practice to mix gas and liquid feedstocks prior to heating. This action has the benefit of reducing the temperature at which the liquid vapourises, thereby giving favourable temperature driving force effects.

Since heat recovery is important to plant economics, the products of the reaction are used to preheat the reactants. The two-phase feed enters the tubes at the base of the exchanger. Within the tubes, the liquid reactants are vapourised, then the gases are superheated before leaving the tower to enter a furnace. Within the furnace, the reactor feed is further heated before it enters a catalytic reactor.

The product stream leaving the reactor consists of a mixture of unreacted feed gases and desired

Retrofit technology enhances heat recovery in feed-effluent exchangers, increases throughput, reduces furnace load and provides cost benefits

PeTeR elleRby and PeTeR DRögemülleR Cal Gavin LtdeDgaR Vazquez-RamiRez and gRaham Polley University of Guanajuato sebasTian eRlenkämPeR BP

products. This is fed to the shell side of the Texas Tower, where it enters at the top and flows downwards. As heat is transferred to the feed stream, the product stream is first desuperheated and then partially condensed. It is then fed to a series of coolers before the desired liquid products are separated from the unreacted gases.

Most catalytic reactors are subject to decay of the catalyst. A common means of maintaining the production rate as the catalyst decays is to increase the temperature of the reaction. Eventually, the operation will reach either the maximum reactor temperature or the maximum firing capacity of the fired heater and the plant will have to be shut down in order to renew or regenerate the catalyst. The load on the fired heater increases across the operating period.

economic assessment of exchanger operationThe obvious way in which the feed/

effluent exchanger contributes to the economics of plant operation is by reducing energy consumption. This benefit is easy to quantify. However, while the saving in energy costs accumulated over the plant’s operating period is substantial, this is not the only benefit. For a fired heater of fixed capacity, the higher the level of heat recovery in the feed/effluent exchanger, the longer the plant’s operating period. This has two major cost advantages: the plant has a higher capacity; and the plant produces more products for a given batch of catalyst.

The economic benefit derived from improving feed/effluent exchanger design is significant. The benefits of revamping existing units include energy saving, reduction in catalyst regeneration cost and improvements in plant throughput in terms of increased production rate and extended operating period.

www.eptq.com REVAMPS 2009 3

Catalyticreactor

Figure 1 Heat recovery around reactor

calgalvin.indd 1 12/8/09 10:44:53

MW associated with condensation is divided between the feed/effluent exchanger (3 MW) and the cooler (3.2 MW).

Problems in Texas Tower designIn the example outlined, we see that in terms of heat transfer mechanisms the feed/effluent exchanger can be divided into three sections. At the base of the tower (cumulative heat load 0–3 MW), condensation (in the presence of non-condensables) is occurring on the shell side with evaporation on the tube side. In the mid-section of the tower (cumulative load 3–11.7 MW), desuperheating is occurring on the shell side with evaporation on the tube side. At the top of the tower, desuperheating is occurring on the shell side with superheating on the tube side.

Heat transfer to and from gas streams dominates the problem. However, the mix of heat transfer mechanisms coupled with the desire to achieve full recovery in a single exchanger leads to a number of problems. The first major problem is obtaining a uniform flow distribution on the tube side of the exchanger. The feed stream entering the unit is a two-phase mixture. The tube bundle presents numerous flow paths. In these circumstances, the two phases will distribute so that the overall pressure drop across the unit is minimised.

Identifying energy benefitsThe thermodynamics of heat recovery in a feed/effluent exchanger can be determined by superimposing curves showing the heat demand (reactor feed) and the heat release (reactor effluent) for the case described above as functions of temperature (see Figure 2).

The curves have been superimposed at a minimum temperature approach of 50°C. The distance the demand curve (the lower curve) extends beyond the release curve at the hot end of the scale shows the demand placed upon the fired heater. In this example, the reactor feed is heated to a temperature of 300°C, the feed has to enter the reactor at a temperature of 390°C and the load on the fired heater is 6.7 MW.

The total demand of the feed stream is 27.5 MW and consists of 11.7 MW associated with vapourisation of the feed stream and 15.8 MW for superheating the stream to 390°C. Given that 6.7 MW is provided by the fired heater, heat recovery provides 20.8 MW.

The heat released by the reactor product as it is cooled to 130°C is 24.0 MW. This figure comprises 17.8 MW for desuperheating the gas and 6.2 MW associated with condensation of product. The amount of heat being absorbed in the cooler positioned after the feed/effluent exchanger is 3.2 MW. Hence, the 6.2

4 REVAMPS 2009 www.eptq.com

Total demand 27.5 MWHeat recovery 20.8 MW

Total supply 24 MWHeat rejection 3.2 MW

Figure 2 Heat recovery around reactor

Our knockout drum needs a mist eliminator!

Can we add one without overhauling the vessel?

Read more on this topic atwww.amistco.com

IT’S A MORE COMMON problem than you might think. A vertical knockout drum removes free liquid from a certain gas stream. But at the time the plant was built, a mist eliminator was not considered necessary. Now mist is carrying over and causing trouble downstream. There is no manway, so adding a conventional mist eliminator would require cutting the vessel open.

mist eliminator inside, securing it

Even when there is a manway,

and old vessels, it drastically cuts installation cost and downtime.

[email protected]


calgalvin.indd 2 24/8/09 09:59:25

http://www.eptq.com/jump.aspx?a=Amistco&p=revamps09

The easiest way of ensuring a reasonable phase distribution is to set the tube count so that the pressure gradient associated with a well-mixed two-phase stream is less than that associated with the hydrostatic head of the liquid phase alone. Under these circumstances, the liquid can only be transported up the tubes as part of a two-phase mixture. However, this can result in another danger: the return of liquid down the tubes at the bundle periphery. So, as a safety measure, phase distributors, such as perforated plates, are often installed in the exchanger headers to ensure gas is present below these tubes.

The next problem is that of maximum tube length. This is typically around 20m. The heat transfer duties are large, typically 20 MW. This drives the design towards large tube counts and large shell diameters.

The limitation on tube length and the need to ensure good phase distribution lead to exchangers with low mass fluxes, relatively low Reynolds numbers (on both sides of the unit) and, as a result, low heat transfer coefficients. However, this is accompanied by a pressure drop that is usually much lower than the permitted value. This results in a significant opportunity to use heat transfer enhancement.

Other problems reported with Texas Towers include the presence of film boiling. This phenomenon is related to the quantity of vapour generated at the tube wall and is another situation where heat transfer enhancement is beneficial. Enhancement results in a lowering of the wall temperature and a reduction in the amount of vapour generated at the exchanger wall, while the amount of vapour generated at the interface between the liquid and vapour streams increases. It should be noted that film boiling often gives rise to fouling.

Finally, it is possible for two-phase flow within the tubes to move from an annular flow regime to one in which the liquid is entrained in a mist. This results in a significant reduction in the convective heat transfer coefficient. Wire mesh inserts

act as demisters and can be expected to improve wetting of the tube wall at high vapour mass quality.

Inserts to enhance heat transferWire matrix inserts are a useful means of enhancing heat transfer in the laminar and transitional flow regions. What is not widely appreciated is that they can also be used in gas flows at quite high Reynolds numbers. Performance tests on hiTRAN (Figure 3) extending well into the turbulent flow region have been carried out at the UK’s National Engineering

Laboratory and at Cal Gavin’s test facility.

In Figure 4, we compare the performance of a low-density wire matrix insert with that of an empty tube. The use of the inserts more than doubles the heat transfer coefficient. However, this is accompanied by a higher flow resistance and, for a fixed Reynolds number, a higher pressure drop. In a revamp, where the Reynolds number is often fixed, the choice of insert is based on pressure drop constraints.

Scope for enhancing heat transferGiven that with plain tubes feed/effluent exchanger designs are driven to gas stream Reynolds numbers that are relatively low (below 100 000) and the heat transfer coefficients achieved in such designs are relatively low, there is significant scope for the use of heat transfer enhancement in both the design of new units and the revamping of existing units.

The use of tube inserts to enhance the heat transfer on the tube side is acceptable for both the evaporation and superheating regions, and it is possible to use different insert geometries in the two regions.

In the case of the shell side, one means of enhancing heat transfer, where gas-to-gas heat transfer is occurring, is the use of low fin tubes. With horizontal units, low fin tubes can be used throughout. However, with vertical units, fins are not acceptable where condensation is occurring. Consequently tubes with low fins in the top part and a plain surface lower down may be used.

Installation of wire matrix inserts is a relatively inexpensive retrofit. The improvement they offer will be limited by the shell-side performance of the exchanger. The use of low-finned tubes on the shell side and inserts on the tube side may also be justified economically. However, this requires a bundle replacement.

For new design and retrofit, the engineer should be aware that, given the variation in heat transfer mechanism along the heat exchanger, the use of a single unit to achieve the required duty may not prove to be the best option. Consideration should be given to the use of one unit to


Figure 3 Tube wall dye stream disruption with a hiTRAN wire matrix insert. Area A shows an uninterrupted laminar flow condition, while area B shows laminar flow disrupted by an insert

The use of tube inserts to enhance the heat transfer on the tube side is acceptable for both the evaporation and superheating regions

calgalvin.indd 3 12/8/09 10:45:25

When You Need Trays....NOW!Why wait 10, 15 or 20 weeks for your delivery! Ask AMISTCO Separation

Products about our Fast Track delivery for trays and a complete range of

tower internals. With complete in-house engineering and fabrication, we

can use your existing drawings or modify them to improve your process.

Valve (numerous options)Sieve or perforated

Bubble cap traysCartridge trays

®

MIST ELIMINATORS

DISTRIBUTORS & SUPPORTS COALESCERS STRUCTURED PACKING

RANDOM PACKING PROFILE WIRE SCREEN

Trays_PTQ.indd 1 7/27/09 6:26 PM


amistco.indd 1 12/8/09 10:18:30

http://www.eptq.com/jump.aspx?a=Amistco&p=revamps09

undertake the phase changes and a second to achieve gas-to-gas heat transfer. In a retrofit, the first unit may be a compact exchanger, while gas-to-gas heat exchange may be achieved in an existing shell-and-tube exchanger fitted with tube inserts.

Wire matrix insertsA recent project in which hiTRAN inserts were used involves the Ruhr Oel plant in Gelsenkirchen, Germany. The plant had been operating successfully for 30 years, but had reached maximum capacity. The operator wanted to increase throughput by 17.5%.

The existing feed effluent exchanger matched feed and effluent associated with an Isomar Process catalytic reactor for paraxylene production. The exchanger comprised 1734 tubes of 20mm outside diameter and 12m length in a 1360mm diameter shell. This gave a heat transfer surface area of 1256m2.

At the established throughput, the feed stream exited the exchanger at a temperature of 327°C. Simulation of the performance of the Texas Tower at increased throughput indicated that the temperature at which the feed left the unit would fall to 310°C. This meant that at end-of-run conditions the load on the fired heater would be above its capacity. Consequently, the operator had either to improve heat recovery or renew the catalyst more frequently. The last option would defeat the object of the revamp, because it would result in a significant increase in operating costs associated with catalyst renewal, which would in turn lead to reduced plant capacity because of increased outage time.

The operator was left with three options: replace the fired heater and burn more fuel; replace the Texas Tower with a new, larger unit; or improve the performance of the Texas Tower. Replacement of the fired heater would involve significant capital cost, increased energy consumption and increased carbon emissions. Replace-ment of the existing tower with a larger unit would incur the capital expenditure associated with the purchase of a new, large unit and could also attract significant civil

encountered in the exchanger• The feed entered at a mass quality of 0.333. The overall mass flux for the unit was low (at 111.4 kg/m2s). Consequently, the liquid film Reynolds number was low at just 2540 and, despite having a two-phase flow, the average two-phase convective heat transfer coefficient was only around 1160 W/m2K. After adjustment for vapour phase resistance, the tube-side heat transfer coefficient was predicted to be 740 W/m2K. The average shell-side heat transfer coefficient at this location was predicted to be 820 W/m2K. The overall heat transfer coefficient across this region was 277 W/m2K. Hence, the tube-side thermal resistance was 47% of the overall (fouled) value• The average heat transfer coefficient towards the tube outlet (where the gases were being superheated) was

Figure 5 Incremental frictional pressure drop and vapour fraction as a function of

tube length

engineering costs associated with the strengthening of existing support structures to accommodate the additional weight of the tower. The operator opted to examine the benefits of using heat transfer enhancement.

Behaviour with increased throughputSimulation of the existing unit under conditions of increased throughput yielded the following observations:• The pressure gradient for uniform distribution of the two-phase flow was significantly lower than the static head gradient for the liquid, so a uniform two-phase distribution can be expected• Application of an overall fouling resistance of 0.000585 m2K/w provided reasonable agreement between predicted and observed performance. This resistance was around 16% of the total resistance

Figure 4 Heat transfer enhancement at high Reynolds numbers


calgalvin.indd 4 13/8/09 15:45:17


predicted to be 730 W/m2K. The average shell-side heat transfer coefficient in this region was predicted to be 918 W/m2K. The overall heat transfer coefficient was 273 W/m2K. Again, the tube-side thermal resistance was 47% of the overall (fouled) value• The gas velocity at the tube outlet was low (at around 10 m/s) and the Reynolds number was about 83 000• The pressure drop on the tube side of the unit was less than 15 kPa, whereas the allowable pressure drop was 110 kPa• The HTRI program predicted that film boiling would be present in parts

of the tube, with mist flow present in others.

Revamp using tube insertsGiven the relatively low Reynolds numbers involved and the availability of pressure drop, it was clear that tube inserts would provide a means of debottlenecking the unit. It was assumed that the fouling levels encountered in the revamped unit would be similar to those estimated for the current unit. Tube inserts often lead to a reduction in fouling levels due to a large increase in wall shear. Elimination of film boiling, which happened in this case, can also be expected to lead to a reduction in fouling levels. These potential benefits

have been ignored in the exchanger analysis, so the results were deemed to be conservative.

Simulation of the tube-side behaviour of the unit showed, for about half of the tube length, the presence of a boiling two-phase mixture. After this region, the fluid is superheated before entering the fired heater (see Figure 1). Fitting a low-density insert (Figure 5) into the flow boiling region of the tube resulted in a modest increase in the tube-side heat transfer coefficient from 740–825 W/m2K. This is a result of heat transfer being dominated by the

resistance presented by the vapour phase.

Placement of a slightly higher density insert into the upper section of the tube (where single-phase heat transfer is occurring) resulted in the overall heat transfer coefficient being increased from 273–405 W/m2K, for 48% increase. The pressure drop encountered on the tube side of the unit rose from around 15–85 kPa, but was still well within the allowable value of 110 kPa.

Overall heat recovery was, indeed, increased, such that the duty required of the fired heater was within its capability. The revamped unit has been back on stream for over 18 months and is performing well.

After evaluating the incremental pressure drop distribution over the length of the exchanger (see Figure 5), it was decided to apply a joined insert with two different packing densities over the whole tube length: a low-density insert for the two-phase region and a slightly higher density insert for the superheated region. The revamp was quickly engineered and the installation was completed within the normal shutdown period.

Summary of benefits As a result of the installation of wire inserts, heat transfer in the flow boiling region improved and, according to the HTRI program used, film boiling was eliminated. This, in turn, means that fouling associated with film boiling would also be eliminated.

Heat transfer in the superheating region increased by 48%, while the reduction in fired heater duty delivered energy savings of about 0.8 MW, which represents a reduced CO2 output of about 1700 t/y. Increased heat recovery led to energy cost savings, which paid for the revamp within 18 months. Large-scale investment in new equipment was avoided.

Isomar is a mark of UOP LLC.

Peter Ellerby is Engineering Manager at Cal Gavin Ltd, UK. He is experienced in the design of heat exchangers and a chemical engineer specialising in heat transfer enhancement. Email: [email protected] Drögemüller is Research & Development Manager at Cal Gavin Ltd, UK. His research focuses on the use of turbulence promoters in film flows and the development of hiTRAN software. He has a PhD in chem engineering. Email: [email protected] Edgar Vazquez-Ramirez holds a master’s degree in mechanical engineering. He is working towards his PhD in boiling heat transfer and two-phase flow at the University of Guanajuato, Mexico.Graham Polley is a professor at the University of Guanajuato, Mexico, where he helps supervise the development of heat exchanger design methods. He is a past president of the UK’s Heat Transfer Society.Email: [email protected] Erlenkämper is Plant Manager of the Aromatics Complex at BP Gelsenkirchen Refinery, Germany. Email: [email protected]

Installation of the joined hiTRAN Insert into the feed effluent exchanger

.During the planning and implementation of the expansion project he was Process Engineer for the plant.

calgalvin.indd 5 12/8/09 10:45:51

TIGER TOWER SERVICES

Hunting for a new and bettertower service?

Call the Tigers.Our experienced team can jump on your tower service

project and execute it reliably, safely and cost effectively.

We provide all the specialized skills you need for tray/pack-

ing installation, tower maintenance and plant turnaround

projects. We are not allied with any one manufacturer and

can work with all types of separations products.

We can completely retrofi t your vessel from its external

pipe fl anges to even the most complex internals. No detail

is overlooked. We remove and install trays and/or packing,

provide tray support modifi cations, clad liner repair/installa-

tion, nozzles, installation and tower resections.

We can help you creep your production by providing

engineering services for pressure vessels and vessel modi-

fi cations. Our design/analysis also includes calculations of

induced loads and we provide AutoCad® drawings of all

vessel/vessel modifi cations along with integrity analysis

and re-rating.

Our keen eye for

quality includes

inspections at all

applicable stages by

CCWI and API 510

qualifi ed personnel.

When you need tower services of any kind; HUNT NO MORE.

www.tigertowerservices.com

A RepconStrickland Company

TigerTower FullPage Ad_8-4-09.indd 1 8/4/09 12:49 PM


tiger tower.indd 1 12/8/09 10:21:04

http://www.eptq.com/jump.aspx?a=Tiger&p=revamps09

Now, styrene producers don’t have to choose.

CRI’s new AstraCat catalysts offers new levels of activity and selectivity at traditional

Steam to Hydrocarbon Ratio. When you want to both maximize productivity and

lower your raw materials costs, AstraCat delivers unrivaled performance.

www.cricatalyst.com© 2009 CRI Catalyst Company. KataLeuna is a trademark of CRI Catalyst Company. (CCC905)

AstraCat Catalysts

CCC905_PTQ_ad.indd 1 8/4/09 6:16:14 PM


cri.indd 1 12/8/09 10:19:12

http://www.eptq.com/jump.aspx?a=CRI&p=revamps09

Stepwise simulation of vacuum transfer line hydraulics

W hen designing a vacuum transfer line, a robust hydraulic model that

predicts velocity and a corresponding pressure drop is crucial. A stepwise approach to hydraulic modelling of vacuum transfer lines increases accuracy and enhances the under-standing of two-phase fluid behaviour. Vacuum gas oil yield, reliability and operability depend on correct design of the vacuum transfer line.

With the depletion of global conventional oil, refinery feedstocks are becoming heavier and often rely on unconventional heavy oil. Canadian oil sands-derived bitumen — a heavy, unconventional oil — is being increasingly processed in Canadian upgraders to produce synthetic crude oil as part of the crude slate in US refineries. Processing unconventional heavy oil like bitumen is challenging. Unconventional heavy oil is usually unstable and prone to coking at high temperature. To maintain reasonable run lengths, a temperature limit is applied to the heater outlet, which indirectly puts a limit on the heater tube’s inside film temperature.

Design objectivesA typical objective in the design of a vacuum unit is to maximise the yield of vacuum gas oil to improve a refinery’s profitability. The vacuum overhead system, column flash zone, vacuum transfer line and the charge heater have to be optimised as a single system to ensure that design objectives are met during unit operation.

Based on the steam and cracked gas loads, the vacuum overhead system is configured so that a low absolute

A stepwise hydraulic calculation determines the pressure profile of a vacuum transfer line by linking the hydraulic model to process simulation results

HArry HA Fluor Canada MAttHeW reisDorf Fluor EnterprisesAbDullA HArji Fluor Canada

pressure is ensured in the flash zone to allow the unit to reach target oil vapourisation at an acceptable heater outlet temperature (HOT). A pressure drop in the vacuum transfer line sets the heater outlet pressure (HOP), which in turn determines the HOT by vapour-liquid equilibrium.

The HOT is limited to an acceptable value to avoid coke formation inside the heater coils. A high pressure drop in the vacuum transfer line increases the temperature difference from the heater outlet to the flash zone. Given a

fixed HOT, an increased pressure drop in the vacuum transfer line decreases the vacuum gas oil lift in the flash zone, resulting in a lower yield of vacuum gas oil. Therefore, the vacuum transfer line’s hydraulics plays a crucial role in achieving the desired product yields and operational reliability.

A study1 showed that one extra kPa added to the total pressure drop of a transfer line reduces the gas oil yield by about 0.2 vol%. For a refinery with a 100 000 bpd throughput, each kPa of pressure drop in the vacuum transfer line implies a significant loss of

revenue. From a process design point of view, the pressure drop in the vacuum transfer line should be as small as possible to maximise the yield of vacuum gas oil. This usually leads to a large transfer line and increased heater passes, resulting in significantly increased capital costs. Therefore, it is essential to select the most cost-effective design, which meets both process design and mechanical requirements.

The vacuum transfer line is a large, elevated line that routes the vacuum unit feed from the charge heater outlet to the vacuum tower flash zone. Depending on capacity, the line’s diameter can range from 48–84 inches inside diameter and its length is typically 40–70 ft. A typical piping layout for a transfer line includes either individual heater pass outlet piping discharges into the main line routed to the vacuum tower, or half the heater passes discharge into a manifold and the two manifolds discharge into the main line. The piping design group should be consulted to establish the preliminary transfer line routing, including approximate lengths and allowance for thermal expansion. Since transfer lines have a low allowable pressure drop, pressure loses due to fittings should be minimised.

The number of parallel heater tube passes is determined by the required cross-sectional area at the heater outlet to accommodate the large volume of two-phase flow. At the heater outlet, there are typically four to eight separate heater tube passes from one or more cells. While cost-effective heater design favours using fewer tube passes, the need to stay below


A typical objective in the design of a vacuum unit is to maximise the yield of vacuum gas oil to improve a refinery’s profitability

fluor.indd 1 12/8/09 11:12:21

two-phase flow hydraulics.2 Reviews and model evaluations have been conducted extensively by researchers and developers of hydraulic calculation tools. However, no universal model has been widely accepted. Depending on the criteria and the system studied, different users favour one model over others. This article will focus on how two-phase flow hydraulics can be simulated, provided the model is chosen.

Following a comparison of numerous model performances, the Dukler-Taitel model2 with liquid hold-up from the HTFS method3 was selected to calculate two-phase flow hydraulics with the aid of Fluor’s proprietary hydraulic software. The Dukler-Taitel model was derived from dynamic similarity analysis with experimental data, which is dependent

on liquid hold-up but independent of flow regime. By assuming constant slip for two-phase flow, the pressure drop through a conduit can be calculated with adequate accuracy when employing a better correlation of liquid hold-up. For details of the method, refer to the original paper.2 The liquid hold-up, defined in the HTFS 1992 design report,3 was developed from HTFS graphical correlations by optimising an

analytical function against the original graphical curves. The correlation was tested with the HTFS data bank and was claimed to give better results than the previous HTFS correlations.3

For compressible fluids, the volumetric flow of gas changes through a pipe as a result of static pressure variations along the line. The acceleration pressure drop is associated with the expansion of the gas phase as pressure is reduced. It becomes significant in two-phase systems with high mass velocity and low pressures, both of which prevail in a vacuum transfer line. Therefore, the acceleration effect cannot be ignored, and this effect is also calculated in Fluor’s proprietary software. Acceleration effects are most pronounced in tees and expanders, as the static pressure changes significantly across these segments.

the critical velocity necessitates an adequate number of tube passes.

A limiting factor in minimising the vacuum transfer line’s size is the bulk HOT. For a state-of-the-art and deep- cut vacuum unit, the maximum recommended oil temperature in a heater is usually 365–415°C to avoid excessive cracking and coking within the heater coils. To reduce cracking and coking, the HOT must be set low enough to ensure the film temperature does not exceed the maximum recommended oil temperature.

The pressure in the vacuum transfer line keeps decreasing from the HOP to the pressure of the column flash zone, which is normally set at 20–30 mmHg absolute for a deep cut. A typical pressure drop in a vacuum transfer line is about 100–150 mmHg to achieve a deep cut point. Considering the low pressure at the flash zone, the pressure drop of the vacuum transfer line is quite significant. Corresponding to this pressure change, the temperature also changes isenthalpically from the HOT to the flash zone temperature, as the feed goes through adiabatic flashes in the transfer line. Depending on the total pressure drop, a temperature difference of 10–20°C can be expected between the heater outlet and the flash zone. As a result, the vapourisation, density of fluid, volumetric flow and the transport properties all change simultaneously along the transfer line. Therefore, a stepwise, equilibrium simulation is necessary to reflect the continuous changes in a vacuum transfer line.

An important consideration in transfer line design is the two-phase critical velocity, which raises concerns about vibration in the line, especially acoustic-induced vibrations. Field measurements of the vacuum transfer line and calculations using theoretical hydraulic models confirm the existence of critical velocity and its influence on the pressure profile inside a vacuum transfer line.

This phenomenon is difficult to predict for the two-phase flow system. Two-phase critical velocity is much lower than the sonic velocity of the

gas phase alone. Therefore, many transfer lines, designed to run under sonic velocity, actually operate at critical velocity, especially near the heater outlet and at the column entrance. The potential impacts of critical velocity are not trivial. The vibration of shock waves could result in failure of the vacuum transfer line. Critical velocity can also lead to excessive entrainment of liquid droplets to the wash zone, even with a well-designed column feed device.1 Heater designers attempt to limit the velocity in the heater coil to reduce vibration problems, and often target a velocity limit of 80% of critical velocity within the heater coil itself. With modern deep-cut columns, avoiding critical velocity throughout the vacuum transfer line often becomes impractical. To the designer, the task

then becomes one of balancing the risks of vibration and entrainment with the reward of a high yield and low cost.

Hydraulic models for transfer line calculationsThe fluid velocity within the vacuum transfer line1 is normally high — either close to or at critical velocity. It is generally believed that a homogeneous-phase dispersed flow regime is present at design and normal operations, while a separated-phase annular flow regime is observed during turndown cases.

For a better understanding of the flow regime and hydraulics within a vacuum transfer line, a two-phase flow model should be applied. Over the years, many models have been developed to account for different fluid systems and flow regimes for


Total = 4.3 psia Total = 4.3 psia

Figure 1 Example of an expander in vacuum transfer line

fluor.indd 2 12/8/09 11:12:31

Figure 1 illustrates an example of the expander in a vacuum transfer line, where the fluid is travelling near critical velocity.

According to Bernoulli’s law, both a fluid’s pressure and velocity contribute to the total energy contained in the moving fluid. At high velocity at the expander inlet, the fluid’s velocity head is a significant component in this sum. As the fluid moves across the gradual expander, the amount of energy lost to friction is relatively low. The velocity is significantly reduced and much of the velocity head is converted to static pressure. The net result is that the static pressure rises across the expander. This effect has been documented in two-phase flow across a sharp expansion,4 and it would be expected to be even more pronounced across the more gradual expansions in pipe fittings.

For a simple check of the vacuum transfer line’s hydraulic model, a designer should review each point where the velocity drops suddenly —

commonly expanders and tees where two process streams combine. As the fluid passes through the expander, the static pressure should rise at these points. If the static pressure does not

rise at these points, the calculation warrants further scrutiny to confirm if the fluid has hit the critical velocity.

As the fluid moves through the transfer line and the pressure drops, some portion of the liquid vapourises, adding to the gas flow. There is a debate on whether or not the gas-

liquid phase is in equilibrium within the transfer line. Some process designers claim that the vacuum transfer line should be modelled by assuming non-equilibrium between vapour and liquid because of the high transfer line velocity,5 while other designers prefer to assume significant or complete equilibrium throughout the transfer line.

Non-equilibrium models give better predictions of wash rate to provide sufficient wetting of the wash bed.5 However, for the purposes of transfer line hydraulics, the most conservative assumption is to consider the fluid in equilibrium throughout the transfer line, as this produces the highest amount of vapour, the highest velocity and, thus, the highest pressure drop. Therefore, complete gas-liquid equilibrium is assumed within the transfer line in this instance.

Stepwise calculation of transfer line hydraulicsIdeally, an integrated numerical equation should be used to reflect the


For a simple check of the vacuum transfer line’s hydraulic model, a designer should review each point where the velocity drops suddenly

THE COMPLETE FURNACE TECHNOLOGY COMPANY

OnQuest leads the industry in designing, supplying, and installing high-performance furnaces, heaters, and related combustion technologies for clients in the oil refining and petrochemical industries.

We are also a world leader in direct-fired process heater technology and burner management systems, and have particular expertise in lump-sum, turnkey projects, retrofits, and revamps.

For more information, visit onquest-inc.com or call Randy Kessler at (909) 451-0502.

Process. Performance.

SOLUTIONS.

• High-Performance Solutions – Our designs prioritizeproject efficiency, for shorter field schedules, reduced man-hours, and lower costs.

• World-Class Expertise – Our engineers and processexperts have decades of experience in complex design and installation projects.

• Global Capability – With offices in California, Texas,and Calgary, Alberta, and representatives in South America, Europe and Asia, we serve clients worldwide.

180 East Arrow Hwy | San Dimas CA 91773Tel. (909) 451-0500 | Fax (909) 451-0499


fluor.indd 3 12/8/09 11:12:45

http://www.eptq.com/jump.aspx?a=Onquest&p=revamps09

differential changes in properties along the vacuum transfer line. However, such a solution is difficult to find for such complex systems. As a compromise, the vacuum transfer line is split into several segments so that the variation in transport properties is insignificant for each segment. Consequently, properties are approxi-mately constant across the segment, and the hydraulic models described above can be applied to each segment. The total pressure drop is simply the sum of all segments.

This method is similar to the use of a numerical integration method, such as the Simpson method, to solve the differential equations. Applying the Dukler-Taitel model to each segment and summarising total hydraulic changes result in an accurate simulation of vacuum transfer line hydraulics. As a rule of thumb, the pressure change of each individual segment should be no more than 10%

of the inlet static pressure of that segment. Separate segments are suggested for all pipe fittings (elbows, expanders, inlet device, and so on). Most importantly, expanders should be placed in a segment of their own, as the static pressure is expected to rise significantly through the expander. When the two-phase fluid flows into the flash zone through an inlet device, all of the velocity energy of the fluid is lost to friction, but the static pressure in the pipe at the exit is essentially the same as that in the column flash zone. The exit loss is due to the dissipation of the discharged jet. There is no pressure drop at the exit except for an insignificant pressure loss due to the inlet device (3.0-3.8 mmHg).1 There have been cases when extra pressure loss was given to the inlet device in calculations of vacuum

decreases and its velocity increases. If the pressure drop through the

pipe is sufficiently large, the gas velocity exiting the pipe reaches sonic velocity. If the pipe outlet pressure is further decreased or the pipe inlet pressure is further increased, the excess pressure drop occurs beyond the pipe exit. This pressure drop is dissipated in the shock waves and turbulence of the exiting gas. When the ideal gas reaches its sonic velocity, it has reached the maximum mass flow rate that the gas can achieve.

For non-ideal gases and two-phase systems, the maximum mass flow through a piece of pipe is the critical flow. An intuitive way of estimating critical velocity for real gases is to use the method developed for an ideal gas, but with non-ideal gas properties and it usually works well. However, two-phase gas liquid mixtures reach critical flow at a velocity much less than sonic velocity. This phenomenon is better depicted by the separated-phase model. Hewitt and Semeria proposed a separated-phase model to derive the critical velocity of two-phase flow (VC,M) from the critical velocity of gas:6

VC,M

= VC,G [x + (1 – x)C]

: x + (1 – x)D2 C2 (1)

where x is the vapour weight fraction, C is the ratio of gas density to liquid density, D is the ratio of critical velocity in gas to critical velocity in liquid and VC,G is the critical velocity of gas.

The critical velocity of gas can be expressed as (Perry’s Handbook):

(2)

where ρ is the density of gas; p and v are the pressure and specific volume of gas.

Assuming ideal gas behaviour, Equation 2 can be simplified as:

(3)

where k = CP/CV, the ratio of specific heats and Mw = mol wt of gas.


transfer line hydraulics. This extra pressure drop changes the pressure profile and thus the transfer line’s design, which will be addressed in detail in a case study.

When the vacuum transfer line is associated with vacuum tower design, normally the temperature at the heater outlet and the pressure at the column flash zone are predetermined. A backward calculation is required for transfer line hydraulics, step-by-step, from flash zone to heater outlet to heater outlet to meet the constraints. To catch fluid property changes as the calculation moves from one segment to another, a process simulation is done by either Pro-II or HYSYS software to flash the stream adiabatically over the expected pressure range of the vacuum transfer line. The changes in properties are then tabulated over the full pressure range of the vacuum transfer line. When the hydraulic calculation goes

to the next segment, the properties are updated from the simulated property table according to the calculated current inlet static pressure of that segment. The calculation is iterated until the changes of properties for each segment are insignificant compared to the previous iteration; the calculation is then converged.

Another important issue with respect to the vacuum transfer line is the pressure discontinuity that occurs within it when the internal velocity reaches the critical velocity of the fluid. For an ideal gas system, the maximum velocity that can be achieved is limited by the maximum velocity of a pressure wave travelling in the pipe, which is equivalent to the sonic velocity of the gas. As the gas flows through the vacuum transfer line, its pressure

AB C D E

G H

K L

M

N

O P Q R S

F

I

J

Figure 2 Vacuum transfer line layout

VC,G

= 1 e upo ρ uv

s

VC,G

= kRT

Mw

fluor.indd 4 12/8/09 11:12:58

In this study, we compared the previous correlations with the homogeneous model developed by Buthod7 and the theoretical approach derived by Kohoutek et al.8 We found that the calculated critical velocities from these three methods are comparable and consistent. Therefore, the method of Hewitt and Semeria is proposed here because of its simplicity.

Stepwise method vs conventional methodCritical flow is a concern in the design or rating of vacuum unit fired heaters, transfer lines and relief header systems. Incorporated within Fluor’s proprietary hydraulic software, an automation program has been developed that updates the transport properties in hydraulic models from the stream properties simulated by Pro-II or HYSYS. With the aid of these programs, a case was studied for a typical vacuum transfer line used in modern vacuum distillation units. A sketch of the vacuum transfer line’s layout, with a branch line connecting the mega-line to one of the heater passes at the heater outlet, is shown in Figure 2. The case presented, not specific to any plant, demonstrates the methodology and procedures used by designers to address key technical issues.

Hydraulic calculations are conducted from the column flash zone, segment by segment, marked by letters all the way to the heater outlet. The simulated hydraulic profile of the vacuum transfer line is illustrated in Figure 3, with location points referring to the layout presented in Figure 2. The calculated results match the specified process constraints: the operating pressure in the flash zone and the heater outlet temperature. Transport properties and the vapourisation of fluids within the vacuum transfer line are simulated by Pro-II using the Improved-Grayson-Streed (IGS) thermodynamic package.

The study of Laird et al1 showed that the thermodynamic packages have substantial effects on predictions of vapourisation at the flash zone condition of a vacuum tower. The IGS

package gives a vapourisation rate between those given by the Grayson-Streed (GS) and BK10 methods and was chosen to simulate the system of interest.

For a vacuum transfer line with total pressure drops of 100–150 mmHg, the simulated vapourisation changes can be as high as 15 wt% along the vacuum transfer line. Using the automated stepwise approach of this work, the property changes are captured and updated for each segment based on the online calculated static pressure of that segment. The calculated pressure drop

of each segment is less than 10% of the inlet static pressure of that segment, which justifies the assumption that properties are constant throughout each individual segment. For comparison, the hydraulic profile calculated by the conventional method (used by many transfer line and heater designers) is also shown in Figure 3. With the conventional calculation, an extra pressure drop (45 mmHg) is assigned to the inlet device as an allowance. Contrasted to the large pressure drop of the inlet device taken by the conventional method, a small and reasonable pressure drop (3.8


A

D

Stepwise method

Conventionalmethod

G

K

O

S

Figure 3 Hydraulic profile of a simulated vacuum transfer line

AD

B

Critical velocity

Calculated velocity

K

G

O

S

P

C

EH

L

Figure 4 Calculated stream velocity vs critical velocity through a vacuum transfer line

fluor.indd 5 12/8/09 11:13:14

mmHg) is used in this work, which is consistent with the values reported by Laird et al.1

As Figure 3 indicates, the simulated total pressure drop of this work is consistent with the results of the conventional method. However, the hydraulic profile of each segment differs substantially between the two methods. Figure 4 shows the calculated velocities of the fluid mixture compared to the critical velocities along the vacuum transfer line using the proposed stepwise method. The current work clearly indicates pressure discontinuities at five locations where the pressure profile of the conventional method showed no sign of calculated velocity ever reaching critical velocity. How can two calculations come to such different conclusions? The differences in this example are due to several factors, which highlight some of the key challenges of this calculation:• The inlet device typically has a very low pressure drop. The current work uses 3.8 mmHg in its calculation, which is a reasonable estimate for a modern column inlet device.1 In the calculation using the conventional method, a much higher pressure drop (45 mmHg) is assigned to the inlet device, which leads to a different pressure profile at the inlet to the column. Consequently, the vacuum transfer line is undersized, resulting in a potential choke flow at the column inlet in operation and significant entrainment inside the column flash zone. In the design of a vacuum transfer line, the pressure drop of the inlet device should be confirmed by the vendor and no extra pressure drop should be assigned• The acceleration effect across an expander is not taken into account in the conventional calculation. Figure 3 shows a monotonic decrease in pressure throughout the transfer line for the conventional calculation, while the stepwise calculation correctly shows a pressure rise across the expanders• The pressure drop calculated across pipe and fittings differs. The conventional calculation shows a near-constant velocity across each segment, as the calculated pressure drops

slightly at each segment except for the inlet device. The stepwise calculation method reveals the actual pressure changes of each segment and demonstrates the correct response of fluid velocity. The effect of pressure change on velocity is clearly shown at location points C, E, H and L in Figure 4 as the fluid passes through the expanders. As the pressure rises across an expander, a corresponding decrease in velocity is predicted using the stepwise modelling approach.

Theoretically, a vacuum transfer line should be designed to remain under critical velocity. However, as a result of uncertainty in calculating the critical velocity and the complexity of two-phase flow hydraulics, it is not unusual to see a vacuum transfer line running at critical velocity — typically at the inlet to the column and near the heater outlet. A deployment of the

stepwise simulation method is crucial for providing an accurate hydraulic profile to support design, risk analysis and review.

ConclusionA stepwise hydraulic calculation has been presented that determines the pressure profile of a vacuum transfer line by linking the hydraulic model to process simulation results. The velocity profile along the transfer line is also reported in relation to the critical velocity of the fluid, which clearly identifies the choke points along the line where the critical velocity is reached. The method offers a more accurate way to specify the pressure profile and line sizing of any two-phase flow system and provides a

useful tool for engineering design of the vacuum transfer line. Correct vacuum transfer line design has been shown to enhance profitability, operability and safety of vacuum distillation process units.

References1 Laird D, Hauser R, Schnepper C,Vacuumtower design techniques for optimumperformance and reliability, NPRA AnnualMeeting, 23–25 Mar 2003, San Antonio,Texas.2 DuklerAE,TaitelY,Flowpatterntransitionsin gas-liquid systems. Measurements andmodeling, Advances in Multiphase Flow, 2,editedbyZuberN,HewittGF,Delhaye JM,McGraw-Hill,NewYork,1986.3 HTFS Two-phase pressure drop designreport,1&2,Design Report 28,July1992.4 LottesPA,Expansionlosses intwo-phaseflow,Nucl. Sci. Engr.,9,1961,26–31.5 BarlettaT,Golden SW,Deep-cut vacuumunitdesign,PTQ,Q42005,91–97.6 Hewitt G F, Semeria R, Aspects of two-phasegas-liquidflow,Heat Exchangers: Design and Theory Sourcebook, editedbyAfganandSchundler,McGraw-Hill,1974,289.7 Buthod P, How to estimate pressure inheaters,Oil and Gas Journal,1July1957.8 KohoutekJ,ZachovalJ,OdstrcilM,StehlikP,Solvingpracticalindustrialproblemsintwo-phasemulticomponentmixture flow-criticalvelocity,Heat Transfer Engineering, 22, 2001,32–40.

Harry Ha is a Process Engineer with Fluor Canada Ltd, Calgary, Alberta, Canada. He has authored more than 20 technical papers on fluid dynamics, transport properties and characterisation of petroleum for refinery processes. He has a doctorate in chemical engineering from the University of Alberta. Email: [email protected] Reisdorf is a Process Manager at Fluor Enterprises Inc, Houston, Texas. His experience is in downstream refining, including vacuum distillation units. He has a BS degree in chemical engineering from Rice University, Houston. Email: [email protected] Harji is Executive Director, Process Technology, at Fluor Canada Ltd, Calgary, Alberta, Canada. He has more than 35 years of experience relating to engineering and operations support of refining, bitumen/heavy oil upgrading, gas processing and petrochemicals facilities. He has a BSc degree in chemical engineering from Loughborough University, UK. Email: [email protected]

Table 4:

16REVAMPS 2009 www.eptq.com

A deployment of the stepwise simulation method is crucial for providing an accurate hydraulic profile to support design, risk analysis and review

fluor.indd 6 12/8/09 11:13:25

© 2009 KBRAll Rights ReservedK09062 08/09

KBR Technology specializes in developing and licensing process technologies worldwide. From refining to ammonia, from chemicals to coal gasification, from olefins to syngas, KBR Technology helps you accelerate profitability and sustain growth.

For more information, visit technology.kbr.com/PTQor email [email protected]

PTQ ad_4Q_2009_A4.indd 1 8/12/2009 7:41:13 AM


kbr.indd 1 12/8/09 15:46:09

http://www.eptq.com/jump.aspx?a=KBR&p=revamps09

3400 BissonnetSuite 130Houston, Texas 77005USA

Ph: [1] (713) 665-7046Fx: [1] (713) [email protected]

PROCESSCONSULTINGSERVICES,INC.

In troubled times fierce globalcompetition for premium crudesmeans that refinery units musthave the flexibility to handleheavy, viscous, dirty crudes thatincreasingly threaten to dominatemarkets. And flexibility mustextend to products as well ascrudes, for refinery productdemand has become more andmore subject to violent economicand political swings. Thus refin-ers must have the greatest flexi-bility in determining yields ofnaphtha, jet fuel, diesel and vacu-um gas oil products.Rather than a single point processmodel, the crude/vacuum unitdesign must provide continuousflexibility to operate reliably overlong periods of time. Simplymeeting the process guarantee 90days after start-up is very differ-ent than having a unit still operat-ing well after 5 years. Sadly fewrefiners actually achieve this—nomatter all the slick presentationsby engineers in business suits!

Why DoManyCrude/VacuumUnits PerformPoorly?In many cases it’s because theoriginal design was based moreon virtual than actual reality.There is no question: computersimulations have a key roleto play but it’s equally truethat process design needs to bebased on what works in the fieldand not on the ideals of theprocess simulator. Nor should thedesigner simply base the equip-ment selection on vendor-statedperformance. The design engi-neer needs to have actual refineryprocess engineering experience,not just expertise in office-based

modeling. Refinery hands-onexperience teaches that fouling,corrosion, asphaltene precipita-tion, crude variability, and crudethermal instability, and manyother non-ideals are the reality.Theoretical outputs of process orequipment models are not. In thisera of slick colorful PowerPoint®presentations by well-spokenengineers in Saville Row suits,it’s no wonder that units don’twork. Shouldn’t engineers wear-ing Nomex® coveralls who haveworked with operators and takenfield measurements be accordedgreater credibility?Today more than ever before thisis important. Gone are the dayswhen a refiner could rely onuninterrupted supplies of light,sweet, easy-to-process crudes.

If you want to explore these issuesin technical detail ask forTechnical Papers 267 and 268.

A4 QUESTION PTQ AD :10 01 PC PTQ 0107 ADF 4/8/09 8:32 AM Page 1


pcs.indd 1 13/8/09 15:46:27

http://www.eptq.com/jump.aspx?a=Process&p=revamps09

Revamping hydrogen and sulphur plants to meet future challenges

A round the world, increasing demand for clean fuels has led refiners to alter operations for

the production of low-sulphur gasoline and diesel. Furthermore, refinery operators are purchasing larger quantities of heavier, sourer crudes to take advantage of a discount over light, sweet crudes.

To support fuel and crude supply trends, the installed capacity of hydroprocessing technologies (ie, hydrotreating and hydrocracking) has increased steadily in recent years. FCC technology — the capacity of which has been on a similar upward trend — also plays a key role in clean fuels production and is considered a major emitter of H2S and SOX.

The increases in these three processing technologies coincide with a five-year average increase in both refinery on-purpose hydrogen production (and recovery) and sulphur processing capacity. Figure 1 indicates worldwide growth in these areas in terms of refinery production capacity since January 2005. Worldwide refinery hydrogen production capacity has grown at an average annual rate of 2.15%. Average annualised growth for worldwide sulphur processing capacity since January 2005 was pegged at 2.9%.

In the past, the hydrogen production unit and the sulphur plant had been considered supporting units for the major refining processes; however, as hydroprocessing capacity continues to expand along with the processing of heavier crude, the impact of the operations of the hydrogen unit and the sulphur plant has grown. Modification and expansion of refinery hydrogen production and

Revamp opportunities for hydrogen and sulphur plant operations in the era of clean fuels production offer significant benefits for refiners

AdRienne Blume, PAtRick chRistensen, BRett GoldhAmmeR and thomAs YeunGHydrocarbon Publishing Company

sulphur plants via revamp and retrofit projects will be necessary for meeting future process requirements while maximising the value of current plant configurations and assets. The following is intended to provide an overview of available technologies and operational goals that should be considered when revamping the hydrogen plant and/or sulphur recovery/production plant in a modern refinery.

hydrogen plant revampsThe hydrogen production facility can now be considered a major component of the refinery and, like any utility, maximising the economics of producing and consuming hydrogen in the refinery affects the overall success of the plant. Technology for the production of refinery hydrogen is currently dominated by the use of steam reforming of natural gas or other light hydrocarbons. Figure 2 shows a refinery hydrogen network.

The revamp of an existing hydrogen

plant is considered the cheapest way to add 10–30% capacity.1 A number of potential revamp projects, listed in order of increasing investment requirements, can be implemented to augment capacity: employ hydrogen management; increase reformer firing; improve PSA recovery; reduce steam-to-carbon ratio; add CO2 recovery; install a low-temperature shift; and add a pre-reformer or post-reformer.2 However, several important constraints are involved in such revamps; namely, minimum hydrogen product pressure, hydrogen purity, process cooling duty, availability of plot space, available down time for revamps, utilisation of export steam, availability of other utilities, safety, and pollutant emissions.3 In addition, the debottlenecking and expansion of existing hydrogen plants depends on limitations to the reformer, such as tube metal temperature, burner heat release, catalyst bed pressure drop, induced draft/forced draft fan capacity and pressure swing


Sulphur

Hydrogen

Figure 1 Worldwide refinery hydrogen and sulphur production capacity, 1/2005–1/2009

hydrocarbon.indd 1 12/8/09 11:42:20

and major hydrogen consumers (hydrotreating, hydroprocessing). The engineers identified the opportunity to limit spill-over hydrogen — excess gas that allows for continued operation of hydroprocessing units during a hydrogen supply disruption — that is routed to the fuel gas system during normal operations. Following the evaluation, two strategies were implemented to improve the use of spill-over stream: integrate all of the hydrogen-producing and -consuming units; and eliminate the spill-over stream by utilising an offsite storage facility and process optimisation.5

In many distribution systems for refinery hydrogen, hydrogen supply is cascaded through a number of hydroprocessing units in which higher-purity, high-pressure hydrogen-consuming units pass their purge gases to lower-purity, lower-pressure units. As demand for hydrogen increases, the optimum cost-effective recovery of gas streams with marginal hydrogen content is becoming more imperative.6,7 Generally, streams with less than 50% hydrogen are routed to fuel. Streams of greater than 50% hydrogen content, with sufficient pressures, are routed to various purification units like PSA, membrane and cryogenic systems for recovery. The cost of hydrogen recovery can be almost half the cost of production.

For successful hydrogen recovery, the heating value of the fuel gas after hydrogen removal, the impact on the burners, the location where recovered hydrogen enters the network, and the impact on the whole system must be considered.8 Depending on the pressure and purity of residual hydrogen, recovery followed by purification might be an option to supplement production from steam reformers. The primary sources for hydrogen recovery are either high-pressure purge loops or low-pressure, high-purity off-gas streams. As a general advantage, recovering hydrogen that would be routed to the refinery’s fuel gas system will inherently increase the heating value of the fuel gas. Table 2 provides a selection guide to hydrogen recovery processes.9

adsorption (PSA) capacity.4 The relative cost and incremental hydrogen gain of potential revamp activities is summarised in Table 1.

Hydrogen management and/or recoveryImplementing hydrogen management is often the first step to revamping and/or improving a refinery’s hydrogen network. Hydrogen management using pinch technologies and mathematical modelling will result in improved process efficiency, reduced energy consumption, lower operating costs, and improved integration of hydrogen-producing and -consuming units. One drawback to hydrogen management is that the available capacity gain is constrained by the limitations of the current system. Hydrogen management

services are becoming more prevalent as refiners look for economical ways to meet hydrogen demands. These offerings primarily take a phased approach to balance use of refinery hydrogen (often by hydrogen pinch techniques), to identify potential projects for improvement (new units, revamps, recovery schemes and so on) and to implement or evaluate changes.

As an example, engineers from Indian Oil Corporation discussed a novel in-house methodology to improve hydrogen management in a refinery. The process begins by defining the hydrogen network in the refinery: hydrogen production units (steam methane and/or naphtha reformers); purge gas and offgas streams as recovery sources (for instance, a catalytic reformer);


Hydrogen production plant

*Fuel gas streams contain <50% H2

Figure 2 Refinery hydrogen network

Revamp project Incremental hydrogen, % Relative costH

2 management 10–15 Low

Increase reformer firing 7–10 LowImprove PSA recovery 1–2 LowReduce steam-to-carbon ratio 0–4 LowAddition of CO

2 recovery 3–5 Low/high

Replace PSA with CO2 recovery 15–20 Low/high

Low temperature shift 2–5 MediumPre-reformer 8–10 MediumPost-reformer 20–30 High

Hydrogen production revamp options

Table 1


Install pre-reforming technologyPre-reforming technology — installed upstream of the primary reformer —offers refiners the ability to process efficiently a wide variety of refinery-related fuels. The most common feeds to the steam methane reformer (SMR) are natural gas and naphtha. However, the flexibility of the SMR will offer a unique opportunity to save operating costs by using refinery byproducts as feed.10 Market demand and prices, and seasonal considerations, will have an effect on the economics of refinery hydrogen production; most notably, the installation of the pre-reformer will enable processing of a greater quantity of heavy naphtha.

Pre-reforming technology offers several advantages as well as disadvantages and, as a result, the decision to install a pre-reformer should be evaluated on a case-by-case basis. For example, the installation of a pre-reformer incurs a more significant capital cost than some other revamp options such as hydrogen management. However, the presence of this technology will ultimately reduce the capacity requirements of the primary reformer, lowering overall operating costs. Similarly, the installation of a pre-reformer will require the purchase and frequent replacement of pre-reforming catalysts, but the presence of a pre-reformer will also help to remove many contaminants from the reformer feed that may limit the operating life of the primary reformer catalyst. Other important factors that must be considered prior to implementing pre-reformer technology are the effect on the heat balance of the hydrogen plant, and the ease of use and cost of the system.

Understand the effects of capacityexpansionOptions to increase the capacity of a hydrogen plant without involving a radical, full-scale overhaul — including increased reformer firing, reducing steam-to-carbon ratio and replacing reformer catalyst — are limited to a 15–25% increase in capacity.11 Additionally, several problems can arise from increased throughput. In the primary reformer

section, increases in throughput will result in a slight rise in tube wall temperature, as well as a pressure drop in the reformer tubes. The impact of these factors is minimal, but should be considered as a minor limitation to revamp operations.

Increased throughput will have a more severe effect on the flue gas duct/convection section of the reformer. High heat flux conditions in this section will lead to limitations in the capacity of the internal draft (ID) fan and increased tube wall temperatures, and structural problems will coil the bank supports. Increased production of flue gas is also viewed as a deterrent for increasing the capacity of hydrogen production units. If debottlenecking activities are not properly planned, the advantage gained by operating at increased throughput will ultimately be negated.

Catalyst and hardware upgrades require significant capital investment and can result in improved hydrogen production for the refiner. Refiners can overcome the constraint of the reformer’s tube metal temperature and achieve a 3–8% increase in hydrogen production by reducing the steam-to-carbon ratio, lowering the outlet temperature or increasing the inlet temperature. Another option, which can increase throughput by about 10%, is to replace the reformer tubing with a new microalloy. This new metallurgy has higher stress values, enabling the tube walls to be made thinner and higher reformer outlet temperatures to be used. The larger inner diameter of the tubes increases the flow area

and lowers the pressure drop. This type of revamp is especially suitable when tubes are reaching the end of their run life (about ten years) and need to be replaced. Refiners should also consider renewing the outlet system and upgrading the flue gas tunnels for more uniform heat distribution.

At least two problems can occur in the convection section when extra hydrogen production is required; namely, high coil temperatures and limited draft capacity due to high flue gas volumes or temperatures. The removal of tubes from the process preheat coil will reduce heat input and result in additional capacity. Another option is to modify the reformer in order to push some heat back to the radiant section so that the radiant heat and firing rate are reduced and the entire reformer can be unloaded. In one case, capacity was

boosted by 10% without raising the arch temperature.

Another debottlenecking option is to modify the PSA unit in order to improve hydrogen recovery by 1–2%.12 One such adjustment involves altering the PSA cycle, which consists of adsorption, depressurisation, dump, purge and repressurisation steps. Other options include reducing the purge gas pressure, performing an adsorbent change-out or utilising additional adsorbent vessels. Finally, existing PSA capacity can be increased through the addition of an upstream CO2 removal system. However, the net hydrogen benefit, which can reach 5%, is generally not high enough to justify this option unless CO2 is a desired byproduct.


Process Types of ROGPSA Hydrocrackerlow-pressureflash,hydrocrackervent, hydrodesulphurisationvent,catalyticreformeroffgas

Membrane FCCoffgas,hydrocrackerlow-pressureflash,hydrocrackervent, HDSvent,catalyticreformeroffgas

Cryogenicseparatororcoldbox FCCoffgas,hydrocrackerlow-pressureflashPSA-membrane Hydrocrackerhigh-pressureventPSA-cryogenic Catalyticreformeroffgas,FCCoffgas,hydrotreateroffgasMembrane-cryogenic Gasoilhydrotreateroffgas,FCCoffgas

Offgas sources for hydrogen recovery

Table 2


Cogeneration with steam reformingDepending on the operational setting of the steam reformer, different amounts of export steam can be produced for utilisation in the SMR and throughout the refinery.10 Furthermore, a significant opportunity to cogenerate power and steam via waste heat recovery is available in steam reforming. In many plant configurations, substantial benefits can be realised by installing a heat recovery steam generator and/or a turbine during a revamp project. Power recovery will not only improve the energy efficiency and environmental performance of the hydrogen plant, but it will also effectively lower the cost per unit of hydrogen being produced. The key to effectively cogenerating usable utilities from a steam reforming unit is to meet efficiently the demands of the plant and to reduce any additional losses. Each refinery has a unique utility balance, and finding the optimum ratio of hydrogen to steam to power production is important to ensure efficient and economical operations in the hydrogen plant.

Compare revamp and replacementIn some cases, rising demand for hydrogen cannot be met by simple

million investment for the new plant, the price of hydrogen was calculated again. The new plant still showed a significant improvement on a cost-per-unit basis — $1.602/kscf of hydrogen for the new plant, against $1.996/kscf of hydrogen for the older facility.

Sulphur plant revampsExisting sulphur recovery technologies are numerous, to say the least, but most solutions are linked to the modified Claus process. For this reason, revamp options are grouped into three main categories: acid gas removal, Claus unit and tail gas treatment. The synergy of these processes can be seen in Figure 3.

Acid gas removalHydrogen sulphide (H2S) and carbon dioxide (CO2) are acid gases that should be removed from refinery streams for numerous reasons.14

While the former is poisonous and combusts to produce the primary reactant in acid rain formation, the latter lowers the capacity and efficiency of the sulphur recovery unit (SRU). Commercial processes offer various features ranging from types of chemical and physical solvents —methyldiethanolamine (MDEA), diglycolamine (DGA), diisopropanol-


revamp activities, and significant capital investment in new reforming technology is necessary. It is important to evaluate the previously mentioned revamp options prior to determining the most suitable upgrade method for each plant.

In a study, engineers from CB&I Howe-Baker considered and evaluated alternative options for increasing hydrogen-producing capacity: debot-tlenecking and revamping an old plant or installing a state-of-the-art hydrogen production unit.13 The engineers attempted to evaluate and compare the overall hydrogen production cost in the old plant and the modern plant over the entire life cycle of the unit, including the significant capital investment required for a new plant. For this evaluation, a modern plant (producing 99.9% pure hydrogen) and an old-style plant (producing 95% pure hydrogen) — each with a capacity of 90MM scf/d — were compared. Both plants were based on natural gas feed. The estimated utility costs show that the efficiency improvement of the modern plant resulted in a significant annual savings ($16.4MM). This finding, however, does not include the capital investment required to construct the new plant. After allowing for a $5

One or twoadditionalcatalyticstagesre

bros

bA

Incinerator(rxns 1 and 2)

Clauscatalyticreactor(rxn 2)

rebr

osb

A

tnevlo

Sr

otare neg er

rotcae r n

oit cude

R)6 – 3 s nxr(

t ne vlo

Sr

otare neger

Figure 3 Generalised sulphur removal process


Table 3

amine (DIPA), N-methyl-2-pyrrolidinone (NMP), polyethylene glycol dimethyl ethers and sulpholane — to two-absorber or two-regenerator configurations. Often, a reduction in utilisation cost via increased amine circulation is mentioned as the key benefit in these processes.

Upgrade solvent for increased capacityMDEA and DGA can be used to replace existing amines in order to achieve an increase in sulphur unit capacity. This type of conversion also enables a higher amine concentration to be used before corrosion becomes a problem. As a result, the circulation requirement of MDEA and DGA is lower than that of solvents like monoethanolamine (MEA) or diethanolamine (DEA).15 MDEA also provides the added advantage of higher selectivity for H2S in the presence of CO2, while DGA offers high CO2 selectivity.16 Amine conversion can be performed online or during unit shutdown.17 According to Dow Chemical Company, it is important to purge the system entirely and start clean when implementing a new amine because a mixture of two amines will not provide the full benefits offered by the individual components.17 Over the years, several refiners have commented, at past NPRA Q&A sessions, on solvent change-out in the AGR unit. In three cases, a switch was made from DEA to MDEA with the goal of increased capacity. In two of these cases, low concentrations of solvent were utilised to avoid losses in the liquid contactors. Two other refiners commented on the

switch from MEA to DGA with the goals of increasing capacity and reducing conversion, respectively. In the first case, a 15% increase in plant capacity was observed. In the second, the switch to DGA resulted in significant solids accumulation in the flash drum.

Explore design modificationsTypical amine-based acid gas removal processes consist of a counter-current absorber followed by a flash stage (for high-pressure sour gases) to remove dissolved and entrained hydrocarbons from the rich amine. The rich amine is then sent to a steam stripper to regenerate the amine solvent and produce acid gas to be sent to the SRU.18 Technologies for revamping absorbers are available from several licensors.

Claus unitThe predominant sulphur recovery technology is the modified Claus process.19 Commercially, several designs of the Claus plant are available: oxygen enrichment, sub-dew point and direct oxidation. Selection between the various Claus configurations or switching from one to another via revamp may provide significant benefits to a refiner, depending on plant configuration.

Utilise oxygen enrichmentA sizeable portion of the volumetric flow to the SRU is the air feed. The high fraction of nitrogen in the air contributes to a significant amount of this gas in the reactor, leading to limitations in operational pressure

drop.20,21 Oxygen enrichment can be used to eliminate the feeding of unnecessary nitrogen to the SRU, thereby reducing the pressure drop through the unit and increasing its processing capacity. This option is especially attractive to refiners with an existing excess oxygen supply.

Several refiners are using oxygen enrichment technologies to increase the processing capacity of the SRU. These technologies are divided into three categories: low-level, mid-level and high-level oxygen enrichment.22

Low-level enrichment involves the injection of oxygen through a diffuser into the furnace combustion air. The concentration of oxygen can be raised to a maximum of 30 vol%, resulting in a 20–30% capacity increase in the SRU. Oxygen concentrations greater than 30 vol% require special materials. Mid-level oxygen enrichment technologies send pure oxygen into a specially designed burner through ports separate from those through which the air flows. These technologies generally utilise an overall oxygen concentration of up to 45 vol% to increase capacity by up to 75%. High-level oxygen enrichment requires that certain modifications be made to maintain the temperature of the reaction furnace at an acceptable level. To avoid unacceptable temperatures, the acid gas must not be burned directly with the enriched air stream. These types of processes can operate at oxygen concentrations of 45–100 vol% and increase SRU capacity by up to 150%. The benefits and drawbacks of oxygen enrichment are listed in Table 3.

Benefits Drawbacks• Retrofitting a unit for oxygen-enriched operation costs 5–25% less • High temperatures and poor mixing in the furnace can cause than building a new unit and takes a relatively short amount of time refractory damage23

• The flame is hotter, which often eliminates the need to use a • If the oxygen is not introduced to the SRU in the proper manner, portion of the acid gas to raise its temperature breakthrough can occur, resulting in damage to the waste heat boiler• Operating flexibility exists in that the oxygen-enriched SRU does • The cost of oxygen delivery to the reactor can be substantial, especially not have to run at full capacity for remote locations• The higher furnace temperatures destroy more ammonia, which • The increases in the sulphur flow rate and the H

2S content of the sulphur

can form salts that can cause plugging problems in downstream may necessitate a revamp of the sulphur handling, degasification and Claus beds forming facilities24

• Emissions are lower due to the improved performance of the absorber • The higher temperature of the sulphur could lead to combustion, unless a • The conversion of H

2S in the Claus process is higher cooling method is utilised to lower its temperature prior to entry into the pit

Benefits and drawbacks of oxygen enrichment



Switch to sub-dew point operationSulphur produced in the Claus reaction is a gas at temperatures above 130°C. At operating temperatures below 130°C, sulphur condenses and is adsorbed on the catalyst. Loss of sulphur to the catalyst reduces its partial pressure in the vapour phase, which in turn produces more sulphur via the Claus reaction to restore gas phase equilibrium. Operating in this temperature regime, however, also favours sulphur deposition on the catalyst, which can result in deactivation of the catalyst and the development of a large pressure drop through the bed. To solve this problem, several companies have developed processes that utilise Claus reactors followed by reactor beds that alternate between sub-dew point operation and online catalyst regeneration.25

Implement direct oxidationDirect oxidation processes increase sulphur recovery by using a selective oxidation catalyst that promotes the irreversible reaction of H2 S with O2

and limits the reversible reaction of H2S with SO2. Since the desired reaction is exothermic and high temperatures increase the occurrence of side reactions, the feed gas to the oxidation reactor must have a low H2S concentration. This technology typically works well as an addition to the end of a Claus train to meet regulations on SO2 stack emissions. It is necessary to ensure that the sulphur in the feed is in the proper form for oxidation, which can be accomplished by using a hydrogenation step to convert the sulphur compounds to H2S or by operating the Claus combustion system with an excess of H2S.

Increase sulphur yield by adding a reduction oxidation sectionReduction oxidation processes focus on reducing SO2 to elemental sulphur with H2 and/or CO in addition to selectively oxidising H2S to sulphur and water downstream of the traditional Claus converter. Using a selective reduction catalyst helps reduce side reactions that lead to H2S formation. This method is important

because the H2S concentration in the selective oxidation reactor must be kept relatively low to avoid temperature excursions and the shortening of catalyst life.26

Tail gas treatmentTail gas treaters are used to increase sulphur recovery by up to 99.9+% from the SRU while lowering operating costs and hydrogen consumption. Tail gas treaters can reduce SO2 emissions, improve reliability and increase run lengths.

Treatment of Claus tail gas is now a necessary part of the SRU to maximise sulphur recovery from residual H2S, SO2, COS and CS2, and to meet strict SO2 stack emission standards that are generally <10 ppmv. Typically, Claus tail gas is reduced and hydrolysed

over a CoMo catalyst to convert SO2, COS and CS2 to H2S before absorbing the H2S out of the stream and sending it back to the feed to the Claus unit. Another alternative is to selectively oxidise the tail gas feed to elemental sulphur.

Recently, Alon USA installed a new tail gas treating (TGT) unit at its refinery in Big Spring, Texas, to reduce SRU emissions. In designing this unit, the company chose not to rely on the existing TGT design at its other sulphur plant, as this unit experienced several operational problems. Instead, various design elements were modified in order to improve reliability, run length and operator approval. Adjustments were made to the air blowers, H2S/SO2 ratio, the analyser at the outlet of the amine

contactor, pumps, heat exchangers, lean/rich amine exchangers, TGT unit contactor, reactor catalyst support, reboiler tube bundle and return line, sight glasses, and process control system.27

Sulphur productionAdditional processes can be implemented to improve the value of recovered sulphur products from the refinery sulphur plant.

Adding a redox process for increased sulphur recovery Iron redox and chelated iron processes are among the alternatives to Claus technologies.28 These processes have demonstrated their ability to remove up to 80 tonnes/day of sulphur from gas feeds, but they are typically used to yield less than 20 tonnes/day. Operational issues related to these processes include the high cost of chemicals, chelate degradation during regeneration and foaming. Approximately 30–90% of the product sulphur obtained from these processes is obtained as elemental sulphur and can be used as chemical feedstock, while the end use of the remaining portion (10–70%) may be limited to the fertiliser market because of the presence of entrained water.

Improve sulphur yield with high-activity catalystTo increase sulphur recovery, refiners can modify the catalytic section of the SRU by adding improved traditional Claus catalysts, adding catalysts to improve sub-dew point operation and upgrading catalysts for oxidation and/or reduction processes.25 High-activity catalysts for Claus converters and TGT units are commercially available from a number of companies. High-grade titania or activated alumina are most often used in Claus converters to provide high conversion of CS2 and COS to H2S, and H2S to elemental sulphur via the Claus reaction.

Produce sulphuric acidImprovements in sulphuric acid-based processes are providing refiners with an economic option for increased sulphur recovery outside of Claus

Table 4:


Tail gas treaters are used to increase sulphur recovery by up to 99.9+% from the SRU while lowering operating costs and hydrogen consumption


technologies.29 Hydrogen sulphide is combusted to form sulphur dioxide, which is subsequently oxidised to sulphur trioxide (SO3). The SO3 is reacted with water in the vapour phase to form H2SO4. A unique advantage of producing sulphuric acid is the inherent synergy with the alkylation unit, which uses sulphuric acid as a catalyst. Where demand exists, the product-quality acid can be sold.

Case studyAs an example of a revamp that integrates several solutions, WorleyParsons was awarded a project to increase a large refinery’s SRU capacity.30 Two reactor trains, each consisting of two Claus reactors and a Shell Claus off-gas treating (SCOT) unit, produced 681 tonnes/day of sulphur at 70% capacity. The project involved increasing each train’s capacity to 1300 tonnes/day and improving unit reliability. The changes made to increase capacity are summarised in Table 4.

ConclusionAs refiners are faced with the global recession, impending legislation restricting CO2 emissions, the increasing use of heavy sour crudes and additional clean fuel specifi-cations, hydrogen production and sulphur plant technologies will need to be adapted to meet changing market demands.

In terms of hydrogen production —an energy-intensive process — the increase in demand for clean fuels will, ironically, conflict with efforts to curb CO2 emissions. Furthermore, the use of clean-burning natural gas as fuel for energy production and transportation may have an effect on the feedstock for hydrogen production as power plants and chemical companies are eyeing the same fuel/feedstock to reduce their carbon footprints.

The economics of sulphur plants are adversely affected by demand destruction, which could stretch over the next few years, putting further pressure on already poor refining margins. For the long term, refiners should examine emerging, novel

hydrogen production and sulphur recovery technologies for plant expansion and revamp opportunities as a means to satisfy environmental legislation and market requirements.

This article was prepared with excerpts from Hydrocarbon Publishing’s Worldwide Refinery Processing Review covering hydrogen production, purification and recovery, and sulphur plant technologies, as well as excerpts from the company’s multi-client strategic report, Future Refinery Operations to Meet Fuel Supply Security and Environmental Requirements.

References1 Fleshman J D, Cost-effective hydrogen plant revamps, AM-01-36, NPRA Annual Meeting, New Orleans, LA, 18–20 March 2001.2 Patel N, Insert flexibility into your hydrogen network — part 2, Hydrocarbon Processing, Oct 2005, 87.3 Ratan S, Increase hydrogen potential from existing plants, AIChE Spring National Meeting, New Orleans, LA, March 2002.4 Abrardo J M, Hydrogen technologies to meet refiners’ future needs, Hydrocarbon Processing, Feb 1995, 45.5 Luckwal K, Mandal K K, Improve hydrogen management of your refinery, Hydrocarbon Processing, Feb 2009, 55–61.6 2002 NPRA Q&A, question 137, p24.7 PTQ&A, PTQ, Winter 2004, 28.8 Zagoria A, Refinery hydrogen management — the big picture, Hydrocarbon Processing, Feb 2003, 45.9 Wentink P, Enhance refinery off-gases through integration, AIChE Spring National Meeting , April 2001.10 Patel N, Insert flexibility into your hydrogen network — part 1, Hydrocarbon Processing, Sept 2005, 75.11 Gupta S, Reformer technology for hydrogen, PTQ, Q2 2007, 73.12 1999 NPRA Q&A, question 10, p121.13 Boyce C A, Time for a new hydrogen plant?, Hydrocarbon Engineering, Feb 2004, 67.

14 Turk K, Treating tail gas, Hydrocarbon Engineering, Feb 2003, 89.15 2000 NPRA Q&A, question 10, p204.16 1998 NPRA Q&A, question 5, p149.17 1997 NPRA Q&A, question 3, p162.18 Kohl A, Gas Purification, Gulf Publishing Co, Houston, TX, 1997, 58.19 Sulphur technology review, Hydrocarbon Engineering, April 2005, 39.20 Fenderson S, Increasing sulphur processing capacity, Hydrocarbon Engineering, Dec 2002, 43.21 Menon R, Simple oxygen enrichment, Hydrocarbon Engineering, Feb 1999, 64.22 1999 NPRA Q&A, question 12, p197.23 1998 NPRA Q&A, question 6, p151.24 Tonjes M, Project strategies, Hydrocarbon Engineering, Feb 2003, 23. 25 Chen J K, Desulphurising fuels: know the basics, Chemical Engineering, Sept 2002, 66.26 Rameshni M, WorleyParsons, PROClaus: The New Standard for Claus Performance, 29 Aug 2007.27 Sanghavi K, Balance tail gas operations, Hydrocarbon Processing, June 2006, 71.28 Heguy D L, An iron grip on emissions, Hydrocarbon Engineering, July 2003, 18. 29 Cavalca C, Solutions for a cleaner future, Hydrocarbon Engineering, Aug 2007, 60.30 WorleyParsons, Sulphur Recovery Unit Expansion Case Studies, 28 Aug 2007.

The four authors are with Hydrocarbon Publishing Company:Adrienne M Blume is Managing Editor and holds an MA in English and Publishing.Email: [email protected] Patrick J Christensen is Technology Analyst with a BSCHE degree. Email: [email protected] P Goldhammer is Technology Analyst with a BS in Material Science and Engineering. Email: [email protected] W Yeung is Principal Consultant. He is a registered professional engineer and holds BSCHE, MSCHE and MBA degrees. Email: [email protected]

Desired outcome Revamp modificationReduction of pressure drop • Replaced reaction furnace burner • Replaced quench tower trays with random packing • Replaced TGU absorber trays with structured packingMinimise amine solvent loss • Water wash section was installed at the top of the amine absorberMore catalyst required for • Installed more catalyst in Claus reactors and tail gasincreased throughput hydrogenation reactorIncrease capacity without • MDEA solvent was replaced with ExxonMobil Flexsorb, allowingmodifying amine regenerator 33% more tail gas to be processed at a 20% reduction in amine circulation rateExtra cooling water circulation • New circulation water pump and extra air coolers were installed

Summary of an integrated SRU revamp

Table 4




mbi.indd 1 11/6/09 20:35:37

http://www.eptq.com/jump.aspx?a=MB&p=revamps09

A novel process to reduce aromatics and benzene in reformates

A reduction in aromatics, by 25–30 vol%, and benzene in finished gasolines, without

any octane barrel loss, is essential to further improve the quality of reformulated gasoline. As Table 1 shows, the majority of the quantitative aromatics (around 60 vol%) and benzene content (around 80 vol%) in finished gasolines is found in reformates, which in turn represent 35–60 vol% of the composition of all gasoline components.

This article looks at new ways of dealing with these issues and focuses on the reconfiguration of existing semi-regenerative reforming units at Moscow Oil Refinery. The aim of the research was to change the rates of major reactions to obtain novel routes, by loading various catalysts into the conventional reactors to obtain reformates with a reduced aromatics content and without octane barrel loss. The normal routes for removing benzene from reformates are either pre-fractionation of the reforming feedstock (to separate sources of benzene formation from the hydrocarbon feedstock) or processes that upgrade reforming post-fractionation cuts (to remove benzene through conversion to other hydrocarbons or through extraction for application to petrochemicals production).

Reducing aromatics content in reformates Figure 1 shows the components of finished gasolines in US reformulated gasoline (34 wt% of the total gasoline pool), which are major contributors to the aromatics content in reforming gasoline, catalytic cracking gasoline,

A catalytic scheme for reducing the aromatics and benzene content in reformates without octane barrel loss uses existing reforming units at Moscow Oil Refinery

MikhAil levinbuk, A Meling, v ZubeR and A lebedev Moscow Oil Refineryv khAvkin All-Union Scientific Research Institute of Petroleum Processing


Sources of RFG-gasolinecomponents

Gasoline pool (100 vol%)(RFG-gasoline – 34% US

gasoline market)

Reforming

Catalyticcracking

(FCC)

Naturalgas

Domesticagriculture

Figure 1 Components of finished gasolines in US reformulated gasoline

Table 1.

indicator euro-3 2002 euro-4 2005 euro-5 2009 Russia 2009 uSA (California) Content: Benzene, vol%,not more than 1.0 1.0 1.0 1-5 0.62 (2011)Sulphur, ppm, not more than 150 50 10 500 30Aromatics, vol%,not more than 42.0 35.0 35.0 – 25Olefins, vol%,not more than 18.0 14.0 14.0 – 6.5Oxygen, wt%,not more than 2.3 2.7 2.7 – 2.0Rvp, kPa,not more than 60.0 60.0 60.0 66.7 60.0

Modern quality standards for motor gasolines

Table 1

moscow.indd 1 24/8/09 09:58:42

alkylate gasoline, methyl tert-butyl ether (MTBE) and ethanol. The largest contributors to the total aromatics content are reforming gasolines, and limits on the solution of aromatics by high-octane components are connected mainly to the total oxygen content of the finished gasoline (not more than 2.0 vol%). Therefore, following the ban on using MTBE (with an oxygen content of 18 wt%) in US reformulated gasolines, the total ethanol content (oxygen content 32 wt%) can substitute only 50% of MTBE’s share in the structure of the entire gasoline pool (see Figure 2).

To preserve octane barrel and overcome these limitations, the alkylate component would need to increase considerably. This would require the construction of alkylation units with a combined annual capacity of about 8–10 million tonnes of alkylate. To reduce the capital investment, it would be more acceptable to take a number of innovative steps involving semi-regenerative reforming units. The combined capacities of such units worldwide is about 60% of the total required capacity. To define the focus of scientific research needed for these developments, certain thermodynamic parameters of the main reactions in the conventional reforming process need to be first considered.

The two main reactions dominating the reforming process are naphthene dehydrogenation and dehydro-cyclisation of n-paraffin hydrocarbons. The latter reaction, achieved using a process with a fixed catalyst bed, can be dealt with using a new process with a moving catalyst bed at a reduced pressure and higher temperature. This provides an increase in reformate yields and octane number valued by an increase in aromatics content in reformates. One of the main tasks when creating a new reforming process is the replacement of n-paraffins dehydro-cyclisation reactions from the conventional reforming process with a reaction that supports high octane numbers in reformate at the expense of non-aromatic hydrocarbons.

Figure 3 shows the dependence of octane number values on the fractionation composition of reforming


Other15% Other

17%

MTBE12%

Alkylate17% Alkylate

23%

Reformate17%

Reformate20%

Ethanol 6%

FCC gas39%

FCC gas34%

Gasoline component formulationwith ethanol under

California state standards

Gasoline component formulationwith MTBE under

California state standards

Figure 2 Gasoline component formulation (California state) with MTBE and ethanol

n – C7H16n – C8H18n – C9H20

Figure 3 Change in feedstock reforming octane number, depending on boiling temperature

Light naphthaisomerisation

Catalysts

n n

nstrong

Lubricant oilisomerisation

Catalysts

n

nweak

Medium naphtha and diesel fuel isomerisation

Catalysts

nmedium

Figure 4 The general concept of isomerisation catalysts development, depending on the content of carbon atoms in paraffins molecular

moscow.indd 2 12/8/09 12:26:28

feedstock. An “octane number pit” in the graph is determined by the 20–25 vol% of low-octane C7-C9 n-paraffins in the reforming feedstock that must be isomerised (rather than cracked) to increase octane number values and support acceptable reforming yields. The general concept of developing catalysts for isomerising n-paraffins with differing hydrocarbon chain lengths is presented in Figure 4. Depending on the length of the paraffin chain (C4-C6, C7-C18 and C25-C55), it is possible to develop three types of catalyst with different matrix acidities and the same noble metals content (for dehydrogenation) reaction.

This development requires the creation of a new catalyst for isomerising medium-chain paraffins (n-C7-C9) and the reloading of semi-regenerative reforming units with both new and established catalysts (see Figure 5). With the arrangement shown in Figure 5, isomerisation of C7-C9 n-paraffins and hydrogenation reactions to convert aromatics into naphthenes (with an increase in octane number of 71–72 RON) occur in the first reactor. Naphthene dehydro-genation reactions to form aromatics (with an increase in octane number of 95–96 RON) take place in the second reactor, which contains conventional catalysts. This new process is called hydroisoforming.

Figures 6 and 7 compare the results of upgrading narrow fractions from straight-run naphtha IBP-180ºC using traditional processes and hydro-isoforming: IBP-70ºC fraction using isomerisation; 70–100ºC and 105–180ºC fractions using traditional benzene and gasoline reforming (Figure 6); and 70–100ºC and 110–180ºC fractions using hydroisoforming (Figure 7).

The best results for reforming straight-run gasolines are seen in the hydroisoforming of the 110–180ºC fraction. In this case, the aromatics content is reduced by 10 vol%, compared with the traditional process, and the benzene content falls from 4–5 vol% to 1.2–1.5 vol%. This enables the application of the resulting reformates as reformulated gasoline. During hydroisoforming of the 70–100ºC fraction at low octane number


Reactor 1 Reactor 2 Reactor 3

Figure 5 Different options for loading new isomerisation and traditional catalysts in reforming reactors for the hydroisoforming process

IBP – 180°C

Isomerisation unit

Total gasoline pool:

Traditionalreforming unit

Traditionalreforming unit

Figure 6 Traditional upgrading of straight-run naphtha fraction IBP-180˚C by isomerisation, benzene and gasoline reforming

IBP – 180°C

Isomerisation unit

Total gasoline pool:

Hydroisoformingunit

Hydroisoformingunit

Figure 7 New approach to upgrading straight-run naphtha fraction IBP-180˚C

moscow.indd 3 12/8/09 12:26:52

values (78–83 RON), there is an increase in benzene content (to 1.0–1.5 vol%), which makes this option unacceptable for compounding the reformates into reformulated gasolines. During hydroisoforming of a wide, 80–180ºC fraction, the catalysate contains at least 2.0 vol% benzene, which again means the reformates cannot be compounded into reformulated gasoline. The problem of removing benzene during the reforming of the 70–100ºC fraction has to be resolved by alternate means.

Reducing benzene content in reformateFigure 8 shows different process options for removing benzene from reformates; the options are based on the pre- and post-fractionation of the reforming feedstock and catalysate. The principal drawback of all of these processes is the reduction in octane barrel in the product component of gasoline when benzene or its precursors are removed from the feedstock and reformates. Increased gasoline yields (when reducing benzene content) can be achieved by reforming narrow, straight-run naphtha fractions in two reforming units operating at different pressures. The influence of pressure on octane number and benzene content on the various fractions of reformates during straight-run naphtha reforming is shown in Figures 9 and 10. These


Pre-fractionation Post-fractionation

Figure 8 Process options for reducing benzene content in finished gasolines pre- and post-fractionation

17 atm28 atm

5 atm

Figure 9 Benzene content in reformate and reformate RON at various hydrogen pressures (without removing benzene-generating hydrocarbons from reforming feedstock, fr. 80–105°C)

17 atm

28 atm

5 atm

Figure 10 Benzene content in reformate and reformate RON at various hydrogen pressures (when removing benzene-generating hydrocarbons from reforming feedstock, fr. 105–180°C)

moscow.indd 4 12/8/09 12:27:12


flexitallic.indd 1 8/6/09 14:29:17

http://www.eptq.com/jump.aspx?a=Flexitallic&p=revamps09

figures also indicate that, during the reforming of the 80–105ºC and 105–180ºC fractions, there is a reverse dependence regarding benzene content in reformates when pressure changes in the range of 5–28 bar. This is explained by the contribution of dealkylation reactions involving heavy aromatics, in which the speed of reaction increases with pressure.

Moscow Oil Refinery has two semi-regenerative reforming units, with annual feedstock capacities of 1.0 and 0.3 million tonnes. They operate at pressures of 14 and 25 bar. Hence, the separate reforming of narrow reforming feedstock fractions (100–130ºC and 112–180ºC) in the two reforming units enables an increase in the volume of reformed feedstock

(with a lowered benzene content in the reformate). A comparison of the octane numbers, volumes of reforming feedstock and the contribution of alkylaromatic dealkylation reactions to the benzene content of reformates during the reforming of straight-run naphtha fractions at 92–180ºC and 112–180ºC shows a considerable reduction in gasoline yields when upgrading 112–180ºC fractions. This is represented by a decrease in the volume of reformatted straight-run naphtha from 11.0 to 9.2 wt%.

Separately reforming the 100–130ºC and 112–180ºC fractions in different units changes the contribution of dealkylation reactions to the benzene content in reformates and allows for an increase in the volume of reformatted straight-run naphtha to 10.3 wt%. This flow scheme was implemented at Moscow Oil Refinery in 2006 to 2007.

However, the scheme has two major disadvantages: the volume of reformatted feedstock is reduced by 10 rel.%; and the reformate produced by the smaller reforming unit presents certain limitations because it has a higher benzene content. Therefore, the main challenge in reducing the benzene content of reformates while preserving octane barrel is to develop new techniques for upgrading post-reformed light reformates.

Catalytic hydroisomerisationThe highest benzene content appears during the reforming of light, straight-run gasoline fractions, so the upgrading of light reformates was tested using novel catalysts for the isomerisation of C7-C9 n-paraffins in the hydroisoforming process. The results of hydroisomerising different types of light fraction feedstocks are shown in Figure 11. The octane numbers of the catalysates are higher than those of the initial reformates.

A variety of tests was carried out on upgrading light reformates using hydroisoforming catalysts. Figures 12 and 13 indicate that an increase in the octane numbers of light reformates during hydroisomerisation is achieved by reactions converting benzene into metylcyclopentane and cyclohexane, and by the isomerisation of n-C7

A

B

Figure 11 Process options for reducing benzene content by hydroisomerisation of light reformates (obtained from different narrow straight-run naphtha fractions)

Methylcyclopentaneand

cyclohexane ration

Catalysate RON

Figure 12 Dependence of the ratio of methylcyclopentane and cyclohexane and catalysate RON on temperature in the benzene hydroisomerisation reaction (fr. IBP=90°C reformate) V=1.5 h-1, H

2/CH=500 m3/m3)


moscow.indd 5 12/8/09 12:43:47

paraffins. Based on these results, a decision was made to upgrade various light reformates fractions in two semi-regenerative reforming units at Moscow Oil Refinery. As a result, catalytic processing of the IBP-90ºC fraction of light reformates, obtained from different straight-run naphtha fractions, delivers reformates with less than 0.1 vol% benzene without octane barrel loss (see Figures 12 to 14).

ConclusionThe research programme has introduced a novel scheme for reducing the aromatics and benzene content in reformates without octane barrel loss using two reforming units. As Figure 14 shows, the new concept can be realised if the heavy straight-run naphtha fraction is hydro-isomerised in the first semi-regenerative reforming unit. The benzene content is reduced when light reformate (obtained by reforming light straight-run naphtha in the second reforming unit) is hydro-isomerised in a separate, new reactor. So, this very new concept of reformed straight-run naphtha allows for decreased aromatics and benzene content with minimum capital costs and without octane barrel loss.

Reference1 Meister J, Crowe T, Keeson W, Stine M, Oil and Gas Journal, 11 Sept 2006, 38–45.

Mikhail Levinbuk is Assistant General Director of Moscow Oil Refinery and a professor of the Gubkin Russian State University of Oil and Gas. A Doctor of Technical Science, he has extensive experience in catalytic cracking

operations, synthesis of zeolites and developing catalysts. Email: [email protected] Meling is General Director of Moscow Oil Refinery.V Zuber is Deputy General Director, Operations, of Moscow Oil Refinery.

A Lebedev is Assistant General Director of the scientific production plant Neftechimiya at Moscow Oil Refinery.V Khavkin is Deputy General Director of All-Union Scientific Research Institute of Petroleum Processing.

Table 4: C7

C6Initial feed

Splitter

Splitter

Figure 13 Content of branched isomers (C6 and C

7) in the catalysate of light reformate

hydroisomerisation (fr. IBP-90˚C), depending on octane number

Figure 14 Total scheme for removing benzene and aromatics in the reformate stream through hydroisoforming and hydroisomerisation processes

PTQ1-4-4c_X3.pdf 1 23.02.2009 10:42:13 Uhrwww.ptqenquiry.com for further information


moscow.indd 6 12/8/09 12:44:04

http://www.eptq.com/jump.aspx?a=Buchen-ICS&p=revamps09

The new RecipCOM delivers diagnostics,

protection and therapy for your reciprocatingcompressors.

Spread the word!

For more information please contact

[email protected]

ww

w.h

oerb

iger

.com

Recipcom 06 2009 ptq.indd 1 04.06.2009 17:48:45


hoerbiger.indd 1 8/6/09 14:10:52

http://www.eptq.com/jump.aspx?a=Hoerbiger&p=revamps09

Hydroprocessing upgrades to meet changing fuels requirements

T he current economic environ-ment has imposed a reduction on capital expenditure budgets

for new expansions, along with some erosion in demand for fuel products, which is squeezing refinery margins. However, refiners are still facing challenges to meet stringent clean fuels regulations and maintain requirements for product quality.

Refiners have to think creatively in order to meet specifications and develop projects that are profitable, with returns on investment that are subject to a higher level of scrutiny in view of restrictions on capex. Light-heavy crude differentials have recently eroded, but the expectation is that the differentials will recover and encourage the development of projects that meet fuels regulations with some bottom of the barrel conversion to light products. There also appears to be momentum in the marketplace to increase diesel production, in the US in particular, with an expected increase in the use of diesel-powered engines for personal vehicles. Revamps of existing units can provide the refiner with the ability to generate high returns on restricted capital investments.

Chevron invented the modern hydrocracking process in 1959. In recent decades, Chevron Lummus Global (CLG) has debottlenecked and revamped a number of existing hydroprocessing units to meet changes in fuels regulations and feedstocks, achieve capacity increases and increase the output of light products. This article discusses four revamp configurations in all areas of hydroprocessing technology, some of which have been fully implemented

Adapting installed hydroprocessing units through a variety of schemes enables refiners to shift their fuels slates to meet changing demand and specifications

JAy PArekh and hArJeeT VirdiChevron Lummus Global

and others that are at various stages of project execution.

Conversion from gasoline/jet to diesel An Isocracking licensee operating a two-stage with recycle (TSR) configuration hydrocracker with intermediate distillation commis-sioned a revamp to expand unit capacity and shift the mode of operation from naphtha production to a distillate-selective operation. The objective of the revamp is to increase the nominal feed rate to the unit by 20% and to increase the diesel yield from zero to more than 40% (with a decrease in naphtha and jet yields).

The inherent advantage of a hydrocracker is the ability to shift the yield selectivity of the operation with a shift in the recycle cut point (RCP), which is defined as the cut point between the heaviest product and the

unconverted oil. By increasing the recycle cut point, more distillate products can be recovered in the intermediate distillation section and less unconverted oil is sent to the second stage for further conversion. The end result is less workload on the catalyst and a reduction in chemical hydrogen consumption, with less overall cracking to naphtha as well as a reduction in light ends make. This enables a refiner to increase feed rates while maintaining catalyst life, without having to debottleneck the gas recovery section of the plant. It results in a reduction in hydrogen demand, which fosters straight-forward debottlenecking to achieve higher throughput.

Since the unit was originally designed to co-produce naphtha and jet, a revamp to introduce diesel selectivity requires modifications to


Figure 1 Two-stage recycle (TSR) with vacuum column for diesel recovery

chevron .indd 1 12/8/09 12:51:48

clean second-stage environment, with overall rate constants much greater than the rate constants from the first stage. This second-stage environment permits full conversion of difficult feeds, with less than half the reactor volume needed compared to a single-stage once-through (SSOT) or single-stage recycle (SSREC). The obvious difference between the traditional TSR configuration and the SSRS configuration is that the effluent from the second stage flows directly to the inlet of the first stage, which provides the following benefits over a conventional TSR configuration:• Effluent from the second stage provides a heat sink for the first stage, reducing demand for first-stage quench gas by up to 40%• Excess hydrogen from the second stage is used to supplement the gas-to-oil requirement for the first stage• The overall recycle gas compressor load is reduced typically by up to 70%• Only one reactor furnace is required.

SSRS revamp The economics of a high-pressure hydroprocessing revamp are largely influenced by the recycle gas compressor costs. The SSRS flow scheme imposes a small incremental load on the recycle gas compressor. This is fairly intuitive for the consideration of a SSOT or SSREC revamp of a two-stage unit, but less intuitive for a TSR revamp. CLG recommended this solution for the refiner, and the scheme is shown in Figure 3. In the TSR configuration, a guard bed is added to the first stage, and an additional first-stage reactor is added between the second-stage effluent and the product fractionator. The guard bed was added to increase demetallation and the overall volume of the first-stage reactor to extend the length of the catalyst run. The unit (pre-revamp) is currently running at 133% of original design capacity. The addition of two new reactors will enable the refiner to increase the unit’s throughput by another 42%, for a total of 175% of original design. The revamp will also extend the run length by 30%. This will provide for a 228%

the existing fractionator in order to “lift” the level of diesel product. However, the ability to add another product draw is limited by the size of the column and the available reboiler duty. Therefore, an additional vacuum column will be added downstream of the fractionator to pull diesel product to a 350°C cut point. An additional heater is not required. Figure 1 shows a schematic of the TSR unit with the addition of the vacuum column. The remainder of the unit’s debottleneck involves modification of the recycle compressor, feed pumps and other minor capital expenditures on the heat exchange train. This revamp project is currently in detail engineering, with an expected startup date in 2012.

Two-stage hydrocracker revamp A licensee originally commissioned in the mid-1990s a TSR unit with selectivity towards middle distillates. The refiner was evaluating a project to increase unit capacity and extend run length significantly. A traditional revamp would necessitate additional reactor volume and modifications to the recycle gas compressor. However, a novel solution, using a process recently commercialised by CLG called single-stage reverse sequencing (SSRS), would enable the company to achieve its goals with substantially less capital expenditure.

A schematic of the SSRS flow scheme is shown in Figure 2. Like a TSR unit, the SSRS also takes advantage of a


2nd stagereactor

1st stagereactor

Figure 2 SSRS Isocracking

Newguard

RX

Newcombstage

RX

Existing1st stage

RX

Existing2nd stage

RX

Figure 3 Revamp configuration using reverse staging

chevron .indd 2 12/8/09 12:51:57

[email protected]

A Weatherford Company

AUSTRALIA - ASIA PACIFIC

+61 7 3867 5555EUROPE - MIDDLE EAST - AFRICA

+33 5 4902 1600 NORTH & SOUTH AMERICAS

+1 651 636 3900

Johnson screens has the products and service you need – across the globe. Call us or visit www.johnsonscreens.com to learn how our internals can make your reactor run at peak performance.

EMERGENCY SHUTDOWNS OFFER A REAL CHALLENGEQUICK, EFFICIENT AND LASTING REPAIR FURTHER TO AN EMERGENCY SHUTDOWN

ConsequencesThe whole unit had to be stopped and costly loss of production would be imminent.

SolutionJohnson Screens emergency shutdown procedure was activated on the same day:

• Johnson crew members arrived on scene at the refinery, inspected the faulty reactor and dismantled the damaged parts

• Meanwhile, new parts were being produced at the Johnson factory and were shipped within 48 hours of the first call

• The parts arrived at the refinery and the repair was performed by a crew of four Johnson Screens’ welders.

ResultsThe catalyst was reloaded five days after the first call to Johnson Screens and the catalytic reformer was restarted within a week.

Scheduled replacement of the three remaining sets of reactor internals was performed as initially planned four months later.

A major German refinery had planned the replacement of four sets of reactor internals in a catalytic reforming unit after 20 years of trouble-free operations.

Unexpected pressure drop was witnessed on one of the reactors four months before the planned shutdown date.

pubJS_210x297.ai 6/03/09 15:15:54


ohnson screens.indd 1 8/6/09 14:48:30

http://www.eptq.com/jump.aspx?a=Johnson&p=revamps09

increase in processed barrels per catalyst fill, compared to the original design, and will all be achieved using the existing recycle gas compressor.

This project was scheduled to start up in Q4 2009. The unit will continue to run in a maximum mid-distillate mode.

Hydrocracking for lubes andintegrated fuels and lubes productionThe benefits of hydrocracking to produce feeds for lubricant base stocks — as well as other downstream process operations such as FCCs and ethylene plants — are well known and in use in plants around the world. In

most of these cases, there is a dedicated lube hydrocracker, followed by dewaxing and finishing steps. What is more unusual is the integration of a hydrocracker, primarily devoted to making high-quality fuels (especially ultra-low sulphur, low-aromatic diesel), while also producing excellent feed for a dewaxing/finishing unit. The benefits of such an approach are clear: a lower capital investment than is required to build separate fuels and lubes hydrocrackers; and a lower cost of producing high-value lube base stocks.

This TSR configuration, with its innate advantages for making clean

fuels, can also provide a flexible platform for producing a range of lube base oils. The patented principle is shown in Figure 4, where the configuration can be set up to produce three lube base oil feedstocks, while maintaining to a large degree the predetermined overall conversion level for fuels. Each of the three lubricating oil streams is of different quality and boiling range. Investment to recover each stream for further Isodewaxing/Isofinishing steps can even be phased to match market requirements. Bharat Petroleum Corporation (BPCL), a refiner in India, recognised the potential of the TSR-

configured hydrocracker. With CLG’s TSR process, BPCL designed an integrated Isodewaxing/Isofinishing unit, using the flexibility available in its TSR to supply its forecast lube base oil market.

The company constructed the unit with minimal interruption to the imminent startup of the hydrocracker. Phasing in the design of the Isodewaxing/Isofinishing unit with the hydrocracker project before all of the major equipment had been ordered enabled BPCL and CLG to reduce costs by over 30%, compared to a standalone unit. Several major pieces of high-pressure equipment could be

eliminated for the lubes hydro-processing unit. BPCL’s lubes plant successfully started up one year following feed into the hydrocracker in 2006.

Residue upgradingThe need to process resid feeds with a high metal content, to run high severity operations in resid desulphurisation (RDS) units, or to extend cycle length and throughput via retrofits to plants already constructed, led to the idea of replacing the catalyst while the unit is on-stream. Liquid-filled upflow reactor (UFR) technology with on-stream catalyst replacement (OCR) involves the catalyst moving in countercurrent flow to a mix of residuum and hydrogen, to ensure full use of the catalyst. In operation, the most deactivated catalyst encounters the most reactive feed at the reactor’s (bottom) inlet, and spent catalyst is withdrawn from the bottom. The least reactive feed contacts the most active (fresh) catalyst at the reactor’s (top) exit. Fresh catalyst is introduced at the top during the on-stream catalyst replacement cycle. OCR technology is being commercially demonstrated in four units currently in operation.

As the need arose to extract most of the value provided by the OCR’s upflow reactor, while minimising capital investment and avoiding catalyst-handling facilities, the idea of the UFR emerged. While it cannot provide all of the demetallation to be expected of an OCR, the UFR is able to add utility in other ways. Since it is not a moving bed design, multiple beds of different catalysts can be used, including combinations of HDM and HDS catalysts (Figure 5). The major attribute of the UFR (as well as the OCR) is its very low pressure drop. This avoids typical limitations in the recycle compressor and, therefore, revamping an existing vacuum resid desulphurisation (VRDS) unit is feasible and economically attractive. This was one of the main reasons why the ENI Taranto Refinery in Italy selected UFR technology.

ENI Taranto’s residue hydro-conversion unit was revamped to expand the unit’s capacity by 25%.

gni sseco r

por

dyH

enoz en

oz noit ar a

p es .mt

A

g nikc ar cor

d yH

enoz

Vacuumseparation

zone

Atm.stripper

Figure 4 Isocracking for fuels and lubes


chevron .indd 3 12/8/09 12:52:12

Startup was in 2006. In addition to increased throughput, it has provided the refi ner with the ability to process heavier local crudes such as Tempa Rossa and Belaym.

ConclusionsIn order to provide refi ners with opportunities to revamp hydro-processing units at lower levels of total capital investment, CLG has developed and commercialised a number of process concepts to make the most of existing hardware. This is accomplished by making optimum use of the existing unit’s design and by integration of what were previously thought of as separate hydro-processing plants. All concepts discussed in this article have been designed for commercial units, two of which are currently operating and the other units will start up within the next few years.

Isocracking, Isodewaxing, Isofi nishing, OCR and UFR are marks of Chevron Lummus Global.

References1 Wade R, Vislocky J, Maesen T, Torchia D, Hydrocracking Catalyst developments and innovative processing scheme, NPRA Annual Meeting, Mar 2009.2 Spieler S, Mukherjee U, Dahlberg A, Upgrading residuum to fi nished products in integrated hydroprocessing platforms — solutions and challenges, NPRA Annual Meeting, Mar 2006.3 Steegstra J, Louie W S, Mukherjee U, Innovative and cost effective designs to achieve multiple clean fuels

solutions, BBTC 2004 Antwerp, Oct 2004.

Jay Parekh is Process Engineering Team Leader for Chevron Lummus Global, Richmond, California. He has worked for Chevron for 18 years, with over ten years in the hydroprocessing and refi ning industry. Email: [email protected] Virdi is Hydrocracking Technology Manager for Chevron Lummus Global, Richmond, California. He has worked in the hydroprocessing and refi ning industry for 28 years. Email: [email protected]

Figure 5 UFR vs OCR reactors


3225 Gallows Road, Fairfax, Virginia 22037-001, USAwww.exxonmobil.com/refiningtechnologies

703-846-2568 • fax [email protected]

BenzOUT TM

Low cost alternative to reducebenzene in gasoline.

• Avoid the Need for Hydrogen

• Expand Gasoline Volume

• Increase Octane

• Grassroots or Retrofits

Get Ready for 2011!Meet Benzene RegulationEnhance Product Value


chevron .indd 4 12/8/09 12:52:24

http://www.eptq.com/jump.aspx?a=Exxonmobil&p=revamps09

CFD study of a VDU feed inlet device and wash bed

V apour distribution in a large, packed tower is a critical factor in a column’s performance.

This article presents a revamp case for a vacuum distillation unit (VDU) column in an Indian refinery, where poor vapour distribution was identified with the help of a study involving computational fluid dynamics (CFD). Following the study, a new feed inlet device and its associated internals were installed, resulting in significant improvement to the quality and quantity of vacuum gas oil (VGO) produced, reduced quantities of slop wax and increased column throughput.

The refinery last revamped its crude distillation unit (CDU) in September 2004, to increase its processing capacity from 160 000–220 000 bpsd. However, this revamp did not enhance the VDU’s capacity, so it remained limited to the crude processing level established before the revamp. Although the VDU’s design capacity was limited, higher crude rates were nonetheless processed (180 000 bpsd during 2006 and 2007). The outcome included penalties in VGO quality, accelerated erosion on the feed inlet device, premature damage to the wash zone packing and a high rate of slop wax production.

The velocity at the feed nozzle exceeded the critical level at the chosen rate of throughput. During shutdowns, erosion of the feed inlet device, as a result of this excessive velocity, was observed. Furthermore, distillate yields were lower than expected and the entrainment of heavy ends was high, resulting in under-performance in the wash zone. During the search for a solution to these problems, it was decided that there should be minimal change to the size of the transfer line,

Analysis of vapour velocity profiles at the entry to the wash zone of a VDU column enables a higher feed rate and improved output quality

DebAngsU RAy Indian Oil Corporation AjAy ARoRA and Anne PhAnikUmAR Sulzer Chemtech

because this would involve major modification and a long outage for the column.

To mitigate the problems and to increase the vacuum column’s throughput by 10%, a joint study by the refinery and Sulzer Chemtech employed CFD analysis to assess the quality of vapour velocity profiles at the entry of the wash zone packing.

CFD model and validationThe aim of the CFD analysis was to assess the performance of the vapour distribution systems with increased capacity in the flash zone of a vacuum

column. To conduct a detailed analysis of vapour velocity profiles at the entry of the wash zone packing, a CFD model of the existing configuration (see Figure 1) was prepared. Y-velocity (vertical velocity) distribution across the collector tray above the feed inlet device and at the entry of the wash zone packing is shown in Figures 2 and 3 respectively. The velocity value is characterised by different colours, where blue represents the lowest velocity and red is the velocity peak of the scale. It can be seen in Figures 2 and 3 that the velocity peaks (shown in red) are not evenly distributed across

Figure 1 CFD model of existing configuration


sulzer.indd 1 12/8/09 12:59:40

Recommendations for column internals Apart from recommendations for a larger feed inlet nozzle and feed inlet device, further recommendations covered column packing and internals:• Bed-1 (top pump) With the existing

the column’s cross-section and that they are mainly concentrated at the column’s centre and at the periphery of the column’s inner wall. This non-uniform distribution of vapour above the feed inlet device and at the entry to the wash zone packing was leading to very high entrainment of heavier ends to the wash zone packing, which tallied with the observed problems of reduced product quality, premature damage to the wash zone packing and high rate of slop wax production.

A subsequent CFD run was based on an increased feed nozzle diameter with a larger Shell Schoepentoeter as the feed inlet device. The results of the rerun led to a recommendation to carry out modifications to enlarge the feed nozzle from ϕ52 inch to ϕ68 inch and replace the existing conventional feed inlet device with a Shell Schoepentoeter that was compatible in size with the new feed nozzle. Figure 4 shows the CFD model of the recommended arrangement.

Enlargement of the feed nozzle would help to reduce the feed inlet velocity to an acceptable level of ≈60 m/sec, which is significantly lower than the critical velocity. The Y-velocity distribution across the collector tray above the feed inlet device, and at the entry of the wash zone packing in accordance with the recommended configuration, is shown in Figures 5 and 6 respectively. Velocity peaks

are distributed evenly across the column’s cross-section. A uniform velocity profile at the inlet of the wash bed would significantly reduce the entrainment of heavier oil and thereby improve the performance of the packing bed.


Figure 2 Y-velocity profile across collector tray of existing

configuration

Figure 4 CFD model of new configuration

Figure 3 Y-velocity profile at packing entry of existing

configuration

sulzer.indd 2 12/8/09 12:59:51

random packing, the flood value was high at 90%. The recommendation was to replace this with structured packing to remove the bottleneck and minimise the pressure drop• Bed-2(LVGO-HVGOtopfractionation)The recommendation was to replace structured packing with next-generation packing to further reduce the pressure drop• Bed-3 (LVGO-HVGO bottomfractionation)No changes were found to be necessary• Bed-4 (HVGO PA) The recommendation was to partially replace the existing structured packing with an increase in bed height to achieve the same spare capacity of the other beds• Bed-4 (wash zone) The recommendation was to replace

structured packing with next- generation packing.

ImplementingrecommendationsSulzer’s recommendations were reviewed by the refinery, with the aim of keeping investment costs to a minimum. The following changes were settled upon:• Enlargement of the feed nozzle with an appropriate expander in the transfer line (ϕ52-68 inch)• Replacement of the feed inlet device with a 68 inch Shell Schoepentoeter inlet devices• Since the feed inlet device was to be enlarged, the chimney collector tray needed to be relocated. The refinery’s decision was to dismantle the chimney tray and install a new one with appropriate positional changes.

The new chimney was designed to help improve distribution of vapour to the wash bed (see Figure 4) • The shutdown period would be limited, so the recommendations made with respect to the column packing were closely examined and it was decided to replace 40% of the Bed-1 random packing with larger-sized random packing to avoid flooding in the bed bottom. The rest of the packed bed replacements were deferred until the unit’s next turnaround.

These modifications were carried out during the refinery’s July 2007 turnaround. The process of concept to commissioning was completed within six months.

TestrunafterrevampFollowing the revamp, the refinery


Figure5Y-velocity profile across collector tray of new

configuration

Figure6Y-velocity profile at packing entry of new configuration

Attribute Earlieroperation TestrunathigherRemarks (HS$run) t’put(HS$run)

Atmosphericvacuumunit(AVU)

CDU t’put, m3/hr 1250 (1088 TPH) 1400 (1218 TPH) No limitation observed in CDU section for operating the column at this t’putCDU COT, ºC 360 364 No limitation observed in CDU furnaceVDU t’put, m3/hr 540 599 No limitation observed in operating the unit at the required t’putVDU COT, ºC 410 412 No limitation observed in VDU furnace for maintaining the COT at this t’putVGO yield (wt% on crude) 22.5 24.2 Removing the critical velocity bottleneck in the transfer line enabled consistent improvement Vac slop wax flow, M3/hr 60 6.7 There had been significant improvement in the flash zone vapour liquid separation, resulting in a big reduction of entrainment in the slop wax collector tray

Keyperformanceresultsbeforeandafterrevamp

Table1

sulzer.indd 3 12/8/09 13:00:08

carried out a test run. Table 1 shows the unit’s key parameters before and after the revamp.

A further increase in crude throughput was restricted by limitations in the vacuum residue pumping capacity (Tables 2 and 3).

ConclusionsThe refi nery achieved a number of benefi ts as a result of the revamp. No physical constraints were observed in operating the AVU and VDU at the desired throughput of 1400 M3/hr (a 12% increase). The VGO yield in the test run was 24.2 wt% (based on the crude supply), compared with a yield of 22.5 wt% before the modifi cations were introduced. The vacuum residue and slop wax yield fell to 26.2 wt%, against a pre-modifi cation yield of 28 wt%, a clear indication of improved performance of the feed inlet device.

The refi nery was able to increase its crude throughput for signifi cant fi nancial benefi t from a limited investment.

Debangsu Ray is Deputy General Manager of Indian Oil’s Panipat refi nery. He has over 25 years’ experience in refi nery operations and technical services with various Indian Oil refi neries and been responsible for implementing projects to improve energy conservation and refi nery distillate. Email: [email protected]

Ajay Arora is Assistant Manager, Business Development & Technology for Sulzer India Limited. He has a master’s degree in process engineering and plant design from IIT Delhi. Email: [email protected]

Anne Phanikumar is a Senior Mechanical Engineer for Sulzer Chemtech, Singapore. He has a master’s degree in computer-integrated manufacturing from Nanyang Technological University, Singapore, and experience in the design and analysis of heavy equipment. Email: [email protected]

Product Normalised, wt% Overall AVU material balanceLPG 1.5Naphtha 14.0HN 2.2ATF 15.3HK 3.3HSD 13.3HVGO 24.2SR&Vac Slop 26.2Total 100.0

Test run yield pattern

Product Attribute Value RemarksLVGO Recovery (@365ºC) 95% Normal Density, kg/m3 888.8 Normal

HVGO HVGO quality was signifi cantly superior in quality with colour at ASTM 6.5 against earlier results of >8. ASTM 95% and EP gap narrowed down to 4°C against earlier value of 20–25°C Density, kg/m3 928.1 Normal Sulphur, wt% 2.3 Normal Basic nitrogen 308.3 ppm Normal CCR 0.39 wt% Normal Metal content, ppm Ni = 0.15 Cu = 0.04 Ni = 0.08 Fe = 1.09 V = 0.45 Normal

SR PEN, 1/10 mm 72 Signifi cant improvement observed in SR PEN. Earlier value was 150–200. Better separation in the fl ash zone and effective stripping steam contributed to this improvement

Product quality during the test run

Table 2

VFF

-IN

NO

VATION OF

THE

YE

AR

VFF-Twin-Pak:dream body – amazing qualities.

best mass transfer

highest mechanical strength

lowest pressure drophighest capacity

P. O. Box 552, D - 56225 Ransbach-Baumbach, Phone +49(0)2623/895-0, [email protected], http://www.vff.com

www.vff-twin-pak.de

+++Tower Packings+++Catalyst Supports+++Internals+++Droplet Separators+++

Table 3



sulzer.indd 4 13/8/09 16:34:50

http://www.eptq.com/jump.aspx?a=Vereinigte&p=revamps09

Application of CFD in NOx reduction

Most refineries processing crude oil employ a large number of combustion units,

including fired heaters for heating process fluids and boilers for steam generation. Refinery off-gases and natural gas are the fuels of choice for combustion. Combustion is a process in which fuel reacts with oxygen to produce carbon dioxide and water. It also results in a visible flame and useful heat. The adiabatic flame temperature is about 1650°C, which results in the formation of NOx.

The term NOx is used in general to describe a group of compounds of oxides of nitrogen, such as NO, NO2, N2O2 and others. NO2 is widely used in combustion calculations as a representative of NOx. Three types of NOx are formed during combustion: thermal NOx, fuel NOx and prompt NOx. Thermal NOx is formed when N2 and O2 combine at high temperatures. Nitrogen from fuel combines with oxygen to form fuel NOx and prompt NOx. Figure 1 shows the various sources of NOx that are emitted to the atmosphere.

It is necessary to reduce NOx because it is a greenhouse gas. It produces ozone (O3) and forms acid rain. The US government passed the Clean Air Act in the 1970s to reduce air pollution and set air quality standards. The Environmental Protection Agency (EPA) and state pollution control agencies, such as the Texas Commission of Environmental Quality (TECQ), enforce the emissions limits. Over the last 40 years, there have been a number of additions and revisions, which have set stricter emission standards for air pollutants, including NOx. Currently,

A study of parameters affecting NOx formation in an ageing boiler unit confirmed

the measures needed to meet emissions targets

ShAhebAz MullA and G RAMAN Technip USA

for the state of Texas, TECQ NOx emission limits for old boilers with a duty of more than 40 MMBtu/hr are set at of 0.06 lb/MMBtu (HHV basis).

The formation of NOx increases with a higher combustion temperature. Fuel gas containing nitrogen produces more fuel NOx. Air also has a linear relationship with NOx formation. In theory, NOx can be significantly reduced if stiochiometric air is supplied for combustion and all oxygen reacts with fuel to form carbon dioxide and water. In practice, some nitrogen reacts with oxygen to form NOx. This results in incomplete combustion, formation of carbon monoxide and loss of useful heat. Hence, it is necessary to provide excess air for combustion. In natural draft systems, it is common to provide 15–20% excess air. In a preheated, forced draft system, the excess air is about 10–15%.

NOx reduction methods1,2

NOx reduction methods can be broadly classified into two groups: pre-combustion and post-combustion.

Pre-combustion methodsThese methods apply various techniques to reduce NOx formation during the combustion process: • low NO

x burners One of the most

popular methods is installing low NOx burners. Staged air, staged fuel, external flue gas recirculation and ultra-low NOx burners with internal flue gas recirculation are commonly used. Each type works on a different principle and is best suited to a certain combustion system in achieving reduced NOx. Most of these burners help in reducing flame temperatures and residence time, thereby reducing thermal NOx.

Selecting a particular type of low NOx burner depends on several factors, such as type of fired

Courtesy: EPA Technical Bulletin 1999

Figure 1 Sources of NOx1


technip.indd 1 12/8/09 13:17:16

NOx emissions results, which applies to the case reported here. A comparison of NOx emissions from a typical boiler burner in various circumstances is shown in Table 1.

Case study A study was carried out on an old boiler designed to produce superheated steam at about 40 bar. Tangentially gas-fired burners were employed to provide the necessary duty. To increase boiler efficiency, the unit had been fitted with two

equipment, type of draft system (forced or natural), air preheat or no air preheat, gas or liquid fuel. A combustion specialist can best advise on the selection of a suitable type of burner• Flue gas recirculation External flue gas recirculation (FGR) reduces NOx by recirculating part of the flue gas back into the combustion zone. FGR reduces the flame temperature by diluting the oxygen content of the combustion air and distributing the heat of combustion into a larger amount of flue gas. It is recommended that less than 30% of flue gas generated should be recirculated. FGR not only reduces NOx, but also reduces the boiler heat duty. To be able to produce the same amount of steam before flue gas injection, the boiler has to fire more fuel. Therefore, the burners have to be designed for higher firing rates. FGR is frequently used along with low NOx burners in systems where air preheat is present• Water or steam injection Water or steam injection can lower NOx levels, but result in a decrease in thermal efficiency of 3–4% and a decrease in the capacity of the boiler. It is advised to use water/steam injection for NOx reduction only under special operating conditions.

Post-combustion methods These methods reduce NOx after it has formed rather than controlling it during the combustion process. Selective catalytic reduction (SCR) is a method in which a mixture of flue gas and ammonia passes through a catalyst at a specific temperature (230–510°C) to achieve NOx reduction to N2 and water vapour. Three types of catalysts are used in SCR, depending on the flue gas temperature. Low-temperature, platinum-based catalysts are very effective in an ideal temperature range of 260–340°C. Medium-temperature, vanadium-titanium-based catalyst has an ideal temperature range of 290–400°C. For high flue gas temperatures, a zeolite-based catalyst, with an ideal operating temperature range of 450–510°C, can be used.

Selective non-catalytic reduction (SNCR) is another method, in which

ammonia or urea is injected directly into the firebox or duct containing the flue gas at a high temperature. SNCR requires a temperature of about 870–1370°C. NOx reduction of 75% can be achieved with SNCR, whereas with SCR a reduction of 90% can be achieved.

There are other methods available and more information on them can be found in EPA1 and American Petroleum Institute (API)2 publi-cations. Two NOx reduction methods are often combined to meet the stricter


Figure 2 Boiler and flue gas ducting before modifications

Figure 3 Recommended modifications

Air preheat Steam inj FGR Gas/oil fired NOx, lb/MMBtu (HHV) (approx) ✗ ✗ ✗ Natural gas 0.25 ✗ ✗ ✗ Oil 0.5 ✓ ✗ ✗ Natural gas 0.50 ✓ ✗ ✗ Oil 0.55 ✗ ✓ ✗ Natural gas 0.10 ✗ ✓ ✗ Oil 0.4 ✓ ✓ ✗ Natural gas 0.20 ✓ ✓ ✗ Oil 0.45 ✗ ✗ ✓ Natural gas 0.04

NOx emissions for a typical boiler burner

Table 1

technip.indd 2 12/8/09 13:17:25

Ljungström-type air preheaters, arranged in parallel. The rotary type air preheaters used exiting flue gas to heat up incoming combustion air. Although the efficiency of the boiler was increased by means of an air preheater, control of NOx output was

adversely affected because a rise in air temperature increases thermal NOx

formation. A forced draft fan was used to supply the necessary combustion air through a ducting system to the boiler. The boiler was currently producing NOx in the range of

0.07–0.1 lb/MMBtu (HHV basis). The goal was to meet the NOx emission limit of 0.06 lb/MMBtu (HHV basis) with air preheat conditions. Figure 2 shows the boiler and duct system for which reduction of NOx was to be addressed.

RecommendationsInstallation of an SCR unit requires significant capital investment in the reactor, catalyst, soot blowers, fan, ducting and expansion joints. A large plot space is also required for all of this equipment. The SCR catalyst is sensitive to temperature fluctuations. High-temperature flue gas can sinter the metals, closing the pores and increasing the system’s pressure drop. The low temperature of the flue gas can result in deposition of ammonium salts on the catalyst bed. The installation of a SCR along with ducting also requires a longer shutdown of the boiler. After reviewing the economics of this option, it was decided to explore other routes.

The next option was to use ultra-low NOx burners (ULNB). The technology in burners for NOx

reduction has advanced such that ULNB can achieve NOx emissions of 0.04 lb/MMBtu without air preheat. Since the installation under study had preheated air, to achieve a 0.06 lb/MMBtu NOx limit, it was recommended to replace the existing burners with ULNB and recirculate part of the flue gas from the boiler; this would further lower the combustion zone temperature and assist in lower NOx formation.

About 20–25% of the flue gas would be drawn by a new FGR fan and mixed with combustion air before being sent to the boiler for combustion. Proper mixing of the air and flue gas is important for NOx reduction. The burners would be designed for multiple refinery gas compositions and provide a stable flame for the fuel gas air and flue gas combustion mixture. This option was economical and had less turnaround time compared to the SCR option and it was therefore selected. Figure 3 shows the proposed scheme to meet the recommended NOx levels.


26" WC

0" WC

Figure 4 Pressure profile

Figure 5 Velocity profile

Figure 6 Oxygen profile

0.2 mole frac

0.049 mole frac

150 ft/s

0 ft/s

technip.indd 3 12/8/09 13:17:39

ResultsIt was estimated that these modifications, along with proper mixing of the flue gas with air and equal distribution, would result in lower boiler temperature and hence lower NOx formation. To validate our results, two computational fluid dynamic (CFD) studies using Fluent flow modelling software were carried out. The first study was carried out to confirm that flue gas and air were adequately mixed in the ducting system before entering the burners. Inadequate mixing and flow to the burners could cause problems with flame shape, stability and NOx

emissions. A second CFD model would show the decrease in boiler temperature following the addition of FGR.

The CFD ducting results show various profiles taken across a plane in the ducting system. The pressure profile in Figure 4 was used to estimate the pressure drop for adequate sizing of the fan and new FGR ducts. The velocity profile in Figure 5 confirmed that the existing air ducts were adequate for the higher gas rates. The O2 profile shows the mixing of the flue gas into the air after the air preheater. Overall, the study showed that the design of the new ducting system and the tie-in to the existing system were adequate, and produced good air and flue gas distribution to the burner windboxes.

In the CFD study of the boiler, a comparison was made between a base case in which the boiler was modelled with new burners, with no FGR and with 25% FGR. The flow patterns from the tangential burners showed a strong swirling motion in the boiler. The addition of 25% FGR resulted in better mixing conditions, which promoted good combustion rates for the refinery gas and rapid combustion of CO. The lower peak temperature and lower peak heat transfer to the walls resulted in a lower exit temperature. The addition of flue gas was not harmful to the combustion, heat transfer or flue gas flow from the boiler, but would aid in NOx reduction with lower flame temperatures.

The vertical and horizontal views in

ConclusionThe CFD studies adequately pointed out the behaviour of important combustion parameters that affect NOx formation. The combustion zone temperature was reduced using ULNB and flue gas injection. The drop in temperature would result in lower levels of thermal NOx. Nitrogen from the fuel gas was eliminated in order to reduce fuel NOx. Thus, it was concluded that installing low NOx

burners along with FGR in this boiler system, with air preheat present,

Figure 7 Temperature (F) profile with no FGR

Figures 7 and 8 of the temperature profile near the burner show that with FGR the flue gas temperature around the burner area is about 1650°C, whereas with 25% FGR the temperature falls to about 1370°C. The temperature profile confirmed that this drop in temperature will decrease thermal NOx formation.

Adding FGR to the boiler resulted in a drop of boiler efficiency by 2%. For the same steam production before revamp, the fired duty had to be increased by 10 MMBtu/hr.


technip.indd 4 12/8/09 13:17:53


would result in NOx reduction and the achievement of NOx targets.

We would like to thank John Rizopoulos, Anupam Bhaumik and Duraivelan Dakshinamoorthy for their valuable comments.

References1 EPA Technical Bulletin: Nitrogen Oxides (NO

x), Why and How they are controlled.

2 Post Combustion NOx Control for Fired

Equipment in General Refinery Services, API Recommended Practice 536, Second Edition, December 2006.

Shahebaz Mulla is a Process Engineer with Technip USA, Houston, Texas. He has worked on a number of NO

x reduction projects. He

holds a MS degree in petroleum engineering from Texas A&M University. Email: [email protected]

Dr G Raman is a Senior Supervising Process Engineer with Technip USA, where he is responsible for managing process engineering activities. He has 25 years’ experience and is a registered Professional Engineer in the state of Texas. Email: [email protected]

Figure 8 Temperature (F) profile with 25% FGR

technip.indd 5 13/8/09 15:53:09

Boiler Services, Inc.RENTECH

For more information, email us at [email protected]

visit us online at WWW.RENTECHSERVICES.COM

or call us at 325.672.2900

RENTECH Boiler Services specializes in engineered repairs, rebuilds and

upgrades of industrial boilers using headered membrane waterwall design.

We retrofit any style of boiler, making RENTECH your one-source boiler company.

Our work meets NBIC and ASME standards. To reduce operating costs, eliminate

shutdowns, allow faster start-up and cool-down, and reduce emissions, call for

personal service from RENTECH Boiler Services.

O U R E N G I N E E R I N G A T R E N T E C H C R E A T E S


http://www.eptq.com/jump.aspx?a=Rentech&p=revamps09

Our advancedcatalytic engineering

has powerful attractionAxens’ HR series ACE™ technology catalysts deliver sustained, on-spec ultra-low-sulfur diesel. The powerful desulfurization activity of this new series is particularly enhanced in combination with EquiFlow™ reactor internals and Catapac™ dense loading technology.

With its submicron-level control, ACE technology provides dual activity for superior levels of sulfur and nitrogen removal in all applications. Easy to load and activate, HR series catalysts also offer excellent regeneration properties.

Axens’ quality products are based on reference technology – commercially proven, dependable, reliable and cost-effective.

The performance improvement specialistswww.axens.net

For more information

Paris +33 1 47 14 25 14 Houston +1 713 840 1133 Email [email protected]

HR_ptq.indd 1 28/11/08 12:44:39


axens.indd 1 9/3/09 11:41:28

http://www.eptq.com/jump.aspx?a=Axens&p=revamps09

cover revamps - digitalrefining · this revamp was performed by altairstrickland. the main scope...

Documents