lng custody transfer handbook

SECOND EDITION

G.I.I.G.N.L.

LNG CUSTODY TRANSFER HANDBOOK

G.I.I.G.N.L 2001

GIIGNL - DS TML/Z - CG - 2001/10/09

SECOND EDITION

DISCLAIMER

This new edition of the LNG Custody Transfer Handbook reflects best current practice at the time of publication.

The purpose of this handbook is to serve as a reference manual, but it is neither a Standard nor a Specification.

G.I.I.G.N.L , and any of its members, disclaim any direct or indirect liability as to information contained in this document for any industrial, commercial or other use whatsoever.

G.I.I.G.N.L.: Groupe International des Importateurs de Gaz Nature! Liquefie - Paris (International Group of Liquefied Natural Gas Importers - Paris (France))

GIIGNL - DS TMUZ - CG - 2001/10/09

SECOND EDITION

LNG CUSTODY TRANSFER HANDBOOK

1. INTRODUCTION 3

2. GENERAL DESCRIPTION OF THE MEASUREMENT 3

2.1. GENERAL FORMULA FOR CALCULATING THE LNG ENERGY TRANSFERRED 3

2.2. GENERAL SCHEME OF THE MEASUREMENT OPERATIONS 4

2.2.1. Volume 4

2.2.2. Density 4

2.2.3. Gross calorific value 4

2.2.4. Energy of the gas displaced by the transfer of LNG 4

2.2.5. Flowchart for determining the energy transferred 4

2.3. INSTRUMENTS USED 5 2.3.1. For the determination of the LNG volume 5

2.3.2. For the determination of LNG density and gross calorific value 5

2.3.3. For the energy of displaced gas 5

2.3.4. Periodic instruments recalibration 6

2.4. STANDARDISATION 6 3. VOLUME MEASUREMENT 6

3.1. GAUGE TABLES 6 3.1.1. Use of gauge tables 6

3.1.2. Correction tables 6

3.1.2.1 Correction according to the condition of the LNG carrier 6

3.1.2.2 Corrections according to the temperatures in the liquid and gaseous phases 12

3.1.3. Approval by authorities 12

3.1.4. Inaccuracy of the table 12

3.2. INSTRUMENTS AND METHODS FOR MEASURING THE LEVEL OF LIQUID IN THE LNG CARRIER'S TANKS 12

3.2.1. Main liquid level gauging devices 12 3.2.1.1 Capacitance gauge 12

3.2.1.2 Float gauge 12

3.2.1.3 Microwave gauge 16

3.2.2. Timing of the level measurement 16 3.2.2.1 In an FOB agreement 16

3.2.2.2 In a GIF or DES agreement 16

3.2.3. Readings 16 3.2.3.1 Reading of the level with float

gauges 16

3.2.3.2 Reading of the level with capacitance and microwave gauges 16


3.2.4. Correction of readings 16 3.2.4.1 Float gauge 16

3.2.4.2 Capacitance and microwave

gauge 16 3.2.5. Use of spare level gauge 17

3.2.6. Complete unloading (tank stripping) 17

3.3. CALCULATION OF THE VOLUME OF LNG TRANSFERRED 17

3.4. INACCURACY OF THE VOLUME MEASUREMENT 17

3.4.1. Cargo Liquid Lines 23 4. TEMPERATURE MEASUREMENT 23

4.1. LIQUID TEMPERATURE 23 4.1.1. Device 23

4.1.2. Testing and accuracy 25

4.2. VAPOUR TEMPERATURE 25 5. VAPOUR PRESSURE MEASUREMENT 25

6. SAMPLING OF LNG 25

6.1. LNG QUALITY 25

6.2. SAMPLING PRINCIPLES 25

6.3. SAMPLING POINT 26

6.4. SAMPLING PROBES 26

6.5. PIPING ARRANGEMENT BETWEEN SAMPLING PROBE AND VAPORISER 29

6.6. LNG VAPORISER AND CONTROL DEVICES 29

6.6.1. Main devices 29

6.6.2. Description of vaporising devices 29

6.6.3. Auxiliary vaporisation control devices 29

6.6.4. Operating parameters 31

6.7. COMPRESSOR FOR TRANSFERRING GAS SAMPLE 31

6.7.1. GAS SAMPLE HOLDER 31 6.8. GAS SAMPLE CONDITIONING 32

6.8.1. Gas sample bottles 32

6.8.2. Direct piping to gas analyser 32

6.9. EXAMPLES OF GENERAL ARRANGEMENT OF SAMPLING DEVICES 36

6.10. PERFORMANCES OF THE DEVICES 36

6.11. SAMPLING PROCEDURE 37

6.11.1. Sampling period 37

6.11.2. Sampling frequency 37

6.11.3. Purging 37

6.11.4. Sampling parameters 37

1

SECOND EDITION

6.11.5. Utilisation of gas sample bottles 37 6.12. SPOT SAMPLING DEVICE 38

7. GAS ANALYSIS 38 7.1. TYPE OF GAS CHROMATOGRAPH 38

7.1.1. General arrangement 38

7.1.2. Columns 38

7.1.3. Detectors 40

7.1.4. Carrier gas 40

7.1.5. Quality of the separation of components 40

7.2. INTEGRATOR AND DATA PROCESSING 41 7.2.1. Integrator system 41

7.2.2. Data processing 41

7.3. CALIBRATION 41 7.3.1. Calibration procedure 41

7.3.2. Calibration gas/working standard 41

7.4. QUANTITATIVE ANALYSIS 42

7.4.1. Response factors 42 7.5. ENVIRONMENT FOR A GAS

CHROMATOGRAPHIC SYSTEM 42

7.6. ANALYSIS OF REGASIFIED LNG AND RETAINED SAMPLES 42

7.7. INACCURACY OF GAS ANALYSIS 43

7.8. IMPURITIES 43

7.8.1. Carbon dioxide 43

7.8.2. Sulphur 43

7.8.2.1 Total sulphur 43

7.8.2.2 Sulphur components 43

7.8.3. Mercury 43 8. DENSITY 44

8.1. GENERAL 44

8.2. DENSITY CALCULATION METHODS 44

8.3. REVISED KLOSEK-Mc KINLEY METHOD 44

8.3.1. Limits of the method 44

8.3.2. Formula 46

8.3.3. Charts available for calculation 46

8.3.4. Example of LNG density calculation 46

8.3.5. Rounding off 46

9. GROSS CALORIFIC VALUE 46 9.1. GENERAL 46

9.2. METHOD OF DETERMINATION OF THE GROSS CALORIFIC VALUE 47

9.2.1. Determination with the help of calorimeters 47

9.2.2. Determination of GCV by calculation 47

9.2.2.1 Examples of formula 47


9.2.2.2 Examples of charts of basic physical constants 48

9.2.2.3 Example of calculation 48

10. ANALYSIS REPORT 49 10.1. IDENTIFICATION 49

10.2. BASIC DATA 49

10.3. RESULTS 49

11. ENERGY OF GAS DISPLACED DURING LOADING OR UNLOADING OPERATION 49

11.1. ENERGY OF GAS DISPLACED FROM THE TANKS OF THE LNG CARRIER 49

11.2. ENERGY OF GAS CONSUMED AS FUEL BY THE LNG CARRIER 50

12. ENERGY TRANSFER MEASUREMENT 50

13. OVERALL INACCURACY OF THE ENERGY TRANSFER MEASUREMENT 50

13.1. VOLUME 50

13.2. DENSITY 50

13.3. GROSS CALORIFIC VALUE 50

13.4. GAS DISPLACED 52

13.5. TOTAL INACCURACY IN THE DETERMINATION OF THE ENERGY TRANSFERRED 52

14. LNG SALES CONTRACT CUSTODY TRANSFER CHECKLIST 52

ENCLOSURE 1: Conversion factor table 54

ENCLOSURE 2: ISO Standards 55

ENCLOSURE 3: Other relevant standards & 58 references

List of figures - List of tables 59

References 60

Appendices 61

2

SECOND EDITION

1. INTRODUCTION Following the publication in 1985 by the N.B.S. of its study "LNG Measurement - A User's Manual for Custody Transfer" [8], the Executive Committee of the G.I.I.G.N.L. (Groupe International des Importateurs de Gaz Nature! Liquefie) considered it would be useful to write a handbook, as simple and as practical as possible, aimed at organisations involved with the measurement of the energy transferred in the form of LNG in the context of a LNG purchase and sales agreement, whether this salebeF.O.B. orC.I.F.

During its session of October 1987, the General Assembly of G.I.I.G.N.L. decided that this practical handbook should be drawn up by a Study Group comprising companies of the G.I.I.G.N.L. and co- ordinated by Distrigas S.A (B).

The methods described in this handbook could serve to improve existing procedures. They could also be used in purchase and sales agreements for the G.I.I.G.N.L. members and serve as a reference in new import agreements.

This handbook is based on the measurement methods most used by G.I.I.G.N.L. members.

The apparatus used is accepted as is, and detailed tests of this apparatus can be found in "LNG Measurement Study" of N.B.S. [8].

We wish to thank the companies - BG (UK) - Distrigas Boston (USA) - Enagas (E) - Kansai Electric Power Co (JP) - Snam (I) - Tokyo Electric Power Co (JP) - Tokyo Gas Co Ltd (JP) - Ruhrgas (D) - CMS Energy Trunkline LNG (USA) for their co- operation in producing this handbook, and more particularly Gaz de France for drawing up Chapters 6 and 7 of this manual and Osaka Gas Co Ltd for co-ordinating the studies of the Japanese companies.

SECOND EDITION OCTOBER 2001 Following the publication of the ISO 13398:1997 standard "LNG - Procedure for custody transfer on board ship", the G.I.I.G.N.L. General Assembly requested the G.I.I.G.N.L. Study Group to revise the original edition (March 1991) of this G.I.I.G.N.L. LNG Custody Transfer Handbook, particularly taking into account this new ISO standard.

All 13 sections of the original edition have been reviewed and updated where appropriate. The following sections have been thoroughly revised: 2. General description of the measurement 3. Volume measurement

6. Sampling of LNG 7. Gas analysis

Moreover, a new section was added: 14. LNG Sales contract custody transfer checklist. Worked out examples for LNG density and GCV have been rearranged in Appendices 1 and 2.

We wish to thank all companies and organisations and their delegates who together contributed to this second edition, viz. (in alphabetical order):


Advantica Technologies Ltd. (UK) BG International (UK) CMS Energy Trunkline LNG Company (USA) Distrigas (B) Enagas(E) Gaz de France (F) Nigeria LNG (Nl) NKKK (JP) Osaka Gas (JP) Rete Gas Italia (I) SIGTTO (UK) Tokyo Gas (JP) Tractebel LNG North America (USA)

2. GENERAL DESCRIPTION OF THE MEASUREMENT

2.1. GENERAL FORMULA FOR CALCULATING THE LNG ENERGY TRANSFERRED

The formula for calculating the LNG transferred depends on the contractual sales conditions. These can relate to an FOB sale, a GIF sale or a DES sale (Incoterms 2000).

In the case of an FOB sale, the determination of the energy transferred and invoiced for will be made in the loading port (FOB: Free On Board). In the case of a GIF or a DES sale, the energy transferred and invoiced for will be determined in the unloading port (GIF: Cost Insurance Freight - DES: Delivery Ex Ship).

In all cases, the formula can be summarised as follows:

E = (\/LWG.DLNG.GC\/LNG)-gas displaced

E =the energy transferred from the loading facilities to the LNG carrier or from the LNG carrier to the unloading facilities. In international LNG trading, the energy transferred is most frequently expressed in millions of British Thermal Units (106 BTU or MMBTU); although tis is not a SI energy unit, this unit will be adopted in this handbook. A conversion factor table for other commonly used energy units can be found in Enclosure 1.

VLNG =the volume of LNG loaded or unloaded in m3.

DLNG =the density of LNG loaded or unloaded in kg/m3.

GCVLNG =the gross calorific value of the LNG loaded or unloaded in MMBTU/kg.

3

SECOND EDITION

Egas displaced = the energy of the gas in gaseous form, also in MMBTU, which is either:

sent back onshore by the LNG carrier when loading. In most cases, this energy is returned free of charge to the loading facilities: Egasdisplaced is then nil; or received by the LNG carrier when unloading in replacement of the LNG transferred. In this case it is generally taken into account.

2.2. GENERAL SCHEME OF THE MEASUREMENT OPERATIONS

The objective is to measure the quantity of energy loaded from production facilities into an LNG carrier, or unloaded from an LNG carrier to a receiving terminal.

From the above formula, it can be inferred that 4 elements must be measured and calculated:

LNG volume, LNG density, LNG gross calorific value, energy of the gas displaced during the transfer of LNG.

For a graphic overview please refer to the 'measurement flowchart' on page 5.

2.2.1. Volume The method chosen of measuring the volume is based on the LNG carrier's instruments, mainly the use of level gauges and calibration tables.

Usually a quantity of LNG, called 'heel', remains on board after unloading so as to keep the tanks cold. Determination of the volume transferred requires two measurements, one before and one after loading or unloading; so the result will be two LNG volumes. The difference between the larger volume and the smaller volume will represent the volume of liquid transferred.

For an accurate custody transfer it is recommended that LNG manifolds on ship's deck be in an identical inventory condition during both custody transfer surveys (CTS) : either completely filled with LNG both during the opening custody transfer (i.e. before (un)loading) and the closing custody transfer (i.e. after (un)loading), or otherwise be drained during both the opening and closing CTS (custody transfer survey).

In some cases the LNG carrier must be completely emptied, e.g. before a long period of inactivity. In this case a special procedure explained in section 3.2.6 is followed for determination of the volume transferred.

The use of in-line flowmeters has not been considered since, at time of writing, the accuracy of this apparatus has not been acceptable to the industry.

2.2.2. Density The density of LNG is determined by calculation from the composition of the LNG transferred and the temperature of the LNG from measurements in the LNG carrier's tanks.

2.2.3. Gross calorific value The composition of the LNG is used to calculate the gross calorific value.

2.2.4. Energy of the gas displaced by the transfer of LNG

This energy is calculated according to the composition and volume of the gas displaced, and the pressure and temperature of the gas inside the tanks of the LNG carrier before loading , resp. after unloading.

The calculation procedure is explained in Section 11.

2.2.5. Flowchart for determining the energy transferred

GIIGNL - DS TML/Z - CG - 2001/10/09 4

SECOND EDITION

FLOWCHART FOR DETERMINING THE ENERGY TRANSFERRED

2.3. INSTRUMENTS USED 2.3.1. For the determination of the LNG volume For the determination of the LNG volume the following are required:

the calibration tables, including the main gauge tables for each tank and the correction tables for list, trim, tank contraction and possible additional factors according to the type of level measuring devices, the equipment for measuring the level of LNG in the LNG carrier's tanks which are either float gauges and/or capacitance type level gauges and/or microwave (radar) type level gauges. Each cargo tank usually has two level gauge systems installed, one designated as 'main1 or 'primary' and the other as 'secondary', temperature probes distributed over the height of the LNG carrier's tanks, all measuring devices required by the correction factors.

2.3.2. For the determination of LNG density and gross calorific value

The determination of the density and the gross calorific value of the LNG transferred will be made on the basis of the average composition of the LNG obtained by:

continuous or discontinuous sampling of LNG,

gas chromatographic analysis,


a calculation based on the average composition of the LNG, its average temperature and the coefficients given by the National Bureau of Standards for the density [10], a calculation based on the average composition of LNG and characteristics of elementary components (GCV, molar volume, molar weight) given by reference tables or standards for the gross calorific value.

2.3.3. For the energy of displaced gas The energy of the displaced gas can be determined from:

sampling of the gas displaced, a gas chromatographic analysis of this sampling enabling the GCV to be calculated, pressure and temperature measurements within the LNG carrier's tanks.

However, for the determination of the energy displaced, some parameters such as pressure, gas composition and temperature can be estimated from experience and taken as constant for both custody transfer surveys before and after (un)loading.

For instance, the displaced gas may be assumed to be pure methane. This assumption will hardly increase the overall inaccuracy.

5

SECOND EDITION

2.3.4. Periodic instruments recalibration It is recommended that, unless it is specified by the fiscal authorities or by the Classification Society, Buyer and Seller agree on the periodicity of recalibration intervals, e.g. at each drydocking.

2.4. STANDARDISATION International standards exist for the classical methods and techniques used for LNG Custody Transfer such as ISO 6568 for gas chromatography and ISO 6976 for calculation of the GCV of gas.

On the other hand many LNG shipping projects, especially existing ones, specify GPA 2261-72 for gas chromatography and IP 251/76 or GPA 2145-86 for the calculation of the GCV of gas. Buyer and Seller may approve these editions, or a more recent edition. As far as methods and techniques dealing with static measurement procedures for LNG are concerned, it must be noted that international standards have been issued in recent years by ISO (see Enclosure 2 page 55). The recommendations included in these documents and future international standards might be considered for new applications.

3. VOLUME MEASUREMENT 3.1. GAUGE TABLES 3.1.1. Use of gauge tables The gauge tables are numerical tables which, relate the height of the liquid in an LNG carrier's tank, to the volume contained in that tank. The height may need to be corrected taking into consideration various factors.

An independent surveyor usually produces the gauge tables during the building of the LNG carrier. They take into account the configuration of the tank, its contraction according to the temperature of the liquid, and the volume occupied by various devices, e.g. cargo pumps.

The calibration tables are usually divided into: main gauge tables: height/volume correlation in ideal conditions, correction tables taking into account actual conditions of the LNG carrier and its measuring instruments.

For each LNG carrier there is one main gauge table per tank. Generally the volumes are given for heights varying cm by cm, the volume for intermediate heights in mm being calculated by interpolation. An example of a gauge table is given in Table 1 (see page 7). The examples used in this section are taken from a vessel with prismatic cargo tanks. The same principles generally apply to those vessels with spherical cargo tanks.


These tables are established for level measurements using the main level-measuring device installed in each tank.

To avoid these interpolations, which can be a source of accuracy, the most used parts of the gauge tables - i.e. heights between 10 and 60 cm and heights corresponding to a volume between 95% and 98% of the total volume of the tank - sometimes are developed and the volumes will be calculated mm by mm. This then reduces the determination of the volume to a mere reading in a table (see Table 2 page 8, Table 3 page 9).

Various methods exist for establishing the gauge tables. The main methods are:

macrometrology with tapes,

laser measuring system,

photogrammetric measuring system.

It is not the purpose of this handbook to describe the different methods [1 - 8].

For details of calibration procedures for tanks, reference can be made to existing ISO standards (see Enclosure 2 page 55).

3.1.2. Correction tables The gauge tables are completed with correction tables established according to:

the condition of the LNG carrier (trim/list), the temperature in the tank that influences contraction or expansion of the tank, the temperature in the gaseous phase, and/or the density of the LNG, influencing the level measuring devices.

It should be noted that LNG vessels normally have two level measurement devices in each cargo tank (and often of two different types) and that correction tables are specific to a level gauge. Using the correction tables for the wrong gauge can result in significant inaccuracies.

3.1.2.1 Correction according to the condition of the LNG carrier

The gauge tables are established for an LNG carrier with zero list and trim. Therefore, it will be necessary to correct the height reading to take into account a list or a trim which is not zero.

Correction tables are made up according to:

the position of the gauge in the tanks, the list of the LNG carrier (see figure 1, page 10), the trim of the LNG carrier (see figure 2, page 10).

6

SECOND EDITION

These corrections can be positive or negative. So the real height will be equal to the algebraic sum of the height reading, the correction for list and the correction for trim. These tables are made up in degrees for the list and in metres for the trim, with fixed steps of variation. For intermediate values, the correction will be calculated by interpolation.

In practice, these corrections are not used frequently since the LNG carrier's cargo officer will usually manage the vessel's ballast to obtain zero list and trim.

3.1.2.2 Corrections according to the temperatures in the liquid and gaseous phases

The corrections are related to the volume variations resulting from the contraction of the tanks and their insulation according to the temperature of the liquid and gaseous phases.

This phenomenon is significant for LNG carriers with self-supporting tanks. Table 4 gives an example of these tables (see page 11).

3.1.3. Approval by authorities The gauge tables may be approved by either the authorities of the countries concerned with the LNG sale and purchase or by independent sworn measurers.

This approval may be valid for a limited duration, generally 10 to 12 years, provided there are no modifications to the tanks. For the European Community, this approval corresponds to a Community Directive.

When an LNG carrier is put into operation, a list of all works on the tanks must be supplied, and the tanks must be inspected for any modifications which might affect the volume.

In the case of any distortion or modification to a tank, the gauge table must be adjusted accordingly.

3.1.4. Inaccuracy of the table The inaccuracy generally guaranteed by the contractor's calibrations is 0.2% at ambient temperature. The study carried out by the National Bureau of Standards [8] on the calibration of the tanks of LNG carriers shows that the real inaccuracy is far better, and is about 0.05% to 0.1 %.

Also according to this study, the systematic inaccuracy due to the effect of shrinkage of the tanks when they have been cooled down should not be more than 0.07%. Therefore for a tank of 26,000 m3, the maximum guaranteed inaccuracy would be 52 m3 LNG, and even less by half if we consider the work of the NBS.


3.2. INSTRUMENTS AND METHODS FOR MEASURING THE LEVEL OF LIQUID IN THE LNG CARRIER'S TANKS 3.2.1. Main liquid level gauging devices The main types of gauges are:

electrical capacitance gauge (cfr. ISO 8309)

float gauge (cfr. ISO 10574)

microwave type gauge (cfr. ISO 13689)

Any of these gauges can serve as the main and usual instruments for measuring the height of the liquid. Usually (but not always) two of these three types are installed. One of these should be agreed as the main (primary) level gauging device by Buyer and Seller. The gauge not used will be considered as the auxiliary (secondary) gauge. A few LNG shipping projects do not specify the type and only specify the required accuracy (ex. 7.5 mm or better).On old ships other devices may be found, such as nitrogen bubbling devices, but the accuracy of these is generally lower.

3.2.1.1 Capacitance gauge The electrical capacitance gauge (see figure 3, page 13) consists of two concentric aluminium tubes. The inner tube is supported by the outer tube by means of concentric insulators placed regularly spaced intervals along the whole length of the tubes. The resulting assembly forms a series of cylindrical capacitors, having the same total height as the cargo tank of the LNG carrier.

The LNG, according to its level, will fill the space between the concentric tubes. The liquid affects the dieletric characteristics such that, by measuring the change in capacitance, the height of the LNG in the annular space, and hence the level in the tank, can be determined. The contraction of the aluminium tube at low temperature may be taken into account to correct the level measurement.

The accuracy of the measurement resulting from the calibration of the dimensions and the linearity of the capacitor and of the electronics should be, for the gauges as a whole, 7.5 mm [1].

3.2.1.2 Float gauge Measurements are made with a float hanging on a tape or a ribbon (see figure 4, page 14). According to the level of the liquid, the float is displaced, and the tape or the ribbon on which it hangs is unrolled or rolled up on a drum whose cycles of rotation are recorded. This enables the position of the probe, and thus the level of liquid in the tank, to be known.

With float gauges, it is necessary to take into account the shrinkage of the ribbon, according to the temperature of the gaseous phase and the height of the liquid, and the density of the LNG, which will influence the float buoyancy. The correction tables will tabulate the corrections for these effects.

12

SECOND EDITION

FIGURE 3

ELECTRICAL CAPACITANCE TYPE LEVEL GAUGE

1. Outer aluminium tube. 2. Inner aluminium tube. 3. Concentric electrical insulator. 4. Isolation of inner tube sections by a gap or dielectric plug. 5. Isolation from the tank bottom. 6. Bolting together the sections of the outer tube making a single electrical conductor. 7. Transfer line of the signals from the outer tube and each centre of the inner tube to a control junction box

outside the cargo tank. 8. LNG cargo tank.

GIIGNL - DS TML/Z - CG - 2001/10/09 13

SECOND EDITION

FIGURE 4

FLOAT TYPE LEVEL GAUGE

GIIGNL - DS TML/Z - CG - 2001/10/09 14

SECOND EDITION

FIGURE 5

MICROWAVE TYPE LEVEL GAUGE

SECOND EDITION

The corrections for temperature are required only in the case of a stainless steel ribbon. In the case of an invar ribbon, the shrinkage is much less and is generally considered as negligible.

The precision of this type of gauge, designed for marine application, is in the range of 4 mm to 8 mm.

3.2.1.3 Microwave gauge The microwave gauge works on the same principle as a ship's radar (see figure 5, page 15). A transmitter is mounted on the top of the cargo tank and emits radar waves vertically down towards the surface of the liquid. The signal is reflected from the surface, received by the transmitter's antenna and sent back to the control panel. The signal is then processed to determine the distance of the liquid surface from the transmitter and hence ullage.

Since all the level detection components are mounted external to the cargo tank, the microwave system allows for the possibility of changing the gauge in service.

The precision of this type of gauge is claimed to be better than 7.5 mm.

3.2.2. Timing of the level measurement 3.2.2.1 In an FOB agreement The level readings will be made just before starting to load, when the loading arms have been connected and before starting to cool them down.

This level reading will enable the determination of the quantity of LNG remaining on board as cooling liquid, also called "heel".

The second level reading will be made 15 to 30 minutes after the end of loading, when the surface of the liquid is nearly stabilised.

3.2.2.2 In a CIF or DES agreement In the case of a CIF or DES agreement, proceed as above mutatis mutandis. It is also a good idea to wait 15 to 30 minutes after the end of unloading so that the surface of the remaining liquid is nearly stabilised.

3.2.3. Readings It is good practice that all readings are witnessed by both parties. The readings of the levels of liquid in the tanks are taken after the readings of the list (port or starboard), and the trim (bow or stern) of the LNG carrier. The temperatures of the liquid and the gaseous phase are also measured (see 4.1. and 4.2.). For some vessels, the atmospheric pressure is also read.

3.2.3.1 Reading of the level with float gauges A test could be carried out before the reading by comparing the height indicated by the gauge in its stowed position with that given by the last calibration


check in this position. If this test is satisfactory the level readings can be taken. It should be noted that the surface of the liquid is not motionless: the liquid may be in an effervescent state and is subject to the movements of the ship. It is advisable to take several readings, from 2 to 6 according to the amplitude of the movement of the float, the height recorded being the average of the maximum and minimum readings.

If the level indication is unusual, the float may be stuck; it is suggested that it is raised and lowered again in an attempt to obtain the expected reading.

It must be noted that float gauges are always stowed in their fixed upper position when sailing so that the ribbon or the tape does not break due to liquid movements.

The float is released before the first reading which can be carried out either locally, by reading the level indication on the gauge head or, if equipped with a transmitter, by reading it on a digital display in the cargo control room.

3.2.3.2 Reading of the level with capacitance and microwave gauges

The reading is taken in the cargo control room of the LNG carrier. On older capacitance gauges, the level is read for each tank several times, up to 5 times at regular intervals, and the arithmetic average value is then calculated.

For modern systems, a computer processes all the information, including averaging of the level readings over time, temperature and pressure, and draws on computer based gauging tables to produce a printed document containing all the ship-generated information required for the custody transfer, however not always including gas displaced.

3.2.4. Correction of readings 3.2.4.1 Float gauge The readings made on the measurement appliances should be corrected according to:

list, trim, density of LNG, affecting float buoyancy, coefficient of contraction of the material and the insulation of the tanks; this coefficient is applicable in the case of self-supporting tanks (table4, seepage 11), temperature of the gaseous phase if the ribbon is not made of invar.

The corrections are made by using tables.

3.2.4.2 Capacitance and microwave gauge In this case, only the corrections for list and trim and the contraction of the tanks are taken into consideration. For accurate level measurement the contraction of the capacitance gauge at low temperature may need to be considered. For some

16

SECOND EDITION

microwave level gauges a temperature compensation of the microwave guide pipe is necessary.

Modern computer based systems usually can accept trim and list data either manually or from external sensors and automatically apply the corrections.

3.2.5. Use of spare level gauge When the main or primary level gauge cannot be used, the secondary level gauge is used to measure the level of LNG. If the calibration tables of the LNG carrier are available only for the main level gauge, a conversion table is required in order to take into account the respective locations of main and secondary gauges, or the statistical differences between the two level gauge measurements, and to evaluate the corresponding corrections which must be applied to level measurement before using the calibration tables.

3.2.6. Complete unloading (tank stripping) For the complete unloading of an LNG carrier with prismatic tanks, the condition of the LNG carrier will be with a maximum trim in order to concentrate the LNG in the part of the tank occupied by the gauge and the pumps.

A procedure may consist of unloading until the minimum measurable height indicated in the calibration table.

The remaining immeasurable LNG may be unloaded using stripping pumps by agreement with all parties. Afterwards the remaining quantity can be vaporised by warming up to a certain temperature e.g. -80C. At that temperature the LNG carrier is considered to be empty of LNG. These stripping and warming up operations require several additional hours of unloading time.

The energy of this remaining LNG transferred either in the liquid or in gaseous form can also be determined by mutual agreement by all parties.

The technical possibilities of the receiving terminal must also be taken into account.

3.3. CALCULATION OF THE VOLUME OF LNG TRANSFERRED

This calculation is illustrated by an example given in tables 5 (see page 18) and 6 (see page 19) showing the results of the volume determination before and after loading the LNG cargo, with the following hypotheses:

a "Gaz Transport"-type LNG carrier with 5 invar membrane prismatic tanks, in each tank, one float gauge with a stainless steel ribbon.

The example as illustrated is representative of an LNG carrier with an older level gauging system. A


similar procedure would be adopted for the secondary gauging system of a modern level gauging system. For the primary system on a modern instrumented LNG carrier, the procedure is carried out automatically by computer using essentially the same methods, however not always taking into account the gas displaced.

3.4. INACCURACY OF THE VOLUME MEASUREMENT

Table 11 gives an example of the overall inaccuracy in volume determination as a result of the respective inaccuracy of each measurement and of the gauge table.

This example is based on loading of a tank of a "Gaz Transport"-type membrane LNG carrier with a total capacity of 26,770 cubic metres.

Upon arrival at the loading port, the heel represents 1.5% of the total capacity of the tank which is then filled to 98%.

17

SECOND EDITION

iNoies:

(1) For measurement of temperature in the gaseous phase, see section 4.

(2) Correction for ribbon shrinkage of the float level gauge due to the cryogenic temperature in the gaseous phase according to table 7 (see page 20).

(3) Correction for LNG density, established from the density calculated on the basis of the LNG composition, according to table 8 (see page 21).

(4) Correction for the list corresponding to the liquid height according to the correction tables, an example of which is shown in table 9 (see page 21) for tank n 1. In this case the position of the gauge is at the starboard side of the ship's centre line.

(5) Correction for the trim corresponding to the liquid height according to the correction tables, an example of which is shown in table 10 (see page 22) for tank n 1 (an interpolation was made in this case between the correction values for a 50 cm and a 100 cm trim). In this case the position of the gauge is on the stern-side of the sideline of the tank.

(6) Algebraic sum of corrections (2), (3), (4) and (5).


(7) Corrected height resulting from the algebraic sum of the five previous columns.

(8) Determination of the liquid volume given in the calibration tables from the corrected height. These tables are established for heights varying mm by mm (see example in table 1, see page 7) on the basis of certified tables indicating volumes for heights varying cm by cm. An example is given in table 2 (see page 8) for the empty tank n 1 and in table 3 (see page 9) for the full tank n 2.

18

SECOND EDITION

GIIGNL - DS TML/Z - CG - 2001/10/09 19

SECOND EDITION

TABLE 10

EXAMPLE OF TRIM CORRECTION TABLE

SECOND EDITION

1. Volume on arrival in the loading port: 1.5% of the total volume indicated in the gauge tables.

2. Rating inaccuracy of the volume indicated in the gauge tables: 0.2%.

3. Height of the liquid for 1.5% of the tank volume given by the gauge tables.

4. Inaccuracy in the measurement of the LNG level by the level gauge, i.e. 7.5 mm. This inaccuracy takes the tank configuration at the height of 412 mm into account.

6. See n 2.

8. See n 4.

10. The overall inaccuracy is obtained as the square root of the sum of the squares of the inaccuracies from the quadratic combination of inaccuracies (2), (4), (6) and (8). It therefore represents about 0.21% of the quantity of LNG loaded.

NOTE For spherical tanks, owing to their geometry, the volume inaccuracy, as a result of the small inaccuracies of the level gauge, is significantly less than for prismatic tanks.

To illustrate, assuming a nominal 33,500 m3 capacity tank, arrival heel 40 m3, the depth in the tank is about 0.8 m and the inaccuracy due to level gauge is 15 mm ( 7.5 mm). This represents a volume of only about 1.5 m3.

3.4.1. Cargo Liquid Lines The previous section addresses the inaccuracy of volume measurement of the cargo tanks. To have a full picture, the issue of contents of the cargo liquid lines needs to be considered.

For most flat-deck' designs of LNG Carrier (membrane and IHI-SPB), the arrangement of cargo lines on deck is such that, at completion of cargo operations, all liquid left in the liquid lines can drain by gravity back to a cargo tank.


Once the drainage is completed, custody transfer measurements can proceed as described and there is no need to consider liquid in the cargo lines.

For all spherical tank designs and, to a lesser extent, some 'flat-deck' designs where the manifold valves are below the crossover lines, some consideration is needed for undrainable liquid.

Since significant volumes of LNG may remain in the cargo manifolds and crossovers after completion of delivery, the normal approach is to pre-cool and completely fill the cargo lines with LNG prior to the first CTS (Custody Transfer Survey) reading on arrival. The assumption is that LNG volume in liquid lines is the same at the time of both CTS readings, and therefore can be ignored in the calculation.

4. TEMPERATURE MEASUREMENT

4.1. LIQUID TEMPERATURE

4.1.1. Device The LNG temperature is measured by probes placed at different heights in the tanks. These probes are generally three- or four-wire platinum resistance temperature sensors, of which there are typically five per tank.

The variation in resistance, according to the temperature, is converted into degrees Celsius with the help of a data acquisition computer equipped with a printer (table 12, see page 24).

Table 12 shows an example of a printout of LNG temperatures when the tanks are filled to 98% capacity with LNG.

Figure 6 (see page 24) shows a diagram of temperature measuring devices installed on a LNG carrier.

In this example, 5 probes are immersed in LNG in each tank.

The liquid temperature is calculated once the loading operations are over, when it is a matter of determining the quantity loaded, and before the unloading operations, when it is a matter of determining the quantity unloaded.

Thermocouples are not used for LNG temperature measurement within custody transfer because they are less sensitive and often give a less accurate measurement than platinum resistance probes. In addition their installation is more complex (compensation cables...). They may be installed sometimes for control or simple indication (such as cooling down or heating of the tank).

23

SECOND EDITION

4.1.2. Testing and accuracy The probes are tested and recalibrated at regular intervals. The accuracy of the platinum resistance probes varies between 0.1 and 0.2C for temperatures ranging between -145 and -165C.

It is preferable to adjust the measurement for operation at low temperature.

The overall accuracy of the temperature measuring chain can be estimated at about 0.5C (probes, cable, signal converter display or printer).

The influence of temperature measurement accuracies on the determination of LNG density (see paragraph 8.3) is important. For instance, for LNG with an average density in the range 440 - 470 kg/m3, and at a temperature around -162C, the relative accuracy on density calculation, due to an accuracy of 0.5C on temperature measurement, is about 0.15%.

4.2. VAPOUR TEMPERATURE The temperature in the gaseous phase of the tanks is used to determine the quantity of gas displaced during the loading and unloading operations, or the level correction of the float gauge due to ribbon shrinkage.

It is the result of the average value of the temperatures indicated by the probes not immersed in the LNG.

For some LNG shipping contracts, an accuracy of 1.5 C (in the range -145 to + 40 C) is required.

5. VAPOUR PRESSURE MEASUREMENT

Vapour pressure measurements can be taken with a pressure gauge, which indicates the pressure in the gas spaces of the cargo tanks. This pressure is needed to calculate the energy of displaced gas, see Section 11. For this, it is necessary for the pressure to be in absolute terms. If the ship's instrumentation measures pressure in 'gauge' terms, then the atmospheric pressure must be recorded and added to the gauge pressure.

The pressure value is recorded, with the atmospheric pressure if appropriate, at the time of taking the other CTS readings.

In some LNG shipping contracts the required pressure measurement accuracy is specified as 1 %FS(1 % of full scale).

ISO 13398 also addresses vapour pressure measurement.


6. SAMPLING OF LNG 6.1. LNG QUALITY Possible contamination of LNG is a concern because it may have consequences to: the sampling system and analytical instruments, systems and equipment in which the LNG is to

be processed, systems and equipment exposed to the

vaporised LNG.

As used here, contamination is meant to include impurities at levels greater than expected and unexpected impurities. The source of contamination may be at the location where the gas is liquefied, in the transport container, and even in the system which is processing or sampling the LNG.

Examples of contaminants and potential impact include: water - exposed to LNG, water or water vapour

turns into a solid (ice) which can block sampling systems, valves and instrument taps as well as damage equipment

particulates - metal shavings, welding debris, insulation, sand, wood and cloth are typical examples of particulate material. If inert, the most common problem with particulates would be blockages and damage to equipment.

sulphur - a sampling point that utilizes copper or copper alloys may be damaged by contamination with sulphur and/or may impair the measurement of trace sulphur compounds by chemical reaction with copper.

mercury - traces of mercury may damage aluminium components by chemical reaction with the aluminium. A release of gas due to a resultant failure of the aluminium is an example of a possible consequence.

other hydrocarbons - a sampling or piping system that contains, for example, traces of LPG, may result in erroneous analysis, or otherwise in LNG out of specification. Moreover due to the limited solubility of butanes and higher paraffins in LNG, too high concentrations of these may also solidify and clog sampling systems.

inert gases - nitrogen and air may be present in both sampling systems and piping systems from, perhaps, inadequate or poor purging operations. Moreover the presence of oxygen from the air may present a safety hazard.

acid gases - CO2 exposed to LNG, turns into a solid similarly to water and may block sampling systems and damage equipment.

In establishing the components for analysis, the potential exists to ignore a contaminant because it is not normally present at levels that exceed tolerance. The consequence may be damage to equipment and, possibly, customer rejection of the LNG.

6.2. SAMPLING PRINCIPLES In order to determine the quality of the LNG it is first necessary to undertake particular operations to

25

SECOND EDITION

condition the fluid sampled from its initial condition, liquid at low temperature, to a final condition, gas at ambient temperature, without partial vaporisation or loss of product.

Sampling of LNG includes three successive operations:

taking a representative sample of LNG,

complete vaporisation, and conditioning the gaseous sample before transporting it to the analyser,

Sampling is the critical point of the LNG measurements chain: each step must always be undertaken without any modification of its representativity. It is the most complicated phase of the measurements and most of the troubles observed in determination of the energy loaded or unloaded come from the sampling system. The sampling system is not easily changeable during a transfer, so many operators duplicate the system to ensure sample collection in the event of failure of the main system.

Evolution in the LNG industry tends to normalize the sampling processes. The "spot (discontinuous) sampling system" described in the previous edition of this handbook has meanwhile become obsolete in the LNG industry. It is therefore recommended to be used only as a back-up system, in case of failure of the main device and for the limited period of its unavailability. For completeness, description of this system is given at the end of this section (paragraph 6.12).

The sampling processes used in the LNG industry are now mainly of two types. Please note that the terminology continuous/discontinuous is different from the terminology used in the previous edition. This is to reflect the terminology used in current standards, such as EN 12838):

Continuous sampling: sampling process involving continuously collecting LNG during the loading/unloading operation from the main LNG flow, which is subsequently vaporized and stored in a gasholder; gas bottles are then filled with the gas mixture in this gasholder, which is representative of the average composition of regasified transferred LNG, and connected to the analyser. The continuous sampling system is specified in ISO 8943. Discontinuous sampling: sampling process involving continuously collecting LNG from the main LNG flow; the LNG is subsequently vaporized and either analysed by chromatography at regular intervals or sampled in gas bottles at regular intervals which are then connected to the analyser.

European Standard EN 12838 specifies the tests to be carried out in order to assess the suitability of these two LNG sampling systems.

The various elements and options of a LNG sampling chain are summarised in figure 7.


6.3. SAMPLING POINT The sampling point is generally located:

on the main loading pipe, after the LNG pumps send-out, on the main unloading pipe, after the unloading arms.

LNG must be sampled on the whole flow of LNG transferred. It is preferable to install the sampling point as close as possible to the transfer point (arm flanges) so that the characteristics of LNG do not change before it is actually transferred to the purchaser due to heat input. However, generally the influence of heat input is limited, when the flow does not vary too much, in an insulated pipe.

In addition, LNG must be in a subcooled condition at the sampling point. The LNG subcooled condition can be determined by using the method proposed in ISO 8943 (annex A).

6.4. SAMPLING PROBES Two basic options can be found:

direct connection of the sampling tube on the periphery of the LNG header, LNG sampling tube protruding inside the LNG header.

This second option is to be preferred as it avoids the possible effect of a boundary layer at the surface of the main pipe, which could affect the representativity of the LNG sampling. However its design should take into account the risk of possible damages, due to flow-induced vibrations.

Sampling probes and tubes, transferring the sample flow of LNG to the vaporiser, are generally made of stainless steel.

In order to keep the LNG sampling flow in a subcooled condition, the ambient heat input should be minimised. The following lay-outs can be used:

a straight sampling tube inside the main LNG header (figure 8a), a direct connection on the main pipe (figure 8b). Remark: both above lay-outs require appropriate insulation around the sampling tube, vacuum insulation around the Pilot

sampling tube (figure 8c) completed with appropriate insulation of the top part of the probe,

LNG Pitot sampling tube cooled by a permanent internal sidestream flow initiated with natural LNG circulation going back to the LNG header (figure 8d),

Optional: ambient heat input can be further minimised by a permanent external sidestream flow to a purge line of the plant.

26

SECOND EDITION

FIGURE 7

ELEMENTS OF LNG SAMPLING CHAINS

GIIGNL - DS TML/Z - CG - 2001/10/09 27

SECOND EDITION

FIGURE 8

EXAMPLES OF SAMPLING PROBES

SECOND EDITION

6.5. PIPING ARRANGEMENT BETWEEN SAMPLING PROBE AND VAPORISER

It is important to have the liquid sample line, between the sampling probe and the LNG vaporiser, as short as possible, with a small inside diameter (4 or 6 mm for instance) and with good insulation, so that the LNG is kept in a subcooled condition until it reaches the vaporiser.

The maximum recommended length of the liquid sample line between the sampling point and the vaporizer can be calculated by the following formula:

6.6. LNG VAPORISER AND CONTROL DEVICES

6.6.1. Main devices The vaporisation of the LNG must be as complete as possible, so that the gas obtained is representative of the quality of the LNG.

The vaporiser must be designed in order to avoid fractionation, especially if the gas sample is directly taken for analysis. When the device includes a gasholder, filled during a great part of the transfer operation, it enables the components coming from vaporisation to be mixed.

Vaporisation is achieved by heat exchange, mainly in one of the following devices:

atmospheric vaporiser, but in which the heating flow may not be enough to avoid fractionation; this device should only be used to fill a gas holder, water vaporiser, the heat flow being provided by water at ambient temperature, or more often by hot water (hot water flow or water heated by an electrical resistance, for instance), steam vaporiser, steam being used to warm up a water bath where a coil of metal piping, in which LNG flows, is submerged, or steam warming the


coil up directly (Shell-tube LNG Sample Vaporiser) (see figure 9c),

electric vaporiser, the coil of piping being warmed up by Joule effect.

6.6.2. Description of vaporising devices

Water or steam vaporiser

Figure 9 shows diagrams of vaporisers using water circulation (ambient temperature or, preferably, hot water), or water warmed by steam as heating fluid, or water warmed by an electrical resistance submerged in the vessel, or direct low pressure steam.

The LNG sample flows in a tubing coil installed in the vessel and vaporises in it. The coil is usually made of stainless steel (recommended) or, sometimes, copper.

Electric vaporiser

Figure 10 shows an example of an electric vaporiser, which consists of a tubing coil in which LNG flows and vaporises, this coil being the short-circuited secondary winding of a transformer.

When the primary winding is supplied by electricity the Joule effect, which develops in the coil, produces the energy necessary to vaporise the LNG sample flow.

The tubing coil can be made of copper or stainless steel. Stainless steel is recommended and is used in the new generation of electrical vaporiser.

6.6.3. Auxiliary vaporisation control devices

Control devices must be installed to supervise the conditions of vaporisation and to protect the equipment; some of the following ones are mainly found:

on regasifted LNG outlet: pressure regulator, or gas flowrate regulator,

to control the LNG flow to be vaporised independently of pressure or flowrate in the main LNG pipe

anti-pulsation bottle, or mixing accumulator, to absorb the pressure pulses and to create a temporary retention of gas homogeneity

impingement chamber, to prevent the entrainment of possible fine droplets of liquid

flow meters pressure meters temperature detection switches,

corresponding to very high gas temperature (example: no more LNG flow) or to very low temperature (example: failure of heating device)

and associated devices, such as overpressure safety valve, electrovalve to isolate the vaporiser, etc...

29

SECOND EDITION

FIGURE 9

EXAMPLES OF WATER OR STEAM VAPORISERS

a) With water or steam circulation

SECOND EDITION

FIGURE 10

EXAMPLE OF ELECTRIC VAPORISER

on LNG inlet: check valve, to prevent a possible retro-

diffusion of vaporised components to the main LNG pipe,

restriction orifice, needle valve, to control the flow of LNG

(however, it is better to control gas phase flow in order not to create a disturbance of the state of the LNG sampled),

filter

on heating fluid (water or steam or electricity): temperature regulator, or thermostatic

control, to keep constant vaporising conditions according to the LNG flow,

thermometer and thermostats in case of failure of heating devices,

control of the electrical supply of the transformer or of the submerged resistance;

auxiliary safety devices, such as: protection of electrical supply and electrical

devices, which must be of a type designed for hazardous conditions (explosion proof, pressurised box).

6.6.4. Operating parameters Among the operating parameters, the following ones can be particularly recommended:

sampling rate greater than 1.0 m3(n)/h of regasified LNG, sample pressure at the sampling point greater than 2 bars,


vaporiser outlet temperature not below about +20C.

6.7. COMPRESSOR FOR TRANSFERRING GAS SAMPLE

According to the pressure and piping conditions, compressors may be used to transfer the gas sample:

from the vaporiser directly to the analyser,

or from the vaporiser to the gas sample holder,

or from the vaporiser or the gas sample holder to the gas sample bottle filling station.

They must be of the oil-free type and stand-by units are also necessary.

6.7.1. GAS SAMPLE HOLDER

A gas sample holder may be used to store regasified LNG during the sampling period of the LNG transfer operation; the characteristics of the gas contained after completion and mixing is representative of the characteristics of the LNG loaded or unloaded.

The volume of gas sample holder must be enough to fill the gas sample bottles and to purge the connecting lines. It is often between 0.5 and 1 m3 (figure 11aandb).

31

SECOND EDITION

Gasholders can be of two types: water-seal type, the sealing water being saturated with gas by bubbling regasified LNG through it before filling the holder, or waterless type.

6.8. GAS SAMPLE CONDITIONING Two main devices can be found:

6.8.1. Gas sample bottles They are filled:

either directly, at the outlet of the vaporiser, periodically during the transfer operation (figure 11 a), or at the outlet of gas sample line coming from the gas sample holder, possibly through a charging compressor (figure 11b).

These bottles have capacity enough for the analyses that follow, for instance 0.51.

They are generally made of stainless steel with valves at both ends, as shown on figure 11c.

Quick connectors are preferable instead of screw connectors due to frequent handlings of the bottles.

After purging of piping has been performed several bottles may be filled simultaneously or successively, according to the installed manifold. It is extremely important to take care that no air enters the bottle and that the bottle is sufficiently purged before taking the gas sample to be analysed.

6.8.2. Direct piping to gas analyser This device allows analyses to be carried out as frequently as the analyser permits, and to get rid of possible air ingress during gas sample handling.

In this case, a prefarably stainless steel pipe with a small diameter directly connects the outlet of the vaporiser to a manifold at the inlet of gas analysers installed in a laboratory of the loading/unloading facilities. A gas compressor may be required in order to make up for the pressure drop in the gas line.

FIGURE 11

EXAMPLE OF METAL GAS SAMPLE CONTAINER

a) Example of water-seal type gas sample holder

SECOND EDITION

b) Example of waterless-type gas sample holder

GIIGNL - DS TML/Z - CG - 2001/10/09 33

SECOND EDITION

FIGURE 12

a) EXAMPLES OF DISCONTINUOUS SAMPLING DEVICES (first example)

GIIGNL - DS TML/Z - CG - 2001/10/09 34

SECOND EDITION

b) EXAMPLES OF DISCONTINUOUS SAMPLING DEVICES (second example)

SECOND EDITION

c) EXAMPLES OF CONTINUOUS SAMPLING DEVICES

6.9. EXAMPLES OF GENERAL ARRANGEMENT OF SAMPLING DEVICES

As examples, figures 12a and 12b show diagrams of several possible arrangements of discontinuous sampling devices and figure 12c shows diagrams of an arrangement of continuous sampling device.

The various elements are described in the previous paragraphs.

6.10. PERFORMANCES OF THE DEVICES

The following remarks can be made about the design of such sampling devices and the choice of the elements which constitute them:

to prevent or to limit possible contamination or adsorption of the heavy components (Cs+) stainless steel is the preferred material for all parts in contact with LNG/NG flowing between the sampling point and the analyser, since it is less reactive than other materials. It is also recommended that lines, bottles, valves, etc... between the vaporising device and the analyser


are insulated. During the filling of the gasholder or during the transfer to the analyser/bottles, the gas flow should be regulated to limit outside temperature and gas velocity effects on the adsorption/desorption phenomenon;

low temperature LNG vaporisers (atmospheric or water at ambient temperature for instance) should not be used, because they are susceptible to creating fractionation, or they should only be used with a gas sample holder as shown in figure 11a or b;

the continuous and discontinuous sampling devices consisting, for the first one, of filling a gas holder during the sampling period in order to obtain a gas mixture equivalent to the average vaporised transferred LNG (figure 12c), and the second one, of direct repetitive analyses of vaporised LNG performed automatically during the sampling period (figures 12a and 12b), provide the best representativity of the LNG transferred;

it is important to ensure that the sampling and vaporising parameters remain as constant as possible during the whole sampling period, mainly LNG or gas flowrate, LNG pressure, and temperature of vaporisation.

36

SECOND EDITION

As the sampling devices have not gone trough the same tests, it is difficult to compare their operating performances. Examples of calculation of operating performances of continuous and discontinuous sampling devices are given in European Standard EN12838.

Yet, considering the distribution of the calorific values calculated from the composition determined by chromatographic analysis of regasified LNG samples, it is reasonable to consider that the gross calorific value can be evaluated with an accuracy of about 0.25% to 0.30%, including the average accuracy of analysis itself ( 0.10%) with good sampling devices complying with the previous comments (see 13.3).

6.11. SAMPLING PROCEDURE 6.11.1. Sampling period It is recommended that the LNG be sampled when the LNG transfer flowrate is sufficiently settled. It is necessary to exclude the initial period, corresponding to the starting of transfer pumps and increase of LNG flowrate, until the main pipe is completely full of LNG and biphasic or overheated LNG contained at the beginning of the operation has been eliminated, or until the full flowrate is obtained.

It is also necessary to exclude the final period when LNG flowrate decreases before stopping.

When significant changes in pressure or flowrate occur in the transfer line, it is better to suspend sampling temporarily.

6.11.2. Sampling frequency As far as filling of a gas holder is concerned, sampling is continuous during the sampling period, at a fixed flowrate; spot samples can be collected in addition during this operation, in order to control LNG quality and to monitor the transfer operation, but the corresponding analyses are not used for energy calculation.

When gas samples are taken in bottles during LNG transfer it should be done on a regular basis, depending on the characteristics of transfer lines and equipment, the organisation of operation in the plant, the duration of gas sample analysis, etc.

Example: frequency often around 1 hour, which makes about 8 samples for a normal LNG transfer duration of 12 hours, sampling starting about 2 hours after the beginning of transfer and ending about 2 hours before the end of transfer.

When regasified LNG is sent directly to the laboratory for analysis, gas sample analysis frequency depends on the available analyser (see paragraph 7.6). Example: 1 chromatographic analysis every 15 to 20 minutes during sampling period, if a chromatograph is dedicated for such an operation and if components higher than C6 are not separated.

GHGNL - DS TML/Z - CG - 2001/10/09

6.11.3. Purging

It is recommended that purging of sampling devices (probe, line, vaporiser, gas holder) and sample conditioning equipment (line, bottles,...) is carried out before any LNG or gas sample is taken into account

before starting sampling: purging of sampling probe if necessary

(double-flow, ...), circulation of LNG, vaporisation and

circulation of regasified LNG in vaporiser, pipe and either to atmosphere (small gas flowrate) or to a boil-off gas pipe of the plant, if there is no gasholder, or in gas holder with possible gas bubbling in the sealing water, and then evacuation to boil-off gas pipe;

before filling as gas bottle: connection of the bottle(s), successive operations of filling and

emptying each bottle (3 times or more) before gas sample is collected,

isolation and removal of the bottle(s).

If samples are taken periodically in gas bottles, it is better to keep the sampling system in service between operations, so that the equipment is continuously purged and ready for a new sampling with the same operating parameters.

6.11.4. Sampling parameters

It is important that the operating parameters of the sampling device (pressure, temperatures, flowrates) are kept as constant as possible throughout the sampling period, in order to obtain a smooth operation which enables representative and repeatable sampling.

6.11.5. Utilisation of gas sample bottles

Gas samples collected in bottles are:

on the one hand, directly analysed in order to determine the average composition of LNG transferred,

and, on the other hand, possibly given to the other party concerned with the transfer (purchaser or seller according to the type of gas purchase contract,) or even kept for further investigations, in case of dispute for instance, during a period defined in the contract (several weeks).

When the sampling device includes a line whereby the regasified LNG is directly piped to the gas chromatograph, an additional system may be designed to collect spot samples (gas sample bottle filling station) which are then only used for control, these samples being taken on a diversion pipe at the outlet of the vaporiser with the sampling parameters being adjusted accordingly.

37

SECOND EDITION

6.12. SPOT SAMPLING DEVICE An appropriate quantity of LNG is injected in a property purged chamber previously cooled down by LNG circulation. The chamber is thus partially filled with LNG, then isolated. The LNG sample is then brought to ambient temperature and vaporises. Thus the regasified LNG filling the whole volume of the chamber which is designed to withstand the corresponding pressure increase. Gas samples are then withdrawn from the chamber via pressure reducing valves to fill gas sampling bottles.

7. GAS ANALYSIS Just like all natural gas, regasified LNG, is analysed by gas chromatography in order to determine its composition. The reason for this is to be able to calculate the energy content based on its components. A direct energy content measurement by e.g. calorimeter would be less precise and would also not give the useful compositional information to calculate other properties (like density or Wobbe index).

Also gas chromatography can be used to determine some of its impurities like sulphur components. A different set up is often required than for its main components.

Other trace impurities, like mercury, require a different analytical technique. With most impurities sampling is critical and special precautions and sampling materials are required.

Gas chromatography is a technique, which is normally applied on a comparison basis; the quality of the analyses is generally only as good as the quality of the calibration gas.

Although it is not the aim of this handbook to describe in full detail what gas chromatography is, a general description of the important aspects involved are mentioned.

This technique is classical to determine the composition of gases and can be directly applied in case of regasified LNG.

Many methods exist in the open literature, for example in international standards series, like ISO (e.g. ISO 6974), national institutes like BS (BS 3156) or methods from institutes like ASTM (ASTM D 1945) or GPA (e.g. GPA 2261).

This chapter describes the general set up of gas chromatography systems, on line or in the laboratory, which can be used in LNG facilities, both from buyer to seller, to determine the quality of transferred LNG, their operation and data processing.


7.1. TYPE OF GAS CHROMATOGRAPH

7.1.1. General arrangement Among the various arrangements that can be found, the following ones are mentioned as examples:

1 chromatograph with 2 or 3 columns to separate selectively the components, for instance: 1 column for N2, Ci to C5 and 1 for Ce+, or 1 column for N2, Ci, C2 and CO2 and 1 for C3, C4 and Cs~, two or more chromatographs, each one specialising in the analysis of a few components; for example: 1 for Ci, 1 for C2 to C5 and CO2, and 1 for O2 and N2.

It should be noted that the significant heaviest components in LNG are generally limited to C$ and that the need for good resolution between hydrocarbons is obvious.

The primary purpose in resolving oxygen peaks is to detect contaminated samples and failures of the sampling device or leaks in the gas chromatograph tubing.

It should be realised that some methods do not resolve oxygen (and argon) from nitrogen; for low oxygen level (ppm level) a dedicated meter should be used.

In addition to the determination of the previously named components, the detection and quantitative analysis of impurities, such as sulphur components, like mercaptans COS or H2S, may also be required; they are carried out according to standardised methods (e.g. ISO 6326 under review and replaced soon by ISO 19739) and may use a specific chromatographic detector.

Figure 13a and b (see page 39) shows two arrangements of gas chromatography systems, respectively with one and several chromatographs.

7.1.2. Columns The choice of the type and stationary phase in the columns depends on the general arrangement chosen for gas chromatographic analysis and on the constituents to be analysed.

The following types of columns can be found: packed columns, consisting of tubes, generally made of stainless steel, filled with a stationary phase (packing material) which may be: a support made of solid particles

impregnated (coated) by a stationary liquid phase,

or a solid adsorbing material (not coated), such as molecular sieves;

38

SECOND EDITION

FIGURE 13

TWO GAS CHROMATOGRAPH SYSTEMS

a) with one gas chomatograph

SECOND EDITION

capillary columns, consisting of open tubes of very small diameter, in which the stationary liquid phase is directly coated on the wall. There are different types: wall coated open tubular columns (WCOT), support coated open tubular columns (SCOT) and porous layer open tubular (PLOT) columns; both in glass (fused silica) and in steel.

Packed columns typically have lengths between 0.5 and 9 m and internal diameters between 2 and 6 mm, whereas normal capillary columns generally have lengths between 5 and 100 m and internal diameters between 0.1 and 0.5 mm. Packed columns are well adapted for the qualitative and quantitative aspect of analysis. Packed columns are well adapted for the analysis of natural gas and are applied in established methods (GPA 2261 and ISO 6974 part. 3, 4 and 5).

Capillary columns (WCOT columns) generally provide the best separation of constituents, and can handle only a very low quantity of gas sample. Splitting of the sample is often applied (ratio 1:10 to 1:50).

The PLOT column has a relatively high capacity with good separation characteristics. Molsieves (for separation of e.g. He, H2, N2 and O2) are now used in both capillary and packed column applications. The latest methods (ISO 6974 part 6) are now using combinations of WCOT, PLOT and capillary molsieve columns to optimise separation characteristics in analysis of natural gas. GPA 2186 now applies a combination of both capillary (WCOT) and packed columns.

The operating temperature may be controlled either at constant temperature or using temperature programming. Temperature programming can reduce the duration of analysis, but isothermal conditions are preferable when thermal conductivity detectors are used because flow variations caused by temperature variations may create baseline drift with these detectors. In any case, the temperature of the oven containing the columns has to be well controlled.

7.1.3. Detectors Thermal conductivity detectors (TCD) are often applied in gas chromatographs used in LNG facilities, mainly because they are sensitive to all the components of natural gas and have a fairly linear response. They are kept at a constant temperature. Detection levels nowadays can go as low as 0.001 mol%.

Flame ionisation detectors (FID) can also be used. They generally have a greater sensitivity, and can detect much lower percentages than thermal conductivity detectors. This however, is not always necessary for the composition of LNG. FID's are useful for the heavy ends (Ce+ fraction, identification and quantification of hydrocarbons up to Ci2), but cannot detect inert gases (N2, O2, He, Ar, etc.).

GHGNL - DS TML/Z - CG - 2001/10/09

7.1.4. Carrier gas Helium and hydrogen are normally used as carrier gases. Their purity must be higher than 99.995%v. Helium is recommended for safety reasons.

The following examples of packing column applications can be given:

according to ISO 6974 part 4:2000 (the former ISO 6568-1981), a standard describing a simple method for the analysis of natural gas (nitrogen, carbon dioxide and hydrocarbons up to pentane and Ce+), using 2 columns whereby: column 1 = 9 m long, 4.75 mm internal

diameter, stainless steel or copper, column 2 = 0.45 m long, 4.74 mm internal

diameter, stainless steel or copper, Stationary phase of both columns: support Chromosorb PAW (porous polymer,

particles of 250 to 315 m), liquid phase = silicone oil DC200, degree of loading = 28 g per 100 g of

support, Operation of the application at a constant oven temperature of 110C 2C, The application uses helium as carrier gas and a TCD as detector. according to ISO 6974 part 3:2000, a standard describing a method for the analysis of natural gas (hydrogen, helium, nitrogen, oxygen, carbon dioxide and hydrocarbons up to octane), using two columns whereby: column 1 = 3 m long, 2 mm internal

diameter, stainless steel, packing = Porapak R with particle size between 150 and 180 um (for separation of hydrocarbons),

column 2 = 3 m long, 2 mm internal diameter, stainless steel, packing = Molsieve 13X, particle size between 150 and 200 um (for separation of He, Hz, N2 and O2),

Temperature control between 40C and 200C, with a linear programmer providing a rate of temperature change of 15C/min, The application uses both helium and argon as carrier gas and two detectors (TCD and FID). (TCD = Thermal Conductivity Detector; FID = Flame Ionisation Detector).

7.1.5. Quality of the separation of components The following parameters, which influence the quality of the separation of the components, must be taken care of when a chromatographic system is installed:

operating temperature: when temperature increases the retention time decreases, but the resolution decreases at the same time, carrier gas flow control by inlet pressure and the carrier gas flow rate should be optimised towards the highest resolution of peaks, valve switching, a problem can create contamination or cutting off of components in the chromatogram.

40

SECOND EDITION

7.2. INTEGRATOR AND DATA PROCESSING

7.2.1. Integrator system An integrator is connected to one or more gas chromatographs in order to determine the retention time, the area of each peak obtained by the detector and to print the results, or send them to an auxiliary data processing system.

Integrators can also, according to the sophistication of their software programs, take into account the calibration of chromatographs, detect and identify the peaks, calculate the percentage of each component in the gas mixture and even control the operation of the chromatographs.

The components of a gas are identified by their retention time. Integration of a peak is sometimes difficult when peaks are not fully separated. It is recommended to use equipment where the actual integration line can be verified.

Another function of the integrator may be to generate alarm.

7.2.2. Data processing Additional processing of the chromatographic data must be done before the analysis is reported. Both integration function and data processing are nowadays carried out in a computer (standard PC). Also additional calculations of physical properties are normally done in this computer (heating value, density, etc., according to ISO 6976-1995).

In calculating the end result (expressed in mol% per component) the computer normally reports both raw (not normalised) data and normalised data. The raw (not normalised) data are the result of the calculation of the sample with the calibration table (resulting from the analysis of the calibration gas). The total of the raw results should normally be between 98 and 102 mol%. If the total (of not normalised data) is outside this window, it normally means that something in the analysis is wrong. The calibration may not be valid or there is a problem in the analysis (injection, column switching, peaks missing etc.). Careful examination of the equipment is than required. If the accuracy is found, a re-calibration should be carried out.

7.3. CALIBRATION The application of gas chromatography in natural gas is a technique that requires calibration. Calibration is carried out with standard gases of a similar composition as the samples (often-single point). Validation of the GC equipment can be done according to ISO 10723.

7.3.1. Calibration procedure After the initial set up of the gas chromatograph and the installation of the integrator/computer software, the gas chromatograph has to be calibrated. This operation consists of successive analyses ( 4 - 6 for


instance) of a certified calibration gas (see paragraph 7.3.2.) in order to determine:

the retention time of each component in the respective columns: identifying the components, the response factors (arithmetic average of factors calculated after several analyses, 10 or more) or the parameters corresponding to reference components (see paragraph 7.4.): quantifying the components gas sample.

T h e s e p a r a m e t e r s a r e s t o r e d i n t h e integrator/computer system.

The quality of calibration gas is determining the quality of all the related measurements. A traceable gas, preferably of Primary Reference Material (PRM), needs to be used from a reliable source (see ISO/DIS 13275).

In addition to these determinations, a repeatability test is recommended to evaluate the random accuracy of the system. Several analyses are carried out successively (at least 20) in the same conditions on the same gas. The distribution of the results (percentages of each component or calculated heating value) is then examined and the standard deviation calculated.

If this parameter is too high, for instance more than 11 kJ/m3(n) on calculated gross calorific value around 44.4 MJ/m3(n), it must be interpreted as an anomaly of the conditioning of the system, or something wrong with the integration parameters, or a failure of an element of the system; and the system then needs to be repaired or its operating conditions need to be modified.

7.3.2. Calibration gas/working standard The gas mixture used as a calibration gas must include all the components found in the regasified LNG to be analysed, within close percentages.

The preparation of such a gas mixture must be undertaken with great accuracy. It is recommended that gravimetric standard methods are used. The preparation of the calibration gas may be performed according to standards such as ISO/DIS 13275.

The preferred quality of the calibration gas is primary reference material (PRM). This gas is expensive, and careful use is advised in view of the relatively high cost (about USD 6,000 for a cylinder of 5 litre at 50 bar).

It is useful to have a working standard, which is analysed against the calibration standard. This working standard needs to be stable and is used to check the performance of the analyser system on a regular basis (e.g. weekly, or prior to each cargo).

Working standards can be made by filling a bottle with gas taken on the LNG facilities (e.g. 50 litres cylinder filled at 50 bar). Of course also lower grade reference gases can be purchased for this purpose.

If the analysis of the working standard indicates a need for re-calibration, the PRM should be used to

41

SECOND EDITION

recalibrate the system (after proper analyses of the cause of this). After each change on the hardware or software, re-calibration should take place. Typically re-calibration should not be needed more than once per 3 months or so.

7.4. QUANTITATIVE ANALYSIS The basic data used for quantitative analysis are the areas of the peaks corresponding to the various components. The most common method is described below according to the ISO. 6974 part 2. This methodology can only be used if the response

of the detector is linearly changing with concentration, or if the concentration of the calibration gas is very close to that of the component being analysed. After determining the concentration of all components, the total is calculated (raw result, nor normalised). Ideally the total is close to 100%. Small deviations may occur from random inaccuracies (e.g. injection volume by variation due to atmospheric pressure). Normally the sum is not exactly 100%. Normalisation (to make the total exactly 100%) is a way to deal with this problem. The assumption with normalisation is that the problem is the same for all components. All components are divided by the raw total and multiplied by 100, given a total (normalised) result of 100%. The window in which this allowed is 98 - 102 percent for the not normalised total. Higher or lower levels are an indication of a problem.

It should be noted that there might also be other problems, which are related to individual components. An integration problem can give a false peak area; integration of peaks should be the same as for the calibration gas. This is why the calibration gas should have about the same composition as the sample to be analysed.

GHGNL - DS TML/Z - CG - 2001/10/09

7.5. ENVIRONMENT FOR A GAS CHROMATOGRAPHIC SYSTEM

The practical requirements for the installation of such a system are the same as those required for any high-accuracy analysis device and mainly involve:

installation in a closed and temperate (not necessarily air-conditioned) building, sheltered from the sun, heating sources or draughts, or outside for process chromatogrpahs typically between -10C and +50C, appropriate and constant temperature of calibration gases and sample (injected mass/constant volume), permanent and secured electrical supply, without interferences, shrouding and earth connections of the electrical connections between the chromatograph and the integrator.

7.6. ANALYSIS OF REGASIFIED LNG AND RETAINED SAMPLES

During a normal LNG transfer operation (for instance, transfer duration of 12 hours, sampling period duration of 8 hours), the following analysis procedures can be carried out:

in the case of direct connection between vaporiser and chromatograph: the analyses can be made successively during the whole sampling period, with a frequency equal to the duration of each analysis by the chromatographic system, Example: with 1 analysis every 20 minutes, 24 analyses are available during the sampling period. For the new generation of chromatographs, the duration of each analysis is reduced to 5 minutes; in case of periodic filling of sampling bottles: 1 or more (often 2) analyses can be carried out successively on each gas sampling bottle, with a comparison of results, and possible additional analysis or new filling of sampling bottles in case of important thresholds, and then calculation of the arithmetic average of the percentages of the components determined by the analyses considered or the determination of the average composition of the sample. Example: 2 analyses on each bottle filled every hour and calculation of the average, so 8 average analyses during the sampling period; in the case of filling of a gas holder (ISO 8943): at the end of the sampling period, 3 bottles are filled, 1 for each of both parties (Seller, Buyer) and 1 kept for further investigations (e.g. a third party, in case of a dispute); 1 or more (generally 2) analyses can then be carried out on the same sample and retained, if there is no significant threshold.

42

SECOND EDITION

Then, the average composition is calculated (arithmetic average of the percentage of the various components) with all of the results taken into account (direct result of analysis or mean of several analyses on the same sample). This value can be used in calculations of any additional parameters (LNG density, calorific value) for instance by the data processing unit (see sections 8 and 9). It is common practice that the analysis resulting from the sample collected by continuous sampling is only applied to the quantity of heat calculation.

7.7. INACCURACY OF GAS ANALYSIS

The inaccuracy is related to the level of a component in the sample. Provided the system is chosen, installed and operated with care and in compliance with the method and recommendations, calibration with a high quality calibration gas (PRM), the total inaccurancy in the computed gross calorific value can be as low as 0.1% (relative) for the main components (level 10-100%).

For lower concentration levels, the inaccuracy (in terms of repeatability) is 1% (relative) for a level of 1-10% and 10% (relative) for a level of 0.1-1%. On a level between 0.01 and 0.1 percent, the repeatability is around 30% relative.

7.8. IMPURITIES The impurities in regasified LNG which sometimes are specified in LNG contracts are carbon dioxide, sulphur components (hydrogen sulphide, carbonyl sulphide and mercaptans) and mercury. The sulphur and mercury components are normally trace impurities at respectively 0-1 mg/m3 or, in case of mercury as low as 5 ng/m3. Sampling cannot be done in normal cylinders for these trace impurities since they are chemically reactive and will be absorbed by the wall of the sample cylinder.

For sulphur impurities, special materials to minimise absorption are commercially available (silicosteel).

The determination of trace impurities requires a special approach. This can hardly be underestimated; set-up, operation and maintenance are an area for specialists. All aspects are critical: sampling, calibration and analysis. Validation and verification of results is strongly advised before using the results of the analyses for commercial purposes.

7.8.1. Carbon dioxide The carbon dioxide content is normally determined by GC analysis; the specification limit is often around 0.01 mol%. The GC is capable of analysing down to 0.01% or even lower.

7.8.2. Sulphur Sulphur can be specified as total sulphur and/or as specific sulphur containing components. Hydrogen


sulphide (H2S), carbonyl sulphide (COS) and mercaptans (RSH, where R is an alkyl group; e.g. methylmercaptan, CH3SH or ethylmercaptan, C2H5SH) are mentioned. The level of these impurities is normally on a 0-25 mg/m3(n) level.

Sampling for trace sulphur components is not so easy; special precautions are needed in order to avoid absorption of sulphur components to the wall of the sampling system devices. Sampling in bottles is preferably made according to the standard method described in the ISO 10715. The incremental sampling is very delicate for sulphur components. The interior face of cylinders must be made out of a material which doesn't react with sulphur components.

The sample cannot be conserved for more than 8 days. The material "silicosteel" is suitable for this application but is very expensive.

7.8.2.1 Total sulphur Total sulphur can be determined by combustion techniques, like the ISO 4260 (Wickbold combustion), where all sulphur is converted into SO2 which is trapped and quantified. Instrumental techniques like microcoulometry (ASTM D 3146), pyrolyses/chemoluminescence, or hydrogenolysis/ rateometric colorimetry (ASTM D 4045) are sometimes applied. ISO 4260, which is often specified in contracts, is cost effective but not the safest to apply. Many accidents have already occured due to the rather violent combustion of gas in an oxygen/hydrogen flame.

Therefore the use of the instrumental techniques is recommended.

7.8.2.2 Sulphur components In order to determine the sulphur components separately in a gas, these components must be separated first. A gas chromatograph can be used for this in combination wi

lng custody transfer handbook

Documents

sampling of lng

lng volume

volume of lng

transfer of lng

lng energy

lng vaporiser

lng quality

level measurement