You are on page 1of 4

,

es, the new controlled s a control nd its ~ e output r by efficiency

nts used J-T ants are rove the e has >d to further to provide iciency and indicate that achieved ' contradicts iis article upposed t not to uith that of the system. le operation id expanders valve maximum I and by so ~amic ,t for the

he growth in LNG trade has continued at a significant level over the last decade. The market for LNG, however, is tightening u p and thus potential investors are looking for lower capital and operating costs for new facilities with increasing intensity. This has led to greater interest in alternative technologies for pre-treatment, liquefaction and storage. The Base Load LNG industry has up to recently been dominated by one process technology, the Air Products & Chemicals Inc.(APCI) precooled Mixed Refrigerant Process (known as MCR). This process currently has an unrivalled track record of over 50 plants in operation, using wound coil cryogenic heat exchangers. For many years Technip, in combination with L'Air Liquide and others also offered a technically sound and commercially attractive wound coil heat exchanger Mixed Refrigerant technology, the Tealarc Process, but did not succeed in penetrating the base load LNG plant market. Recently, the situation has changed with the established technology being aggressively challenged by processes using plate fin heat exchangers (PFHEs). These include the Phillips Optimised Cascade for the Atlantic LNG project at Point Fortin in Trinidad. Also, the Pritchard - PRlCO mixed refrigerant process has been operating at Skikda for many years and modern designs offer a further alternative process for the Base Load LNG market. In addition, the Peak Shave LNG market is seeing a revival through a significant potential for replacement of older, less efficient plants operating at peak shaving facilities around the world. It has been

Liquefaction
b y Terry Lavin, L i n n h o f f M a r c h L i m i t e d ; Harry lsalski a n d Philip Hindley, Tractebel I n d u s t r y E n g i n e e r i n g S.A.; a n d David Linnett, C o n s u l t a n t i n C r y o g e n i c s & G a s Processing (UK). This article is a n updated, edited version o f a paper presented a1 the Conference organized by IIR Limited : "Market Opportunities for LNG Projecls", London, October 14-15, 1997.

satisfied by a much larger number of suppliers of technology including APCI, Pritchard, Linde, L'Air Liquide, Gaz de France, Snamprogetti, CBI, BOC-and. Costain. The Peak Shave LNG market \ has probably seen the widest range of liquefaction technologies applied, including a variety of mixed refrigerant processes comprising one to five stages of refrigeration, classic cascade refrigeration processes, inert gas and natural gas

men1 of ;ses by using levenlh uefied ;tilute of Gas


ulled Turbine

Terry Lavin is a Principal Consultant at Linnhoff March. His speciallies include cryogenics, gas processing a n d distillation. I n previous positions w i t h the BOC Group a n d British Gas, he w a s responsible for design a n d commissioning of several LNG peak shave plants. Recent w o r k includes detailed thermodynamic analysis of LNG base load designs. Harry lsalski has over 2 7 years experience in the conceptual design,
delailed engineering a n d commissioning of all types of gas processing, cryogenic a n d LNG facilities. This includes all design, engineering a n d startup of t w o LNG peak shave plants a n d process optimisation studies of mid-sized a n d base load LNG plants. H e is Technical Director at Traclebel lndustry Engineering, OK Branch Office.

expansion cycles. Some of these processes are also suitable for mid-sized and base load applications. As the range of potential technology increases, the operating companies have a much more difficult choice to make when evaluating the various schemes and making the optimum choice of process route for a specific application. This article describes various methods of process review and the perceptions that influence process selection. It discusses simple efficiency benchmarking and recommends a more structured way forward which includes Exergy Analysis as a technique that gives a clearer picture in any comparison methodology.

Criteria for Process Evaluation


Natural gas liquefaction consumes a large amount of refrigeration energy to bring the gas from elevated pressure and ambient temperature to a liquid at near atmospheric pressure and minus 162'C. The way the refrigeration is applied has a dramatic effect on the capital and operating costs of the liquefaction plant. Currently, common methods of evaluating LNG liquefaction processes use: specific power, kwhftonne of LNG specific capital cost. $ or Sftonne of annual capacity past experience in these plants The above methods are useful as an initial assessment, but with the number of process options available today and the pressure on establishing low life cycle costs, a more rigorous approach is

teering, '. Lld., U.K. rolled Cryogenic I d Pumps,


.E.: "Power

nic Proceedings lwer echnology "Method pansion in ; Palenl


I.. :

David Linnett has been a n independent consullant in cryogenics a n d gas processing since 1987 a n d w a s previously Process Engineering Manager w i t h BOC Process Plants. He has over 30 years experience in LNG plant technology, including major assignments i n many LNG projects of all sizes from peak shave l o base load. Philip Hindley graduated w i t h a n honours degree in Chemical
Engineering from Loughborough University, a n d has 4 years o f g a s processinq He is a Process Enqineer w i t h Tractebel - desiqn - experience. . lndustry Engineering SA.

jnics: An Y.A. and McGraw-Hill

News in brlq
LNG
M o r e from EiIGT

-I
necessary. Other qualitative aspects used for comparing processes include: process complexity number and type o f machines driving the process type of heat exchangers used (wound tube or plate fin) what ambient cooling methods should be adopted what design margins should be applied I n any refrigeration process, the refrigerant compressor dominates the capital and operating costs. When the plants are large, the economies of scale help in increasing the machinery efficiency considerably. For the larger refrigerant compressors, polytropic efficiencies over 86% can now be reached. This means that for a given power input, more refrigeration can be provided and more LNG can be made. If gas is cheap, the view could be taken that process efficiency should not be a critical issue. This may hold true for the gas turbine drivers in the liquefaction plants because it may be that the more efficient driver may be more expensive to buy. When coupled u p together, the compressor and driver is a high proportion of the LNG plant liquefier capital cost. A lower cost, more robust gas turbine driver may be the most cost effective solution even though it may have a lower efficiency in terms of conversion of natural gas into compressor shaft energy. However, the process efficiency will always matter once a given compressor and driver has been chosen. This is because the designer has chosen a machine for a given specific cost per kW of power supplied to the drive shaft. Once that energy is at the shaft, it must be put to the best use possible in order to maximise the production of LNG. This will have the effect of driving down the specific cost of the plant per tonne of LNG. Therefore, the design of the liquefaction section itself needs very careful attention. In selecting a liquefaction process, the designer has many parameters that he can change to optimise the process, including: type of cycle; mixed refrigerant, pure component, expander, cascade, precooled or a combination of these which components to use in the refrigerant number and temperatures of stages of refrigeration pressures at the suction and discharge of the main compressor ambient cooling media pressure drops in equipment Given previous experience, the choices can be narrowed down considerably. However, it is wise to analyse the options. particularly when the nature of the LNG market, and the locations and sizes of plants are changing significantly. A more rigorous approach for the highly competitive environment of the 1990's is warranted.

O SlGTTO (.the Society t3f Gas TE ~ n k e.r... ; ,~ produced and Terminal Operators) nas
~ ~

further reports. They include : '., $ "Inspection Guidelines for Ships ' --.: Carrying Liquefied Gases in Bulk -.' . Produced jointly with OCIMF, published i i n 1998. Cost G 15. .,. . ;. i.5.1 . , "An Introduction to the C)esign ar Maintenance of Cargo Syste m Pres: Relief Valves o n Board Gas Carriers J : . Cost E 20. To be publishw a n September 1998. These two publication:j are avai from: Witherby & Co. Ud. (Lonoon). 171 251 5341. Fax.44 171 2 5 1 Others being produced include .. "A Glossary of Terms Used i n .. ! -;J,! Liquefied Gas Shipping". To be .: , published in September 1998. ,..? 4 "A Risk Based Approach for the . Evaluation of Fire Fighting Equipmaent 1 o n Liquefied Gas Jetties". This .: I .--A incorporates a report p r o d u c ~ u 01, ' behalf of SlGTTO by AEA Technology,' ; and is due to be published later in 1998.. j For more information contact Roger ' Roue (SIGTTO) : Tel. 4 4 171 628 1 124. Fax. 44 171 299 1818. . . ;. !

;:.: ;I

3::
',I

".

Thermodynamic Approaches
The simplest thermodynamic approach i s to consider the law of conservation of available energy. This has often been used in natural gas processing to determine a given process' yield of heating value. Putting this simply, the ratio of calories out to calories in (the caloric ratio) gives a benchmark of how much energy has been lost in the process through compression power and fuel gas to generate heat and steam. In a base load LNG plant, this figure is often used as a yardstick for a given plant performance. When applying this to an LNG plant, values in the range of 89% - 94% can be found. This always looks a reasonably high value and can give false comfort to management. Plant improvements may not present an attractive improvement to an already high number. A second benchmark sometimes used to compare process routes is an overall entropy based efficiency (E) term, E = Wm/Wa, where W m is the minimum work required to liquefy the gas, and Wa is the actual work input to the process. A similar ratio was used by Bronfenbrenner (1). The use of this equation is a good starting point for process comparison because it is quick and needs few calculations; as long as the comparison is on the same basis, it provides the user

. .

'

i,

LNG 1 2 ProcCeu~~ ~gs O The Proceedings o f the 12th . ,, International Conference o n Liquefied
,

Natural Gas (LNG 12) are o n sale either as a two-volume set o f books, or o n C D ROM, from : Colleen Taylor Sen, Institute of Gas Technology. 1700 S. Mount Prospect Road. Des Plaines. IL 601 18, Fax. 1 847 768 0842. Tel. 1 847 768 0512. E-mail sen@igt.org

.
I
\

EU market
0 Energy liberalisation is set t o radically change the shape of European electricity and gas markets, as major EU I power companies and new entrants fight j for a slice of a USS125 billion market, according to a new report from Reuters Business Insight (RBI).The passing o f' ! the European Commission's electricity 1 directive in February 1997 means thaf'f'r' ' by 1999, at least 60% of the EU ., electricity market will be OF competition.

Table 1 Overall Comparison o f LNG Processes (Cold Section Only) Process Route Approx. Caloric Ratio x 100% Cold Section Thermodynamic Efficiency (%)
55-61 60 - 6 5 63 - 66 60 - 63 35 - 45

Specific Power kW/kg LNG


0.34 - 0.36 0.30 - 0.32 0.29 - 0.3 1 0.32 - 0.34 0.45 - 0.55

I-Stage MRC 5-Stage MRC 5-Stage MRC (fully sub-cooled LNG) C3 Precooled MCR 3 turbine N2 cycle

93.0 93.7 93.9 93.4 90.5

30

cade, pref these e i n the


IS

Ftgue 1: Slngle Stage MLxed Refriserent Cycly

o f stages of

FEED GAS MAIN EXCHANGER

and discharge

nent the choices derably. the options. of the LNG d sizes o f ntly. A more rhly he 1990's is

CONDENSER

LNG
CYCLE COMPRESSOR

:es : approach is vation of ten been lg to ield o f heating e ratio of 5 caloric row much rocess 3nd fuel gas I n a base often used as >erformance. 3 plant, 34% can be asonably comfort to ments may rovement to
?times used an overall term. E = inimum work i d Wa is the ess. A nfenbrenner is a good nparison ;few omparison is s the user

n Cycle wi
NITROGENCYCLE

HEAVY HYDR

NZ COMPRESSOR

EXPANDER 1

EXPANDER 2

EXPANDER 3

with an initial guide to the efficiency o f each route. However, it can only look at the overall picture; it does not identify where the major inefficiencies are. This can be illustrated by comparing several LNG liquefaction processes as shown i n Table 1. It is important to note that this article compares only the cryogenic units o f the LNG plant. The above two methods would most likely be used by an assessor of process routes during preliminary appraisal of the economics o f the project and for initial comparison between potential suppliers of technology. The designer, however, needs something more accurate to help h i m search for the optimum route for liquefaction. It is imperative that the designer can identify the location of the major inefficiencies and see what effect design changes have on the system. Using the principles of Conservation of Energy (First Law), one can use heat exchanger composite curves to some good effect. However, it does not always show the most effective route, as illustrated i n a paper by Lavin & lsalski (2), which compared the composite heat load curves for the cryogenic sections of the single stage and a 5-stage mixed refrigerant cycle for the liquefaction of natural gas. When comparing the heating and cooling curves, at first sight, the 5stage process looked inefficient when compared to the single stage process because the temperature differences were larger, implying greater losses. Therefore, a cursory view may lead to the conclusion that the single stage process is more efficient. The true picture is very different i n that the 5-stage cycle consumes some 20% less power to liquefy the same amount of LNG. The reason is that the flow of refrigerant i n the 5-stage process is much lower at the cold end than at the warm end, making the irreversible losses smaller at the cold end. Also, the final stage refrigerant only comprises the lightest components, nitrogen and methane. which are the most effective at the cold end. In the single stage process. the flow and composition of refrigerant are the same at the warm end as at the cold end. This greatly increases the heat load and reduces both efficiency and operating flexibility. This proves that it is necessary to go into more detailed analysis to illustrate the merits of liquefaction processes. Exergy Analysis provides an excellent tool to achieve this.

Exergy Analysis
Exergy o r "availability" can be defined as the maximum amount of reversible work that can result from the interaction of a closed system with its surroundings. Exergy Analysis is a well-known technique for analysing irreversible losses i n a separation process. By combining exergy

analysis with flowsheet optimisation, one can effectively compare various processes and identify the largest loss locations and or this comparison, a simulation was constructed for each of the schemes so that it look for ways to minimise them. The was able to adjust compositions of refrigerants in order to get to a workable solution. concepts have been adapted during recent Exergy analysis was then carried out to identify the areas where the losses were greatest work to allow easy interaction with HYSYS, and adjustments were made to the process to gain a better solution. The Table uses which is being used for process data from simulations on the same basis, optimised to the same level of accuracy to optimisation. Exergy analysis has been make the comparison meaningful. The calculations assume process streams are cooled used to provide data for this article by to 33C in compressor aftercoolers, etc. Each of these has assumed a 50 barg and 40'C applying it to the process simulations clean natural gas feed containing about 5% by volume of ethane and about 4% of developed for the following processes: propane and heavier. The same compressor isentropic efficiency was used for all cases. a single stage mixed refrigerant The minimum exchanger temperature approaches have also been normalised to about process - Figure 1 2.5"C pinch. The expansion turbines have been assumed to achieve 85% isentropic a five stage mixed refrigerant process efficiency. The mixed refrigerant processes have used a mixture of nitrogen, methane, Figure 2 ethane, propane, butanes and pentanes. If ethylene were used instead of ethane in the a three-turbine nitrogen gas mixed refrigerant process options, a 5 - 10 % lower power would be expected. compression/expansion cycle - Figure 3 The nitrogen cycle is assumed to use pure nitrogen as the circulating fluid. The Note: in practice, the small expansion turbines take flows of the returning heavy Table 2 - Energy Analysis Comparison for 3 Process Routes the bulk of the exergy loss hydrocarbons stream would be as expected. However, the re-heated in the main Plant Items I-Stage MRC 5 - S t a g e MRC 3-Turblne Cycle majority will be recovered exchangers. This is not shown 72 78* 51 Heat Exchangers in the compression section on the diagram. 26 15 Valves 16 through the brakes linked These are compared in 33** 0 0 Turbines to the expander. There are Table 2. Overall irreversible 2 7 negligible Mixers high losses in the warm losses are tabulated as well end heat exchangers, as individual item losses. Total 100 100 100 which could easily be Total Losses 1.58 1.16 1.9 with turbine brake reduced by precooling the (kW/kg mole LNG) 6.0 without brake Heat Exchanger Design nitrogen with say, Some of the recent advances 'As an example -exchangers broken down into 5 sections yield the following loss propane, yielding a in the manufacture of brazed breakdown from warm end to cold end: 31; 18; 13; 14; 2% or a total of 78% relatively efficient process, aluminium plate fin heat '*This example quoted for turbine inclusive of brake recovery albeit with five machines. exchangers (PFHEs) have Figure 3 shows a nitrogen dispelled worries about their cycle with three turbines in parallel, which was only one of the typical starting options use in base load LNG plants. This has considered. Exergy analysis showed that a more efficient configuration is two turbines been manifested in the award of the in series with intermediate reheat, plus a third turbine in parallel. These options have all Trinidad project to a process that uses been considered and used in appropriate applications for many years : however, these extensively. Moreover. there are improved design techniques, including detailed exergy analysis, now enable much better many smaller LNG plants that have comparison of cycle options. Cycle selection and optimisation can then be closely operated using PFHEs for many decades. tailored to specific applications. They are ideal for use in complex arrangements where small temperature differences are required. Their design are significant losses in the heat Exergy analysis has shown that the heat naturally allows excellent heat integration exchangers. It is easy to see where the exchangers have the largest irreversible of multiple heating and cooling streams losses are in the 5stage process. losses in the process and is used to giving excellent process efficiencies. The However, for a single stage process the identify where they are. The use of exergy air separation industry has used PFHEs as analysis does not show where in that analysis as part of the optimisation standard for many years and has exchanger the loss is the greatest. process, together with major advances in maximised these benefits, continuing to Therefore, it is necessary that the the design and manufacture of PFHEs. expand the frontiers of what is a mature exchanger is broken down into several enables high efficiency processes to be technology. Large multi-exchanger segments, to allow segment by segment tailored to meet the specific requirements assemblies approaching the size needed analysis of the exchanger. of individual applications. for a base load LNG plants are already in successful operation in both the air Conclusions References separation and LNG industries. The varieties of LNG processes available I . Bron/enbrenner,J.C., "The A i r In using PHFEs in the liquefaction today require careful selection to ensure Products Propane Precooled/Mixed process, their multi-stream capability that the correct process is chosen for a Refrigeranl LNG Process", The LNG makes them more difficult to assess. given application. This selection will Jo~~rna Nou/Dec, l, 1996. Rigorous analysis using exergy can make depend on the plant size and location. 2. Lauin, J.T. & Isalski, W.H.. it much easier to appreciate comparisons "lmprouing the E//iciency o/ Liquefaclion However, all but the very small plants will between several process options. As is benefit from a rigorous analysis to help Processes lhrough the Applicalion of shown in Table 2, exergy analysis can Exergy Analysis", IIR Conlerence on make the right choice. highlight losses in any process item Condensing mixed refrigerant cycles LNG. London, 15th Ocl 1997. including heat exchangers, valves, and show a considerably higher overall mixing devices as well as identifying efficiency than gas expansion cycles, even Acknowledgements pressure drop losses. when these use three turbines. MultiOur lhanks are due to [he management This analysis can lead to process stage mixed refrigerant processes are of Traclebel Industry Engineering S.A. enhancements, having identified its effect better than single stage ones from the a n d Linnhoff March L l d for permission relative to the whole process. For all the efficiency point of view. to publish this arlicle. processes, we can see from Table 2 there

I Comparison basis for Exergy Analysis of Process Schemes


F

You might also like