The 14th International Conference and Exhibition on Liquefied Natural Gas (LNG 14) took place in Doha, Qatar, March 21-24. It broke records for the number of delegates and the size of the exhibition. Part 1 of this report covers LNG export projects, with an emphasis on liquefaction technology, and safety - topics that were prominent at LNG 14. Part 2 (in our May/June issue) will deal with LNG trade and markets, LNG reception terminals, storage, LNG shipping and transfer.
 

The organization of LNG 14 by Qatar Petroleum was first class, and the facilities at the Sheraton Hotel's conference centre were excellent. There were 2245 registered delegates - over 300 more than for LNG 13 in Korea, and a record number for the series of 3-yearly international LNG Conferences sponsored by the International Gas Union (IGU), International Institute of Refrigeration (IIR) and Gas Technology Institute (GTI). The total number of participants (including also registered exhibitors, accompanying persons, guests, etc.) was more than 3000, from 90 different countries. One reason for the high number of delegates is apparent from an analysis of the list of delegates - the "home crowd" factor: about 24% (over 500) were from Qatar. This compares with only 11% from the host country, Korea, for LNG 13. Japan provided the second highest number of delegates (about 15%), not surprisingly down on the proportion for LNG 13, closer to home.

At the closing ceremony, after thanks were given to the National Organizing Committee, the Steering and Programme Committee, and all those who had contributed to the success of LNG 14, there was an official handover to Juan Pons, Secretary General of the Spanish Gas Association, Sedigas, and Chairman of the LNG 15 National Organizing Committee. LNG 15 is to be held in 2007 in Barcelona, Spain. The Chairman of the Steering Committee will be Michael Dugan of GTI. The Chairman of the Programme Committee will once again be Nirmal Chatterjee of Air Products, representing IIR. It was also announced at the Closing Ceremony that Algiers, Algeria, will be the venue of LNG 16.

The Opening Ceremony

In opening the Conference, His Highness the Emir Sheikh Hamad Bin Khalifa Al-Thani referred to the importance of the optimum utilisation of Qatar's gas reserves (among the largest in the world), and LNG exports in particular, to the development of the country's economy. In the first keynote speech, Qatar's Minister of Energy and Industry, H.E. Abdullah Bin Hamad Al Attiyah, outlined the growth of LNG consumption worldwide and the ambitious plans for further development of Qatar's LNG and pipeline gas exports, and natural gas fuelled GTL and petrochemical projects. Qatar aims to be the largest LNG exporter in the world, with 60 mta by 2010, compared with 18 mta today.

A keynote speech by H.E. Dr. Chakib Khelil, Algerian Minister of Energy and Mines, included a strong warning about the continued need for long-term LNG sales contracts to underpin the heavy investment in new LNG supply projects. A move towards short-term LNG trade carried the risks of price volatility and

of under-investment in infrastucture to provide security of supply. He also pointed out that care was needed in deregulating gas markets: it had not necessarily been successful in lowering gas prices to benefit the consumer - in Europe, for example.

A keynote speech by Mr. Claude Manil, Executive Director of the International Energy, predicted a six-fold growth in LNG trade over the next 30 years. However, the LNG industry would not have a "free ride" and he pointed out many uncertainties. These included: the ability to fund the huge increase in numbers of LNG facilities; prices (too high and demand will fall, too low and financing will be difficult); the need to continue to develop technology to reduce costs; potential competition from nuclear energy; security; the need for Government control over liberalisation; and the "NIMBY" attitudes which inhibited siting of new LNG facilities.

The Conference Programme

As before there were seven paper sessions, a poster session, and four workshop sessions, with paper and workshop sessions running in parallel.

There were 50% more presented papers (excluding posters) on LNG production than the total number on LNG reception, storage, shipping and marine transfer. This is understandable, as there has been much recent work on liquefaction technology and plant design. With LNG 15 being held in an LNG importing country in Europe, and with the high level of current interest in new LNG import terminals in the Atlantic Basin, it is hoped that LNG 15 will be able to accommodate a few more relevant papers about LNG reception terminal projects and technology, and shipping. There is certainly a case for extending the conference in future to a fifth day, to allow more time for many delegates to attend Conference seessions as well as hold business meetings, network, and attend the exhibition.

To avoid repetition, and for convenience of the reader, we structure this report according to topics, rather than following the order of the programme

Existing Liquefaction Technologies

This section deals with liquefaction technologies in export terminals under construction, and operation of existing plants.

A paper by Andrew Jamieson of Nigeria LNG, in conjunction with Kellogg, Brown and Root (KBR), presented an overview of LNG exports from Nigeria, including a description of the liquefaction process in the NLNG Plus train 4 and 5 expansion, now under construction with train 4 due to be operational in mid-2005. This can be regarded as Shell's "current generation design" of plant using the Air Products (AP) propane pre-cooled mixed refrigerant process (C3-MR). It has air cooling (compared with trains 1-3 which are water cooled), and uses two GE Frame 7 air turbine drivers: one drives a propane refrigerant compressor, and one drives in tandem an axial compressor for the low-pressure MR and a two-stage centrifugal compressor for the high pressure MR. Production is enhanced by the use of starter/helper motors supplying additional shaft power to the refrigerant compressors, and the use of hydraulic turbines (liquid turbo-expanders) for pressure reduction in the LNG and MR streams. Trains 1-3 used one Frame 6 and one Frame 7. The capacity of trains 4 and 5 is 4.1 mta. A similar process is used in the Australia NWS train 4 expansion (under construction).

A variant on this process was described in a paper by Don Hill of KBR, in conjuction with SEGAS and Union Fenosa, about the SEGAS (50% Union Fenosa, 50% ENI) LNG export project in Egypt, now under construction. The paper goes into some detail on the reasons why air cooling was selected over water cooling. In order to allow the power split between propane and MR refigeration to be optimised while fully using the power available from the two drivers, AP's Split MR refrigerant compressor configuration is used - the first Frame 7 EA turbine drives a high pressure MR compressor as well as the propane compressor, while the other drives the low pressure and medium pressure MR compressors. This arrangement is also used in the RasGas trains 3 and 4 expansion. With the help of feed gas compression, and other optimisation, the train capacity is increased to 5 mta. The authors claim that this will be the fastest LNG export project between initial concept and first LNG production at the end of 2004 - a major achievement for an operating company with no previoius experience in LNG projects, but owing something to the fact that the gas is bought from the pipeline system in Egypt: no gas exploration and production had to be undertaken by SEGAS.

The liquefaction process that was the first to challenge the C3-MR process in recent years is Phillips' optimized cascade process, as used in the Atlantic LNG Trinidad plants. A paper presented by Phil Hunter of Bechtel, in conjunction with Atlantic LNG and ConocoPhillips, described Atlantic LNG's train 4, claimed at 5.2 mta io be the largest liquefaction train under construction. The larger size compared with trains 1-3 (3.3 mta) was achieved by increasing the number of compressors on the refrigerant cycles in the cascade process. Aero-derivative gas turbines were seriously considered for this project, but the decision was made to stay with Frame 5D turbines, but at a greater number than required for trains 2 and 3. A 10% increase in efficiency, and a 19% reduction in CO2 emissions, will be achieved.

The first plant to use aero-derivative gas turbine drivers is claimed to be the Darwin LNG plant, presented in a paper by Doug Yates of ConocoPhillips. This ConocoPhillips project selected a train size of 3 mta because it would be easier to market the LNG and get the project off the ground than a 5 mta train project: "it is easier to coordinate issues among 2 or 3 buyers compared with 5 or 6". It uses the Phillips optimized cascade process, withGE LM 2500+ aero gas turbines, which are similar in power to Frame 5Ds but have a higher thermal efficiency: 41.1 % compared with 30.3%. This improved efficiency contributes to reduced greenhouse gas emissions. Further reduction is achieved through waste heat recovery on the gas turbines, and recovery of ship vapours.

Another LNG plant under construction which has placed a major emphasis on greenhouse gas emission reductions is the Snohvit project, Norway, described in a paper by Roy Scott Heiersted of Statoil. It uses CO2 reinjection offshore, at a cost of $200 million. This will be the first electric motor-driven LNG plant, using five LM 6000 aero-derivative gas turbines with low-NOx burners, and waste heat recovery, to generate power for the refrigerant compressors. It is the first application of the Statoil-Linde Mixed Fluid Cascade process. This has three cascade refrigerant cycles, each of different mixture composition. The process uses two Plate Fin Heat Exchanger assemblies and two Linde spiral wound heat exchangers. A paper by Manfred Steinbauer of Linde, in conjunction with Statoil, described the testing in operation of the Linde exchanger at the Mossel Bay refinery plant in South Africa. Apart from Snohvit, these exchangers are being supplied for Australia NWS train 4 (the first application), for the Sakhalin project, and as replacements in existing trains of Brunei LNG.

Another new process is to be used for the first time in the Sakhalin II LNG plant, described in a paper by Andy Calitz of Sakhalin Energy Investment Co. The plant has two 4.8 mta trains and is now under construction; it will use the Shell Double Mixed Refrigerant Process. This uses mixed refrigerant in both the pre-cooling and the main cooling cycle, each employing a Linde spiral-wound heat exchanger. The paper describes the design measures employed to prevent freezing of water and other process streams in the very low ambient temperatures, and the need for greater use of low temperature and stainless steels.

Plants of the Future

Several papers were presented that described studies and designs of large liquefaction trains of 7.5 mta capacity or more, with the aim of achieving even greater economies of scale and reductions in the unit cost of LNG production.

Mark Roberts of Air Products presented a paper about the use of the AP-X process for larger trains. The author said that the problem with the C3-MR process at train capacities significantly greater than 5 mta was that new designs would need to be developed for several major equipment items. Air Products has therefore developed the AP-X process, which adds a third refrigeration cycle - a nitrogen expander - at the back end of the two cycles of the C3-MR process, offloading the other two cycles, allowing them to use compressors and exchangers of proven size. It could use various options of gas turbine or electric motor. A cost of $151 /tonne/year for an 8 mta train was quoted compared with $168/tonne/year for two 4 mta C3-MR trains. In answer to a question the author stated that the process had a similar make up of mixed refrigerant as for the C3-MR process, and a similar reliability.

Phil Hagyard of Technip, in conjunction with Total, described studies of the C3-MR precooling cycles for high train capacity. The study came to a different conclusion from Air Products: alternative configurations studied showed that the basic C3-MR process could go to 8 mta using proven equipment and either three Frame 7 drivers, or two larger Frame 9 drivers if Frame 9s were to be qualified for LNG service.

Grant Thompson of the Qatargas II project, in conjunction with ExxonMobil and Qatar Petroleum, presented an overview of the Qatargas II full supply chain.. The project is in the FEED stage. Each of the two 7.8 mta trains will use the Air Products AP-X process. The project developers have satisfied themselves that Frame 9 gas turbines can be used: the plant will use three Frame 9 gas turbine drivers, one for each cycle, with starter/helper motors and heat recovery steam generators. As the plant is likely to supply the UK market, which requires a lean gas, it is designed with NGL recovery using the Ortloff Engineers' SCORE expander-based cryogenic process.

Marc le Metais of Total presented a paper on "Large Capacity Trains", which reiterated the problem in scaling up the standard AP C3-MR process. Propane compressors were close to their technological limits, deficient performances have been experienced with some compressors already constructed. The largest main heat exchanger that AP can make in their current workshops is suitable for

5 mta capacity. The paper compares the AP-X process with the C3-MR process using dual liquefaction strings (two MR cycles in parallel, and one of several options for reducing the load on the propane compressor). The conclusion was that the AP-X process was more complex, with lower availability. Larger capacity trains, e.g. 9 mta, were feasible with a three Frame 9 dual string arrangement.

Other papers based their studies on other liquefaction processses. Sergio Buoncristiano of GE and Amos Avidan of Bechtel, in conjunction with ConocoPhillips, presented a paper about a new generation of larger plants based on the Phillips process. The paper describes studies that concluded that the larger volume compressors required to go up to 8 mta would be within proven design parameters. Also large single shaft gas turbine drivers (Frame 7 or Frame 9) were feasible for the Phillips Optimized Cascade Process.

J. Pek of Shell Global Solutions International presented a paper on "Large Capacity LNG Plant Development" based on the Shell DMR process. The authors claim that for a 5.5 mta train, the Shell DMR process is more efficient than the C3-MR process (11.7 kW/tonne/day compared with 12.4 kW/tonne/day). For an 8 mta train design, Shell has developed the Parallel Mixed Refrigerant Process : the main liquefaction cycle is split into two parallel streams. Three Frame 7 gas turbine drivers are used : one for the precooling cycle and one each on the liquefaction streams. The parallel configuration on liquefaction is claimed to give higher availability than other processes (such as AP-X and Phillips) which use precooling, liquefaction and subcooling cycles in series. The authors are cautious about going to larger Frame 9 drivers. They believe that electric motor drive has potential advantages, but the industry has little experience with the larger electric motors required.

Richard Jones of BP, in conjunction with KBR, presented "BP's Big Green Train: Benchmarking Next Generation LNG Plant Designs". This study took a design basis for an electric motor driven plant, with objectives on plant capacity, CO2 emissions, autoconsumption and EPC Cost reduction, and evaluated designs submitted by four process licensors : Axens (Liquefin), AP (AP-X), Linde (MFC) and Black & Veatch (PRICO). The results were presented in a form that meant that specific processes could not be identified. All met the objectives, or would readily do so on futher optimisation.

The authors had as one objective a lowering of the EPC cost per tonne/year of LNG produced by 25% compared with Trinidad Train 1. In a question, Phil Redding (BG) pointed out that in fact a similar reduction in unit costs had been achieved in expanding the Trinidad plant from one train to three: economies of scale might also be achieved with multiple smaller trains (through savings on common services, etc.) as well as by building one large train.

In Workshop Session 1 on Technical Innovation, co-chaired by Rob Klein Nagelvoort of Shell Global and Paul Sibal of ExxonMobil, there was further discussion on the merits of electric motor driven plant. Although they had lower efficiency, this was compensated for by higher availability. One big advantage is that electric motor drivers decoupled the selection of the gas turbine driver from the compressor, and created more competition - which should drive down costs. At present GE had a dominating postion as leading supplier of gas turbines for mechanical drivers in LNG plants. More vendors were able to bid to supply combined cycle power plants producing the power for electric motor driven plants. Another advantage was the faster schedule as the power plant could be built in parallel with the liquefaction train. Also, it was thought that combined cycle heat recovery on a gas turbine providing mechanical drive added complication in the middle of the liquefaction train, whereas it was a standard part of a separate power plant.

In Workshop Session 4 on "The Commercial and Technical Interface", co-chaired by Charles Durr (Kellogg, Brown & Root) and Graham McNellie (BP), the Panel agreed that although super large trains lead to economies of scale, the train size had to be "fit for purpose": cost savings were lost if the ramp up period was too prolonged because the LNG could not be sold. There were still greater efficiencies and cost reductions to be achieved for smaller trains (5 mta or less) due to design optimisation and new technologies.

The issue of contracting strategy was addressed in a paper by Charles Durr of KBR - "LNG Project Design Competition - A Contractor's Viewpoint". It has been increasingly popular for project developers to use a contract strategy involving a design competition: a number of contractors develop parallel FEED packages with a lump sum bid for the EPC contract. The authors discuss all the issues to be addressed by the project developer to ensure that he achieves the objective of stimulating innovation and cost savings in design, in a process that is fair to all the participants. It is necessary to carefully define the scope of the FEED - what is fixed and what is left to the contractor's discretion, and to define in advance the procedures and criteria the project developer will use for evaluating prototype or improved designs. It is also essential not to share a contractor's ideas with his competitors, unless that is agreed before the contractor decides to participate, otherwise trust and the incentive to innovate will be lost.

Safety

The Skikda Explosion: As a late adddition to the programme, Mr. Bachir Achour of Sonatrach gave a presentation about the major explosion at the Skikda LNG plant in January. The sequence of events was as follows :

• At 6.39 p.m. on 19 January steam pressure was seen to rise in the steam
  boiler adjacent to the 0.85 mta capacity train, GL1-K unit 40. The operator took action to reduce fuel input to the lowest level
  
• A visible vapour cloud was seen at 6.40 p.m. on train 40
  
• The steam pressure in the boiler continued to rise, because unknown to the operator flammable vapour from the external vapour cloud was being drawn into the air intake (aspirator fan), resulting in an explosive mixture in the boiler fire box.
  
• Seconds later, at 6.40 p.m. there was a first explosion (the boiler) followed immediately by a massive explosion and fireball - a vapour cloud explosion
  
• A large fire covered trains 40, 30 and 20, owing to release of flammable gas
  and liquids caused by major blast damage. Emergency medical help was sought for injured persons. The site's fire-fighting team from the unaffected part of the plant, the fire services from the Skikda industrial area and the regional fire brigade took action to protect train 10, nearest to the fire, and the LNG storage tanks. Remaining LNG operations were shut down.
  
• After 8 hours, the fire on trains 20, 30 and 40 was extinguished.

Trains 20, 30 and 40 were destroyed in the explosion, together with the adjoining maintenance and security buildings. 27 persons on site were killed, and 56 injured, mainly due to blast and collapsing buildings, not fire. Train 10 suffered minor damage, mainly from flying fragments of plant from the explosion - it was shielded by trains 20 and 30 from the blast. The LNG storage tanks, LNG loading facilities, and trains 5 and 6 (800 m away) were undamaged. Steps have been taken to establish temporary maintenance, telcommunications and computing services. Trains 5 and 6 are scheduled to start up in May and June 2004. Train 10 with its utilities is being isolated from the destoyed trains, and will be subjected to a detailed inspection, repaired and brought back into operation by October 2004.

In 2005 the Skikda capacity will be 20.2 mta (compared with 23 mta before the explosion). Sonatrach has decided to replace the destroyed trains with one new 3.8 mta train by 2007. This will be followed by a further 6 mta of new capacity by 2010 through the already planned Gassi Touil expansion project.

Mr. Achour stated that the source of the vapour cloud over train 40 was not yet known. A low temperature fluid was released, as the vapour was visible. It appeared to come from a location where there were several fuel gas lines carrying low temperature methane, as well as a refrigerant line. As there was no wind, a denser-than-air cold methane vapour cloud would not disperse quickly into the atmosphere. (In a later Workshop session S. Baussoualem of Sonatrach said that although from the scale of the damage at Skikda it may have been more reactive refrigerant vapour, he could not rule out the possibility that the vapour cloud was methane/LNG vapour - there had been one well-documented case in the past in a chemical works of a vapour cloud explosion resulting from low temperature methane. (Note : Although an unconfined flammable LNG vapour cloud will not support a high flame speed, large scale experiments have shown that given an initial confined explosion as ignition source and sufficient partial confinement/congestion in the vapour cloud - as was the case at Skikda - a vapour cloud explosion can be sustained in a methane/air vapour cloud).

Mr. Achour stressed that this accident was a tragedy for those who lost their lives or were seriously injured and their families, and for Sonatrach. It occurred despite Sonatrach's major programme of plant revamping and upgrading of safety in the 1990s, and for two years its plant had been under a new Directorate. He praised the site's emergency plans, and the disaster plan of the Skikda industrial area, which had been shown to be effective. He promised that the results of the investigation in progress and the lessons learnt would be fully shared with the LNG industry.

Felix de la Vega of Kellogg Brown and Root presented a paper on "Designing Safety into LNG Export/Import Plants". The authors stress the importance of going beyond compliance with regulations, codes and standards, undertaking a range of safety and hazard reviews during design using experienced personnel, and undertaking hazard assessments. Also the principles of inherently safer design should be followed : using lower inventories of hazardous materials, using a safer material in place of a hazardous one, simplifying by reducing equipment, etc. There was a significant question from the audience: what should companies do who do not have the hazard assessment and design tools used in design but have to operate and maintain LNG plants after the EPC contractors are long gone? Mr. de la Vega said that plants needed re-assessment every five years.

Tony Acton of BG Group, in conjunction with Tractebel LNG, Gaz de France, Osaka Gas, and Tokyo Gas, presented a paper "LNG Incident Identification - A Compilation and Analysis by the International LNG Importers Group (GIIGNL)". This is a good example of a thorough co-operative safety study by the LNG industry: 246 incidents of releases of hazardous material, near misses and other incidents of concern over the period 1965 to 2000 have been reported and analysed for GIIGNL members' LNG reception terminals and peak-shaving facilities. Only 11% of the events reported resulted in an explosion, fire or rapid phase transition, and the frequency of reported incidents is low : 0.33 incidents per site-year. There is a trend towards a decrease in the relative number of events where significant quantities of hydrocarbon have been released. GIIGNL is committed both to improving further the reporting of incidents and to maintaining its database up-to-date for the general good of the LNG Industry.

The Workshop Session on Safety and Security of the LNG Chain was co-chaired by Amos Avidan (Bechtel) and Noriyoshi Nozawa (Chiyoda). Each panellist gave an initial brief presentation. Ad Small of Shell Global Solutions mentioned the need to maintain safety standards as installations got older and as new players without prior experience of LNG entered the rapidly growing LNG industry. It was important to underpin the scale up of plants and equipment with R&D to ensure their integrity. Takinao Hojo of Mitsui OSK Lines stated that MOL's LNG Carriers are safety audited from time to time. He questioned the practice of having "LNG" in large letters on the side of ships - could this be inviting terrorist attacks?

Mr. Uchino of Tokyo Gas noted the recent revisons of Japan's Recommended Practice for LNG Receiving Terminals, and seismic design codes. He outlined the extensive efforts in Japan on accident prevention through safe design and construction, high-level maintenance, operator training and safety systems. Mr. Iizuka of Mitsubishi Heavy Industries explained safety concepts in design of storage tanks in Japan. In answer to a question he said that it is not the type of tank (aboveground, inground, etc.) that is important, but the safety features in the design, construction and operation.

Bill Lewis of PTL Consultants said that the definitions of containment in NFPA 59A has resulted in the anomaly that in the USA the authorities require even a full containment tank to have a low impoundment wall (in effect tertiary containment). It was still not clear to what extent specific provisions of the onshore codes and safety requirements would apply to US offshore LNG facilities. FERC require threat assessments and much effort is going into the modelling of damage and LNG spill scenarios.

In answer to a question on the commercial impact of the Skikda disaster on the Industry as a whole, Ad Smaal said that any major accident will have an impact, not on the ability to raise financing, but on insurance costs and on the ability to obtain permits. Bill Lewis said that the Skikda incident is already having an impact on the permitting process for new LNG terminals in the USA.

There was a discussion about impossible "worst-case scenarios" postulated by objectors to new LNG facilties, and what should in fact be regarded as realistic accident scenarios for LNG shipping and storage. It was noted that recent experiments had indicated that penetration of the primary containment in LNG Carriers was unlikely to result from explosions alongside the outer hull.

Towards the end, the Panel members were able to stress once again the importance of safety reviews and hazard assessment not only in design but during plant commissioning and operation, and when modifications were made. Ad Small said a plant that ran like clockwork could lead to complacency - continual vigilance over safety was essential.

Editorial Comment - LNG Hazards

There is a need for the high standards of safety in the LNG industry to be maintained as LNG trade grows; as plants, terminals and ships get larger; and as more companies enter the business. A number of industry groups and individual companies will be reviewing LNG safety issues in the light of the recent Skikda explosion. There are several general issues that need to be addressed. They include :

(a) Is there a need for a worldwide system of independent safety audit and "certification" at regular intervals of operating LNG shore facilities as well as ships ?

(b) Is it necessary to review again all LNG plants, terminals and ships to determine whether there are circumstances where, in the event of a release of flammable vapour, that vapour could enter, or be drawn into, plant, buildings and other confined spaces, and thereby present an explosion hazard ? For example, air intakes of gas turbines and other gas-fired equipment, site buildings, underground ducting, accommodation and machinery spaces on LNG carriers, or ship's ballast tanks. If such circumstances exist, how can the hazard be reduced or eliminated, e.g. through building/air intake redesign, gas detection and automatic shutdown systems ?

(c) In design and layout of new liquefaction plants is it standard practice worldwide to consider the possibility of vapour cloud explosions when considering layout and siting, and the protection of key facilities such as storage tanks, ships, control rooms, etc? What additional measures can be taken in layout (e.g. avoiding partial confinement and congestion), or in provision of safety systems, or in ignition prevention, to help prevent vapour cloud exlosions should an accidental release of flammable vapour or liquefied gas occur? Is it feasible to modify in this way existing liquefaction plants to reduce vapour cloud explosion hazards ?

(d) Is it necessary in design of LNG reception terminals and LNG carriers to consider and study the possibility of vapour cloud explosion hazards in LNG spill scenarios involving partly confined/congested plant areas (and discount them if appropriate to do so) instead of discounting them from the outset based on research into unconfined vapour cloud behaviour, or because of their absence from Code requirements ?

(e) As it is important to consider "inherently safe" plant design, is it reasonably practicable to constrain the size or design of liquefaction trains to minimize inventories of the more reactive flammable refrigerants (such as propane and ethylene) ? Is it feasible to develop a liquefaction process that has substantially reduced inventories of flammable refrigerants, or of the more reactive types of refrigerant, or which does not use flammable refrigerants at all ?

An impressive LNG 14 Exhibition