GB2409707A

GB2409707A - Liquid metal heat recovery in a gas turbine power system

Info

Publication number: GB2409707A
Application number: GB0330198A
Authority: GB
Inventors: Noel Alfred Warner
Original assignee: Individual
Current assignee: Individual
Priority date: 2003-12-31
Filing date: 2003-12-31
Publication date: 2005-07-06
Anticipated expiration: 2023-12-31
Also published as: GB2409707B; GB0330198D0

Abstract

A gas turbine power generation system includes hot gas clean-up of the gas fuel supplied to the gas turbine and a closed-loop liquid metal heat recovery means 12. The hot gas may be quenched by massive steam injection so as to decompose the ammonia content of the gas. The first step of the closed-loop liquid metal heat recovery means 12 may comprise quenching raw high temperature fuel gas, from a gasification process, in a liquid metal quench 9. The gasification process may be oxygen, rather than air, based. The raw fuel gas may then proceed to filter system 13 and then the liquid metal heat recovery system heats the fuel gas in re-heater 31 before it reaches the combustor 22 of the gas turbine.

Description

LIQUID METAL SYSTEMS FOR GASIlICATlON-BASED POWER GENERATION In a co-

pending Application (PCT GB 2003/0030S5) attention was focussed on the gasification of carbonaceous solids using a liquid-metal system for transporting the solid feed materials through a horizontal pressurized gasifies. Further development of this concept is now preserved as well as its extension to a whole power generation cycle, which is uniquely chcer by a close loop circulation of liquid metal serially servicing me requirements of a number of subset systems, which when added together constitute an integrated gasification and power generation process in its totality. It has been discovered that this novel approach is especially advantageous in teens ofthe net power output divided by the heating value of the carbonaceous material consumed in its generation, referred to universally as the cycle efficiency and also in the abatement of air toxic and other harmful emissions normally associated with fossil fuels.

Using a closed loop liquid metal system is also seen as a means for enhancing the performance of certain other integrated gasification and power generation processes now commercially proven, such as the Shell Coal Gasification Process (SCGP) so the concepts developed need not necessarily involve the newly proposed liquid metal transport gasifies, but can come into effect for example by bringing together the raw fuel gas from a Shell gasifies at an elevated temperature, typically in the region of 1500 C, and the circulating liquid metal in an attenuated condition to provide a liquid metal quench (LMQ) as the first step in the closed loop system.

Enhancement of cycle efficiency reduces greenhouse gas emissions, but clearly for successful commercialization the cost of electricity generated must be competitive with other technologies. However, in response to concerns about climate change and global warming, and in the longer term, the power generation industry may be forced to adopt near to zero gas emissions scenarios or otherwise face punitive statutory charges related to the emission of carbon dioxide (CO2). This appears to be one of the factors influencing the study made by J. Lawton as part of the Department of Trade and Industry (DTI) Coal Research and Development Prograrnme in 1997 (Report No COAL R100). Amongst other suggestions, he identified possible overall cost reductions in power generation using flue gas recycle with partial or full Steam Injected Gas Turbine (STIG) technology and recommended that as a long-tenn measure, it would be prudent to construct new power facilitates so that flue gas recycling could be facilitated in the future for back-end CO2 removal. This invention closely relates to this philosophy and provides a practical cost- effective means for achieving this in the near tend so that the high capital costs of air-separation for production of oxygen, an essentffal component in a semi-closed Bmyton cycle of the type envisaged by Lawton, are avoided initially and only contemplated if retrofitting of flue gas recycling to the gas turbine and perhaps associated gasified becomes an economically feasible proposition sometime in the future.

STIG was first introduced in the mid-1950's with steam piped directly to the inside of the combustor liner and this was followed in the 1960's by injecting steam into the compressor discharge of the gas turbine.

More recently, steam teas been injected with the fuel, or around the fuel nozzle tips to control nitrogen l oxide (NEIL) emissions. Also a mixture of steam and reformed natural gas has been proposed as the fuel for a gas turbine in the Chemically Recuperated Gas Turbine Cycle (CRGTC). In the early 1990's a joint EPRI/ShelUGE programme blended steam with oxygenblown SheD syugas in order to lower the heating value of fuel gas in combustion screening tests. Further GE work by Cook, Corman and Todd (ASME Oct.

1995) licks this earlier work with LBTU (Low BTU gas) such as that produced by air-blown gasifiers, by adding steam to fuel gas to reduce its heating value in order to simulate LBTU.

In the present invention, steam is injected into partially pre-combusted fuel gas, which has been given sufficient residence time at elevated temperature in advance of the gas turbme combustors for chemical equilibrium to be closely approached, typically at 1100 C or preferably somewhat higher temperature, so that ammonia in the fuel gas is decomposed heterogeneously not at the typically lower temperatures associated with catalytic ammonia decomposition, but rather in the regime where both chemical and gas- phase transport controls are both important but the latter exerting the greater influence. in this so-called inixed control regime", expensive catalysts are not requited but rather commodity market standard products, such as INCO carbonyl nickel pellets can be used effectively to conduct the heterogeneous decomposition of fuel gas ammonia before the high temperature gas is quenched with steam and transmitted rapidly to the gas turbine combustors using pipes comprised of current state-of-the-art heat resisting alloy materials up to temperatures presently recognized by gas turbine experts as high as 870 C, but preferably at somewhat higher temperatures, dependent on furler development of oxide dispersion strengthened (ODS) materials.

It will be appreciated that the approach described above is distinctly different from the rationale behind STIG in its traditional sense and that this new approach offers a novel basis for cantrolting NO, emissions in power generation using gas turbine combustors similar to those used presently in the so-called 1500 C Class gas turbines utilising canannular combustors and the like.

In this connection it is worthy of note that the GE work previously referred to, assessed fuel gas preheat up to 344 C, but the present invention is calling for fuel gas preheating in excess of 800 C. There has been a great deal published in the technical literature about the benefits of highly preheated combustion air for furnaces, but similar emphasis has not been directed to highb preheated fuel gas in the past. The use of liquid tin-based systems in the present invention makes the use of highly preheated fuel gas a practical proposition, which in turn is reflected in enhanced cycle efficiency being attained for power generation.

In the related but distinctly different approach of second generation Pressurised Fluidised Bed (PFB) combined cycle, employing topping combustion to raise the turbine inlet temperature for enhanced cycle efficiency as described in the articles by Beer and Garland (A coal-fuel combustion turbine with topping con bustion. Trons. ASME, J. Eng Gas Turbines and Power, 119, t997, 84-90.) and Domeracki, Dowdy and Backovehin (Topping combustor status for second-generation pressurized fluidized bed cycle application, Trans. ASME, J. Eng. Gas Turbines and Power, 119, 1991, p 27-33) the syugas entering the topping combustor has been previously cleaned of particulates and alkalis by the hot gas cleanup (HGCU) system but it still contains significant fuel bound nitrogen present as ammonia and other compounds that will selectively convert to NOx, if the fuel is burnt under the oxidising conditions of standard combustion turbine combustors. In the present proposal this problem is alleviated by equilibration for ammonia removal from the high temperature fuel gas under reducing conditions in advance of admission of the high temperature fuel gas to the combustors. This is not an option with the topping combustion system discussed by the authors just referred to. Because the air entering the combustors in their case is at 1600 F (871 C), the conventional type of combustor is not suitable as both emissions and wall cooling concerns preclude the use of conventional design. The combustion a* in the present invention comes directly from the gas turbine compressor so mere should be no problem with state-of- theart combustor technology. On the other hand, both the topping combustors of the second generation Pressurised Fluid Bed (PFB) cycle and those of me present invention involve fuel gas inlet temperatures higher than customarily used.

However, with the present invention this is not exacerbated by vitiated air at a temperature at 870 C or thereabouts arriving at the combustors. While a totally new approach is necessary in one case, a relatively straighffotward modification of the immediate fuel inlet system is all that is required with the approach of the present invention.

HGCU with regenerable sorbents for sulphur retnoval Connally operate at around 650 C. Catalytic ammonia decomposition is requiem at these tempera levels to secure sufficient decomposition. The fact that catalysts are required is irnmediateb indicative of the predominance of chemical reaction kinetic control, but as the temperature is increased markedly there will be a shift in the rate controlling step towards gas phase mass transfer. Accepting that the chemical process of ammonia composition is heterogeneous rather than homogeneous at say 1200 C and that, for example, a nickel surface is required at this temperature, it can be shown that the rate of the chemical reaction at the gas/solid interface is some 700 times more than it would be if the temperature were at 650 C. Accordingly, the rate controlling step can be anticipated to be transport of gaseous reactants and products to and from the reaction interface, rather than the intrinsic chemical reaction rate itself. For simple heterogeous ammonia decomposition, suitable solid packing materials as for example carbonyl nickel pellets are available commercially at the market price for refined nickel metal and are therefore relatively inexpensive in comparison with especially prepared nickel catalysts. These pellets have already a suitable size distribution to form a porous packed bed and are chemically inert at the elevated temperature under discussion provided of course a reasonable level of desulphurisation is attained during HGCU. Neither nickel sulphide nor nickel oxide is thermodynamically stable under these proposed operating conditions.

Chemical equilibrium will be established at the surface of the nickel pellets but ammonia will not decompose heterogeneously at a rate faster then the gas phase mass transport of ammonia molecules from the bulk of the gas to the solid surface. The rate is readily calculable so Hat dimensions of a non-catalytic reactor can be precisely defined in teens of the operating temperature and the physical properties of the gas phase. If some chemical kinetic influence still persists at 1200 C, then clearly the reactor size will have to be increased accordingly.

The prerequisites for effective ammonia decomposition as a precursor to low BOX emissions for the gas turbine's combustor as identified in this invention are partial oxidation of the cleaned fuel gas to an elevated temperature of say at least 1100 C or probably 1200 C or somewhat hotter, provision of gas/solid contacting in an appropriately sized bed packed with relatively inexpensive carbonyl nickel pellets and then a sudden quench with steam back to a temperature level such that ammonia is not given the opportunity to reform, All of this must be achieved before the fuel gas is admitted to the gas turbine combustor.

To ameliorate the reduction in heating value of the fuel gas resulting from partial pre-oxidation of the raw fuel gas in advance of the gas turbine combustors, with fuel gas derived from gasification, it is imperative that the raw fuel gas is already at a relatively high temperature, such as 750 C or above, in advance of ptecombustion to take this temperature to the 12000c level or thereabouts to effect ammonia decomposition in pursuit of ultimate low NOx emissions from the gas turbines.

There are a number of ways of ensuring that the raw fuel gas is at least at 750 C before partial pre- oxidation, depending on whether a traditional approach to HGCU is utilised or whether the approach favoured in this invention is adopted.

With a liquid metal-based system it is advantageous to quench the raw fuel gas emanating from a gasified by direct countercurrent contact with a liquid metal coolant, filter the gas free of particulate solids at a relatively low temperature using conventional filter systems and then reheat the fuel gas to say 750 C or higher, again by countercurrent direct contacting with the liquid metal discharged from the liquid metal quench (LMQ3. By means already identified by the inventor in PCT GB2003/0030S5 the raw fuel gas can be desulphurised and cleaned up of air toxic metals, alkali metals and halides during the LMQ and also the liquid metal medium can itself be rendered "superclean" so that its further contacting with filtered fuel gas does not introduce contaminants during re-heating.

High temperature gasifiers of any type can be adapted to this approach by employing a closed loop liquid metal system with in-line quenching and hot gas clef capabilities for removing chemical pollutants responsible for acid rain and smog and air toxic emissions such as mercury constituting hazards to human health. If tin is the preferred liquid metal, the clean fuel gas can be cooled to 240-235 C for ultimate removal of particulate solids. To preserve thermal efficiency it is then mandatory that the fuel gas is re- heated as discussed at length by the inventor in co-pending PCT GB2003/003055. However, there may still be potential problems with the fuel nitrogen NO, emissions with the approach advocated in the earlier patent application.

The present invention addresses this situation in a reliable and noncomplicated fashion. The essential prerequisites are firstly the availability of quite large amounts of steam and secondly the ability to re-heat cooled fuel gas to an elevated temperature and then to conduct partial pre-combustion and equilibration at an even higher temperature and then finally to effect rapid quenching with steam to freeze the equilibrium composition. Lois is immediately followed by rapid transmission of the quenched fuel gas at the maximum temperature possible with state-of-the-art heat resistant alloy tubing to the especially modified can-annular combustors of a 1 500 C Class gas turbine or whatever upper temperature level pertains to gas turbines of the future. With present alloy materials 870 C is probably the maximum temperature but, with the prospect of advances in ODS alloys, this temperature may be increased in the not too distant future.

Of the various candidate liquid metals potentially available for systems associated with power generation, tin is believed to be superior to lead or lead-bismuth eutectic both on health and safety grounds as well as on technical merit. In more than one of the individual component systems in the overall power generation scheme, a tin based system introduces vastly superior phase separation attributes, which far outstrip those of the other candidates. For example, the desulphmsation system for raw fuel gas in HGCU relies critically on the huge difference in volatility exhibited between stannous sulphide and elemental tin at temperatures in the region of 1100-1500 C to render the tin "superclean" in advance of its use to reheat the cooled filtered fuel gas back to high temperature as well as providing the driving force for desulphurisation itself At equilibrium, the vapour pressure of tin sulphide is more than twice that of lead sulphide at these temperature levels, but the full superiority is better seen by comparing the ratio of equilibrium vapour pressures between the sulphides and the parent metal. This parameter is more than three orders of magnitude greater for the tin-based system than for the corresponding bad-based system. Similarly, iron contamination of the liquid metal transport medium resulting from high temperature reduction of coal ash in contact with carbon during gasification is relatively easily dealt with by vacuum distillation of tin sulphide from the metallic crosses formed when the liquid metal medium is cooled from say 1400 C to 240 C as it is in a liquid metal based transport gasifies flowsheet.

There is, however, a negative attribute introduced by selecting tin as the preferred liquid metal both in transport gasification and in fuel gas reheating by direct contacting with liquid metal, a vital component in the quest for improved thermal efficiency. Oxide volatility is a more serious threat for a tin-based system than for a lead-based system. Tin losses can be dealt with by filtration at the relatively low temperature level 240 C but the more serious aspect is deposition of oxide solid accretions on cooler surfaces within the transport gasified and probably more importantly on turbine blades resulting from volatile stannous oxide being oxidised to relatively non-volatile stannic oxide during combustion of the fuel gas and its carry-over, possibly as an aerosol, into the expander of the gas turbine. This places a severe restriction of the amount of stannous oxide gas that can be tolerated and immediately focuses attention of the oxygen potential of the fuel gas, which from a thermodynamic viewpoint governs the formation of the volatile oxide in the fast instance.

With air-blown gasification, simple heat balance considerations for a liquid-metal transport gasifies dictate Mat a certain amount of the carbon in the solid feed must be oxidised through to carbon dioxide to provide the thermal input to counteract the heat absorbing influence of the otherwise inert nitrogen associated with the air blowing, which passes through the gasification system without contributing to the chemical reactions involved but seriously detracting from the inherent energy output of the gaseous fuel produced.

With oxygen/steam top blowing, on the other hand both oxygen and steam contribute to the gasification reactions without the nitrogen diluent absorbing valuable energy. This means that with oxygen blowing it is possible to limit the concentration of CO2 to an absolute minimum and thus produce fuel gas with exceedingly low oxygen potential. Thus, contrary to initial expectation, the problems of oxide excretions are not exacerbated by substituting oxygen for air in the gasification system, but rather are very considerably reduced. This is provided, of course, that contact between the more aggressive oxidant and molten tin is totally eluninated within the gassier itself This conclusion focuses attention on the absolute necessity for screening the surface of molten tin within a transport gasiffer from contact with the highly oxidismg gases used to gasify the carbonaceous material until reaction has reduced the oxygen potential to relatively benign levels. The copending Application relies on the provision of carbonized material floating on the liquid metal surface to protect it from oxidation, thereby precluding access of gaseous oxidants to the tin surface other than in reduced form. Accordingly, not all the feed is gasified in a single pass, or alternatively a protective carbon-containing hearth layer is deposited on the liquid metal surface in advance of the material to be gasified. The bulk of this carbon containing hearth material is recycled after physical separation from the solid slag eventually issuing from the gasifies.

The above approach is perfectly satisfactory for gasification of biomass and other materials with very little ash content or, indeed, for an ash that has an excessively high fusion temperature requiring an inordinate amount of flux material to effect melting but for most coah better use can be made of the liquid slag sheet Donned and this is now incorporated as a principal aim of the present invention.

By carefully controlling the molten tin circulation rate and the depth of the molten tin in the gamier, a liquid slag sheet can be formed of dimensions such that its width extends over almost the whole width of the tin bath, closely approaching the refractory walls on both sides without actually coming into contact with them.

Regarding the thickness of the slag sheet ultimately formed in the proposed process and referring to Pilkington (The float glass process. Proc. Ray. Soc. Land. A, 1969, 314, 125.) who derived the fonnula for the equilibrium thickness T as a function of the surface tension of both liquid phases tr, and cr2; the interracial tension cry 2 and the phase densities pi and P2 by the force balance equation given as Equation 1, where I and 2 refer to the glass or slag and the liquid metal, respectively: T2 =(t +(-2-J2) P2 [1] gyp, (P2 - P. ) For the float glass process, Pillcington gives the equilibrium thickness T as approximately 6mm for a typical plate glass composition. The consensus among more up-todate sources seems to favour 7mm for current glass formulations and on this basis the glass/tin interracial tension can be calculated from Equation I to be about O.SO N/m. As a first approximation, it is assumed that the interracial tension for a typical coal ash slag and a liquid metal is also about O.SO N/m.

Using relationships given in Smithell's Metals Reference Book (The physical properties of liquid metals, Metals Reference Book, 7 edn, Brandes E.A. and Brook G.B., eds. (Butterworth-Heinemann, Oxford UK, 1998) 14 - 8, Table 14 - 8) the density and surface tension of molten tin at 1500 C are 6223 kg/m3 and 0.45S N/m, respectively and for Daw Mill UK coal ash-derived slag the density is estimated to be about 3000 kg/m3 at 1500 C and the surface tension is in the region of 0.40 N/m based on information given in INCRA Monograph III (The Thermic Properties of copper-slag systems, 1974 Ed. Carlos Diaz, Internatio=I Copper Research Association, Inc. USA, 9-11, 16-2). Accordingly from Equation 1, the equilibrium thickness is calculated to be 7.6mm' so to a first approximation the equilibrium thickness for both float glass and coal ash slag is similar at about 7mm.

Accepting that the ultimate equilibrium thickness is a function of the chemical composition of the ash in that this determines the relevant slag physical properties, on-stream chemical analysis of the feed material upsteam from addition to the gasified is used as the basis for controlling the appropriate molten tin circulation rate so that for the required gasification rate or solids feed rate in response to load variations if the gasiffer is integrated with power generation, so that the primary criterion of protecting the molten tin from contacting oxidant gases is adhered to rigorously at all times. Because the slag formation and melting occurs over a relatively short distance after the appropriate temperature is reached, liquid slag falling onto the slag sheet from the gasifying carbon raft above can be engineered to cease well before the antioxidation shielding surface requirement is needed. By the time the carbonized material forming the floating ran is depleted to such an extent that complete protective coverage is no longer possible, the slag sheet has reached its equilibrium thickness and the tin flow is controlled so that its associated floating slag layer covers virtually all molten bath leaving only a small clearance so that physical contact between slag and refractory is avoided.

Besides implementation of the control measures specified above, the other variable, which has to be under close control, is the depth of the molten tin bath. Electromagnetic devices for controlling melt levels in continuous casting have been developed and an adaptation of these may be used in conjunction with ordinary control means so that the molten tin surface is maintained close to a continuous projecting refractory shelf immediately above and covering the slag free edges on both sides of the slag sheet so that admission of a minimum quantity of inert purge gas into the small gas clearance is sufficient to prevent ingress of oxidising gases coming into contact with the residual exposed surface of the tin bath.

By the means just described, the provision of recycled carbonized hearth material can be completely dispensed with. This is a considerable simplification in terms of ultimate disposal of a premium glassy slag material with a range of possible uses including cement manufacture, building aggregate and other civil engineering applications.

The inert purge gas referred to may be nitrogen from an air separation plant if available, a bleed stream of fuel gas which has been treated to reduce its oxygen potential below that of the main stream product fuel gas or possibly, if readily available, a small purge of natural gas may be the simplest approach, particularly if nitrogen addition complicates an otherwise zero gas emission power generation scenario after carbon dioxide sequestration.

Molten tin is an essential ingredient m a number of subset systems which can collectively constitute a zero gas emissions power generation scheme and therefore in the present context due attention needs to be paid to the total overall system beginning with gasification. However, certain existing commercially established gasification technologies can be combined advantageously with tin-based HGCU to yield worthwhile efficiency gains and overall improvement in performance. This approach should proceed in parallel with the development of new gasification technology such as that outlined in the co-pending PCT application.

Further development of the Shell Coal Gasification Process (SCGP) is particularb well placed in this regard.

With the recently proposed liquid metal transport gasifies system, tin is directly involved from the outset but even here to reach its full potential and if zero gas emission is the ultimate target, important process modifications must be made and issues such as the decontamination of the liquid tin carrier medium, not dealt with previously, must now been addressed.

Also a tin based desulphurisation system in the context of HGCU in a zero gas emission mode needs identification for the purposes of delivering an elemental sulphur product by a continuous process compatible with advanced power generation. The present invention focuses on all these various aspects as individually they must all be capable of being integrated into an overall tin-based scheme with verifiable real prospects for significant improvements over current technology.

Co-firing of sewage sludge, for example, and coal in existing pulverized fuel (p.f.) Fred powered generation plant is now becoming an accepted commercial practice rather than further development of new technology dedicated to the particular waste or other carbonaceous arising by itself. The same co-firing approach has also been advocated for biomass and coal in p. boilers. On the other hand, dedicated plant based on gasification of petroleum refinery residues is universally recognized as the best way forward for the oil industry. Moreover, for advanced power generation (APO) from coal, informed opinion is divided between pursuit of p. firing with advanced steam turbines or gasification. Within gasification itself a consensus does not appear to exist as to whether an oxygen or air-blown system is superior. However, a recent survey undertaken on behalf of the Department of Trade and Industry (DTI) as part of the Cleaner Coal Technology Programme (Project Summary 245) concludes that a p.f. power plant with Flue Gas lyesulphurisation (FGD) is the most economic of five systems evaluated, embracing a range of supercritical and subcritical steam cycles together with integrated gasification combined cycle (IGCC3 options. The hybrid cycle, referred to as the Air Blown Gasification Cycle (ABGC) was found to be the second most economic. However, an earlier report from me Clean Coal Power Generation Group (Coal R089) draws attention to me disadvantageous position of ABGC wing hot gas cleanup (HGCU) in terms of the relatively high emission level of gaseous pollutants, mainly associated with the ABGC's char combustor, because there will always be a base line emission level emanating from the circulating fluidised bed combustor (CFBC), which is traditionally associated with the ABGC cycle. The same report also reveals that little interest has emerged in STIG for coal gasification purposes. The study referred to above did not consider the emission characteristics

of the five systems evaluated for APG, but it is well known that p.f. firing plus FGD is normally regarded as inferior to the single digit NOx ppm levels expected from gas turbines in the future. Nor is the vexed question of the emission of certain air taxies such as mercury from coal fired plant dealt with.

The deficiencies of prior art described above are overcome if Mitsui Babcock's air-blown gasifies is used in conjunction with a supercritical p.f. power plant and FGD. Using the teachings of the present invention this becomes entirely feasible and is a cost-effective means for substantially upgrading the power output of an existing p.f. plant and its associated advanced steam turbine system, provided the gas turbine downstream of the Mitsui Babcock air-blown gasifier can absorb all the steam generated in the Heat Recovery Steam Generator (HRSG) and elsewhere in the gasification circuit. The char produced by the gasifies, perhaps with some simple mineral beneficiation, substitutes for part or all of the coal feed to the p.f. plant. Under these circumstances the gasifies substantially increases the overall power output and very importantly is seen to be a means for eliminating air toxic emissions such as mercury and other metals and trace elements, which in the future will increasingly pose problems for conventional coal-fired p. plant.

Clearly, the gasifies itself has to be arranged so that its associated sulphur and nitrogen oxide emissions as well as air taxies are themselves negligible. For the p.f. plant and its FGD with low NOx burners will be retained and may possibly in certain cases need supplementation by Selective Catalytic Reduction (SCR) to control NOX emissions to extremely low levels. In the very much longer term, say post 2030 or 2050, climate change and global warming concerns may by then force the introduction of near to zero gas emission technologies, but in the meantime and possibly in preparation for realization of the hydrogen economy, industrial experience gained with equipment such as the Babcock Mitsui gasifier may prove to be invaluable as the possible precursor to hydrogen production from fossil fuel sources in conjunction with carbon dioxide sequestration.

Particular examples of the invention will now be given as applied to electricity generation based on fuel gas derived from gasification of carbonaceous materials such as coal, coal char, petroleum coke or biomass, uniquely characterized by the use of closed loop melt circulation of a liquid metal, preferably molten tin in most cases, to effect either gasification and hot gas cleanup together, or alternatively to use other gasification technologies which benefit substantially by incorporation of closed loop liquid metal systems within the overall power generation circuit.

Figure 1 is an overview of the general approach adopted to secure the principal objective, which for advanced power generation is provision of electricity at the lowest possible unit cost, both economically and in teens of cost to the environment.

Figure 2 compares conventional STIG with the proposed cycle. The need for partial oxidation of reheated fuel gas, fuel gas equilibration prior to steam injection followed immediately by transmission of the quenched highly preheated gas with minimal residence time to the gas turbine combustors, is to combat a potential NOX problem.

Figure 3 shows the overall arrangement for power generation with a gas turbine in conjunction with a liquid metal transport gasified and a closed loop liquid metal circuit serving a number of ancillary systems for enhancing overall performance.

Figure 4 shows the overall arrangements for power generation with a gas turbine in conjunction with a supply of high temperature raw fuel gas from a commercially proven gasifies and a closed loop liquid metal melt circulation circuit serving a number of ancillary systems for enhancing overall performance.

Figure 5 defines key locations for flow conditions and chemical compositions, evaluated using the Outokampu HSC computer program.

In Figure I the options illustrated reflect, firstly, the traditional long-cherished view that hot gas cleanup (HGCU) is the way forward for an air-blown gasification system in order to ensure high thermal efficiency and then, secondly, the proposed approach, involving a liquid metal transport medium. The key issue is whether or not a liquid metal system of the type shown schematically in Figure 2 can realistically compete with the traditional approach. The conclusion reached is that better results stem from incorporation of a liquid metal quench (LMQ) to permit ultimate removal of fine particulate solids by conventional filtration rather than using ceramic candle filters, for example, perhaps of dubious reliability for continuous high temperature/pressure operation. This is then followed immediately by fuel gas re-heat to high temperature using dispersed contact between the fuel and the liquid metal in direct countercurrent heat exchange to maximise cycle efficiency.

Referring now to Figure 3, carbonaceous solid materials such as coal, char, petroleum coke or biomass either individually or in association with each other are introduced onto the surface of a molten tin bath I via a barometric leg (not shown) of molten tin containing the entrained solid feed. As viewed in the diagram the liquid metal bath is flowing from right to leR and thus it transports the floating layer of deposited solids as a raft 2 initially into me top blown gasification zone comprised of an array of top blowing jets 3 or perhaps transverse slots which impinge a mixture of air and steam, previously preheated within the body of the horizontal gasified vessel 4 by an arrangement of pipes S receiving direct radiation from the slag sheet 6 floating down the gasifier, in an analogous fashion to the glass sheet produced in the float glass process. The ash associated with solid feed ultimately melts as gasification of the feed progresses and forms a ribbon of molten slag at the so-called equilibrium thickness, which solidifies as heat is extracted from it by the radiant air/steam preheater and it is ultimately continuously withdrawn from the pressurized gasifies using water or air cooled rolls (not shown) which exude the semi-solid slag sheet intact out of the gasifies or alternatively, the solidified slag sheet is lifted office molten tin bath in me same fashion as the glass product in the float glass process, within the gasifies and then fragmentised, perhaps via thermal shock chilling or by mechanical means before being withdrawn from the gasified either as a slurry mixed with water or as comminuted solid material wing a lock hopper or similar device (not shown). The molten tin enters the gasifier bath at 8 directly into bath from the liquid metal quench tower 9, in which the very hot fuel gas enters via the hot-gas omake 10 and is quenched by direct countercurrent contact with the liquid metal irrigating a packed bed (preferably a moving packed bed) with liquid tin entering at the top via an expel closed loop liquid metal circulation system 12. The quenched raw fuel gas then proceeds to a filter system 13 of conventional design at a relative low temperature before passing to the base of the fuel gas reheater 31, another direct contact liquid metal irrigated packed bed, in which the filtered gas is contacted with "superclean" tin, which has been refined in-line within the gasified circuit in a countercurrent liquid metal irrigated packed tower with 8 strip gas at reduced pressure, and referred to as the vacuum desorber 14. This desorber is serviced by inlet 15 and outlet 16 barometric legs with a small amount of inert gas (not shown) injected into the upleg to provide a liquid metal pump action based on the air- lift principle. The associated vacuum pumping equipment and ancillary systems are not shown but are described in co-pending application PCT GB2003/0030S5. To ensure that virtually all of the liquid metal flows through the vacuum desorber, an overflow weir 17 is positioned such that the slag sheet passes across this weir without restriction in association with a small reverse flow of molten tin induced locally in the immediate vicinity of the overflow weir, by returning a small portion of the downleg flow of refined tin back to the top of the broad-crested overflow so that some of it flows counter to the liquid slag sheet, which is itself is pushed forward over the weir by the "ram" effect of the advancing raft of solid charge material floating on the melt well upstream of the broad-crested overflow weir. This 'in force" propelling the liquid slag over the weir (in the absence of bulk tin flow in this direction in the unmediate area of the weir) is the net force pushing the slag Coward, and comes originally from the drag on the submerged solid raft surfaces by the circulating tin over a substantial length of the bath upsilon. This drag force is very much larger than the interracial frictional resistance between the two liquid layers as they flow counter to each other over the weir. This results in small amount of tin being returned to the gasification zone rather than proceeding downstream with the bulk of the tin flow, in order to prevent contaminated tin mixing with "superclean" tin prodded in the vacuum desorber 14. Similarly, provision of some baffling (not shown) of the gas space and a small flow of purge, either inert or sulphur-free gas, prevents sulphur containing fuel gas entering gas phase region 18 downstearn ofthe broad-crested weir 17 and thereby corroding the metallic alloy air/steam radiant preheater tubes 5 as well as re-contarninating any exposed molten tin surfaces in this region. An overflow arrangement (not shown) at 19 permits a continuous withdrawal of "superclean" tin from the bath to allow it to proceed to the top of the fuel gas reheater 31.

Continuing with reference to Figure 3, fuel gas is re-heated after filtration at 240 C or thereabouts to a preheat temperature in the region of 7S0 C or possibly higher, depending on the thermodynamic oxygen potential of the particular fuel gas, by direct contacting in the packed tower 31 with molten tin normally admitted at the top at a temperature around 800 C or possibly higher, again depending on the oxygen potential of the particular fuel gas. The re-heated fuel gas is then partially combusted in a refractory-lined chamber 20 to at least 1100 C and possibly to a much higher Shoe level depending on the particular requirement, using hot air extracted from the gas turbine compressor with sufficient hold-up time provided so that equilibrium is closely approached. The partially combusted fuel gas is then quenched by massive steam injection at 21 and promptly transmitted to the can-annular combustors 22 of a modified Series H General Electric (GE) gas turbine 23, for example, which incorporates closed-loop steam cooling. By these means partially pre-combusted fuel gas is delivered to the gas turbine combustors in a highb preheated state with the 1100 C, or the very much higher temperature level referred to previously, thermodynamic equilibrium composition effectiveb frozen, while the gas itself is at the maximum temperature feasible for the heat resistant alloy tubes or piping, currently considered to be in the region of 870 C, but with further development of ODS alloys confidently anticipated to be in excess of 900 C. Also it is anticipated that the maximum combustor exit and turbine inlet temperature of 1430 C for the current Series H GE gas turbine may in the near term probably be upgraded a maximum combustor temperature of 1500 C as the target turbine inlet temperature. Air admission to the combustor from the gas turbine compressor is then normally the combustion temperature moderating means rather than further steam injection into the combustors themselves.

In the present case the injection of steam upstream of the canannular combustors is merely to permit the equilibrated partially pre-combusted fuel gas to be transmitted to the can-annular combustors using metallic tubing at relatively high velocity so that the residence time is too short to permit significant gas compositional changes to take place prior to full combustion within the combustors. Under these conditions, the very considerably diluted fuel gas has a very low heating value so its entrance into the combwtors at a relatively high temperature is essential to swtain stable combustion. Also as a direct consequence of the exceptionally high level of fuel gas preheat, combustion in stateof-the- art gas turbines can advantageously take place close to the stoichiometric oxygen requirement in the air from the compressor without the need of a wasteful vast excess air being compressed by the gas turbine and its associated energy consumption penalty.

The gas turbine exhaust flows into the Heat Recovery Steam Generator (HRSG) 24 with normally only a single pressure level steam requirement at a relatively lower pressure level (say typically 28 bar) than is customary for a gas turbine with a pressure ratio of 23 and therefore inherently a less expensive option than that required for a combined cycle HRSG. Because the massive steam utilisation involved in bringing back to still a highly preheated fuel gas temperature above 800 C aRer prior equilibration at a very much higher temperature level in advance of transmission to the combustors 22, recovery of water from the turbine exhaust gases would normally be mandatory. Accordingly, in Figure 3 the exhaust gases must next flow to a condensation system 25, preferably a random packed tower or similar device. After requisite water treatment and deaeration, the recovered water is returned to the HRSG for recycling via line 26. Most of the steam generated in the HRSG 24 is returned to the outtake 27 from the gas equilibrator via the line 29 with only a minor amount routed to the gasifies by line 30 to be mixed with the gasifier's air requirements, extracted from the gas turbine compressor and appropriately increased in pressure by the booster compressor 40. The air and steam is mixed just in advance (not shown) of the radiant preheater system 5 within the liquid metal transport gasifies 4.

Again with reference to Figure 3, the refractory lined chamber 20 has two zones in series. The partial preeombustion zone 32 is preferably left as a void space to facilitate backmixing during the initial partial precombustion, followed then by the equilibrator, which is charged with solid packing 33 in recognition of the heterogenous nature of the ammonia decomposition reaction. At temperatures in the region 1100 C to 1300 C, the equilibrator will be packed with ordinary commodity market carbonyl nickel pellets, whilst at higher temperatures clearly a suitable refractory material will be needed.

Referring now to Figure 4, the liquid mete} transport gasifer of Figure 3 is replaced by a commercially proven gasified (now shown) to which is integrated the closed loop liquid metal circulation technology of the present invention. Tfthe gasifer delivers raw hot fuel gas at a temperature say 1200 C or greater, then this gas is hot enough to feed the liquid metal system. Otherwise, a simple partial pre-oxidation of the raw fuel gas can be used to elevate its temperature in advance of the scheme depicted in Figure 4. Figure 4, therefore, starts with a supply of raw fuel gas 34 at an appropriate temperature level to be fed directly to the LMQ 9, an integral part of the closed loop melt circulation system 12, but requiring in this case a pressure seal pot 35 or a pressurized liquid metal sump, which takes the place of the molten tin bath 1 of the gasifies in Figure 3. Regarding the vacuum deorber, the barometric leg 15 and 16 of Figure 3 are arranged somewhat differently in Figure 4. The upleg 15 is basically the same as that in Figure 3, requiring a minor amount of inert gas injection (not shown) to promote air-liR pump action as was the case previously. The arrangement for liquid metal discharge from the Vacuum Desorber 14, however, the down-flow system is now distinctly different. The short downleg 36 is a simply barometric leg accommodating a pressure differential of about I bar, because the radiant Liquid Metal Steam Generator (LMSG) operates with a tin bath or launder at atmospheric pressure to facilitate construction and to simplify any maintenance work associated with the LMSG in later operations. The downleg 12 dips into the liquid tin in pressure seal pot 37 or pressured liquid sump and the liquid tin overflows at 38 via line 39 into the fuel gas reheater 31. The filter system 13 is the same as that for Figure 3 and indeed all the remaining systems are identical to those in Figure 3. These descriptions will not be repeated here, but for clarity the appropriate items are numbered and added to Figure 4.

In effect, the steam quenched partially precombusted gas is the highly pre-heated fuel gas input to the gas turbine with its low ammonia content frozen in so that ultimately low NON emissions from the gas turbine are facilitated.

Example t

Cycle emcieney comparison of liquid-metal based coal gasification/power generation with other advanced technologies In the UK and throughout the world, the majority of electricity is generated by fossil-fired steam boilers in association with steam turbines, typically with steam conditions of 160 bar/568 C/568 C and with overall plant efficiency of 40%. Current state-of-thart plant has steam conditions 300 bar/600 C/600 C/600 C with an overall efficiency of 47.5%. The European Union through its Thermie 700 C Project has a target net efficiency within the range 53-55% for advanced pulverized coal-fired plant, but this is going to require massive development of the necessary materials technologies, based principally on the replacement of iron- based alloys by nickel-based alloys to attain the stated-objective steam conditions of 375 bar/700 C/720 C/720 C by 2020.

For a typical UK coal, CRE Group Lid (CRE) ARACHNE computer programme for Integrated Gasification Combined Cycle (IGCC) yields an efficiency (LHV) of 49.0% for the air-blown gasification cycle (ABGC) base case involving both a high pressure spouted bed gasifies and circulating fluidised bed combustor (CFBC) for second-stage char gasification at atmospheric pressure with limestone addition to the CPBC for sulphur fixation. Under comparable conditions, the Shell Coal Gasification Process (SCGP) base case for IGCC is projected to have an efficiency (LVH) of around 46.8%, based on air separation for oxygen production and conventional lowtemperature fuel gas desulphurisation.

A capital cost breakdown for the base case ABGC, indicates that elimination of the CFBC and the Steam Turbine would reduce the capital cost for ABGC by about 32%. On this basis alone, an airblown liquid- metal based advanced power system employing Crib a gas turbine, may reduce the ABGC costs based on coal gasification so that they are more competitive with advanced steam turbine/pulverized coal power generation. Therefore, in line with the stated objective to reduce the unit cost of electricity, the thermal efficiency of a liquid-metal based coal gasified, in conjunction with massive steam addition in advance of gas turbine's combustors, will now be evaluated, rather than the very much more expensive option of IGCC with both steam and gas turbines. The aim is to see whether comparable or even superior efficiency can be attained for a liquid-metal based air-blown gasification system.

In projecting future efficiency for a new power cycle, it seemed prudent to incorporate the considered views of experts in what should constitute the 'ground rules". Just such an opportunity is afforded by Bannister et al. (Final report on the development of a hydrogen-fueled combustion turbine cycle for power generation. Trans. ASME, J.Eng. Gas Turbines and Power, 121, 1999, 38 45) outlining the final report on the development of a hydrogenfuelled combustion turbine with full-scale 500MW demonstration plant scheduled for 2020. In the meantime, these authors also considered a near-term plant, implying that the technology applied is an adaptation of current technology requiring only moderate development effort. A selection from their quantitative parameters, which are relevant to the current evaluation, is given in

Table 1.

For the near-tenn, an assumption of 93% isentropic efficiency for a preliminary appraisal for an advanced gas turbine is a projection of future capabilities that may be possible with improved aerofoil design.

Bannister et al. stipulate open-loop steam cooling whereas, it is considered that closed-loop steam cooling of 1500 C Class gas turbines is already state-of-the-art. Also the assumed value of 99.2% for generator efficiency is about 0.2 percent greater than the then current (1999) generator technology.

A UK power station coal has been used in detailed evaluations based on the analysis given in Table 2.

Table 3 gives some of the key dimensions and calculated parameters for the proposed gasifies. To put these into perspective it must be noted that they refer to a plant with an anticipated net power output approaching 550MW. For comparison and to give an indication of the physical sizes involved in current gasification technology, the 300MW coaVgas fired power plant at Polk County USA was reported in the trade literature to have taken delivery of a gasifies radiant cooler weighing 470t having a shell with dimensions of 42.4m x 7.0mx7.1m.

Table 4 is an overall mass balance. The small discrepancy between inputs and outputs is thought to be due to rounding off errors in preliminary calculation. Table 5 contains heat balances on the gasifies involved in the present discussion.

In Table 5, the calculated temperature of the air/steam at point of entry into the top-blow nozzles 1098 C indicates that ceramic materials are required for the higher temperature end of radiant air-steam preheater.

However, the join between heat resisting alloy tubes and ceramic tubes, probably silicon carbide, is within the gasifies itself and has to withstand no more than 1.5 bar pressure differential, the velocity head loss envisaged across the jets, so this should not pose any real technical difficulty, especially as a small leakage could be tolerated without serious consequences, but merely would be manifested in a reduction in gasifies performance. The calorific value of the fuel gas produced by conditions selected for Table 5 with a molar ratio of carbon monoxide to carbon dioxide equal to 2.5 is approaching me minimum value recommended for industrial gas turbines. However, this is not a real issue, because by increasing the molar ratio, higher heating values can be obtained, but this is at the expense of melt surface area required, because of the lower levels of CO2 and H2O and thus diffusive driving force for gas phase mass transfer. Clearly, this aspect requires optimization with respect to overall cost. It should be appreciated that the detailed calculations that now follow regarding cycle efficiency ate based on a molar ratio of carbon monoxide to carbon dioxide of nominally 5 to I rather than the 2.S figure used in Tables 3, 4 and 5.

Besides quantitative data of the type already assembled in Tables 1 to 5, additional parameters need to be evaluated in order to estimate the cycle efficiency of Me proposed arrangement. Table 6 summarizes the end results of relatively straightforward calculations needed to progress with the efficiency computation, in this case based on raw fuel gas leaving the gasifies with a molar ratio of carbon monoxide to carbon dioxide equal to 5.0.

Table 7 is a compilation of flow conditions and chemical compositions, evaluated using the Outokumpu DISC computer programme for the key locations defined in Figure 5.

Table 8 gives details of combustion calculations for a 1 S00 C Class gas turbine, in this case with a combustor exit temperature of 1448 C based on 20% excess air above the stoichiometric requirement, while Table 9 evaluates the cycle efficiency.

The cycle jwt outlined with steam quenching at the rate of 9.10 kmol/s is not necessarily the optimum condition. To make available the amount of steam required for this case requires relatively demanding pinch points and approach temperatures. More generous temperature driving forces for steam raising are obtained if steam injection is at the rate of 8.50 kmol/s, implying a precombustion and equilibration temperature of 1100 C prior to steam quenching back to the 870 C level, before transmitting the partially reacted but low ammonia (NH3) fuel gas to the can- annular combustors of the gas turbine. The corresponding net cycle efficiency is then calculated to be 52.23%, a small but perhaps significant gain over the previous figure of 51.65%. However, chemical kinetics may still exert an influence on equilibration at 1100 C, whereas at 13S0 C this is much less likely. Accordingly, for the purposes of preliminary evaluation of feasibility, a net efficiency of S2% is probably a reasonable estimate for the proposed cycle. This is about 2 percentage points above, that projected for ABGC with HGCU and some 5 percentage points higher than the commercially proven SCGP standard oxygenblown cycle.

Nitrogen oxide emissions Regulations relating to smog-causing nitrogen oxide (NOX) emissions are becoming increasingly more stringent. Currently, GE's dry low NOX combustion technology reduces NOX emissions to less than 10 ppm (on a dry 15% oxygen basis). In the US the utility-scale Advance Turbine Systems Program (ATS) objectives for operation on natural gas include 60% efficiency and NOX emission levels less than 9 ppm.

The followon proannne is called the Next Generation Turbine Systems (NGTS) . Included in the NGTS goals are references to coal-derived fuels and in general the reduction of NOX and CO emissions to less than 5 ppm to ensure permitting in the post-2008 time frame.

The NOx emission characteristics of the proposed cycle will now be reviewed briefly. The principal precursor to NOX formation is fuel-bound nitrogen, which, in conjunction with the gasification method, determines the NH3 level in raw fuel gas. The NH3 in reheated fuel gas from liquidmetal based coal gasification will be unacceptably high, so steps have to be taken to reduce this prior to admission to the gas twbine combustors. With oxygen blown-systems and low nue desulphurisation technology, such as SCGP, this is not an issue but in the ABGC case with HCGU, reduction in NH3 level is mandatory.

CRE's ARACHNE Programme, for example, indicates NH3 levels in the vicinity of 1,000 ppm can be anticipated for the fuel gas going to the gas turbine for the ABGC base case.

Partial precombustion followed by gas equilibration is the preferred choice for the proposed cycle to ensure low levels of NH3 are transmitted to the gas turbine combustors. At gems around 1350 C, it is anticipated that NH3 reduction kinetics will be favourabk and expensive nickel catalysts developed for NH3 reduction up to say 900 C will not be necessary. However, it seems prudent to accept that even at 1350 C the reduction chemistry will take place heterogeneously and consideration should be given to the provision of chan nickel surface area provided by packing the equilibrator with Mond Process carbonyl nickel pellets, as produced by lNCO at the Clydach Refinery in Wales, as their main straun primary nickel product. Unlike nickel catalysts, this material does not command a premium above the London Metal Exchange (LME) nickel price. Taking into account the published activation energy (135 kJ/mol, Wang et al. (Kinetics of ammonia decomposition in hot gas cleaning. Ind. & EngChem.Research, 38, Nov 1999, 4175-4182) for the NH3 reduction reaction, the chemical rate constant at 1350 C is some 4.10 thnes greater than that at 1150 C so a shin towards diffusion control can be expected as me temperature is increased to the higher figure. Therefore, at 13S0 C the principal rate controlling influence in a mixed controlled kinetic model is very likely to be gaseous diffusion of NH3 from the bulk gas to theextended surface of a packed bed of Ni pellets, say nominally 10mm diameter. Accordingly, me required dimensions of me equilibrator can be estimated from published mass transfer correlations. For the present case, an acceptable degree of NH3 reduction should be obtained within a squat packed bed 4m diameter by 1m in length, giving rise to a gas-phase pressure drop in the region of 1 bar.

For the case evaluated in Example 1, Table 10 compares me equilibrium level of NH3 for equilibrated gas at 1350 C wim that attained, if the gas were allowed to come back to equilibrium during its transmission at 870 C to the gas turbine annular combustors. At 1350 C the NH3 level is less man 30 ppm, whereas at 870 C the level increases tol40 ppm. Clearly, the aun should be rapid quenching by steam immediately followed by as short a time as possible for transit between the equilibrator and the gas turbine combustors, so that the gas composition is effectively frozen close to the 1350 C condition.

Table 11 shows mat the expected NOX concentration is less than 5 ppm, if a conversion ratio of 0.45 for NH3 to NOX is assigned. In practice, the conversion ratio is likely to be considerably lower than this, as evident fiom data presented by Hasegawa et al. (A study of combustion characteristics of gasified coal fuel, Trans. ASME, J. Eng Gas Turbines and Power, 123, 2001, 22-32) for fuel gases with low calorific values compared with natural gas. On a dry basis this predicted NO,, emission level corresponds to 2.85 ppm(v) at 15% oxygen.

Example 2

An embodiment of the invention will now be considered in which the commercia11y proven Shell Coal Gasification Process (SCGP) is coupled with a liquid metal closed loop melt circulation system in over to provide a lower overall cost alternative to the Integrated Gasification Combined Cycle (I()CC) traditionally associated with SCGP. The objective is to considerably lower the capital costs by removing He need for a Steam Turbine and at the same time substituting liquid-ruetal based fuel gas cleanup to remove sulphur and the various air toxic elements, particularly mercury, normally associated with coa1-fired combustion, by using instead elevated temperature treatment involving desorption of stannous su1phide gas under reduced pressure from the liquid metal at temperatures in excess of 1100 C. In the LMQ the liquid tin absorbs sulphur compounds from the fuel gas under reducing conditions as the gas is cooled. The LMQ in this case takes the place of the gas quench currently used for SCGP, which involves recirculating cooled fuel gas back to the gasifies at a rate somewhat exceeding the rate of production of raw fuel gas in the gasifies and then cooling the combined gas stream at about 900 C in a fuel gas cooler, which is a waste heat boiler of conventional design but still rather expensive. After the waste heat boiler in the current SCGP flowsheet, the fuel gas is treated using a ceramic candle filter system to remove particulate solids at about 400 C, prior to further cooling and scrubbing in advance of an extensive cold gas treatment plant, principally for sulphur removal while eliminating certain other impurities of concern at the same time. This is to be compared with the very hot liquid metal exiting LMQ immediately being desorbed of its sulphur and air tox* elements by gas stripping under reduced pressure to continuously produce a refined stream of molten metal, which can be described reasonably as being "superclean". The strip gas containing the stannous sulphide and air toxic elements can be processed in a compact facility to yield a saleable high-grade (low iron) zinc sulphide concentrate, if a metallic zinc-based system is adopted to regenerate in-situ the dissolved tin sulphide back to elemental tin in a independent closed loop system, which precludes any transfer of zinc, whatsoever, to the main fuel gas stream and, therefore, obviates completely any concern about zinc contamination being transmitted to the gas turbine. Alternatively, the fuel strip gas can be cooled and filtered of its solid stannous sulphide and treated on-site or treated across-the- fence to produce elemental sulphur, employing the chemistry of what traditional metallurgists refer to as a roast-reduction reaction.

This reaction in the same context has been studied by workers at Karlsrube University in Germany (Desulfurization of rnanufacred gases with liquid metals. Int. Gas Research Cot., Toronto, Canada, 1986, 11241133). The stannic oxide product of the roast-reduction reaction can be readily incorporated into the solid gasffier feed for SCGP, eventually being discharged from the SCGP gasifies as gaseous stannous sulphide in the raw fuel gas to be absorbed into the closed loop molten tin servicing the LMQ, thereby permitting its recovery in the vacuum desorber as outlined above. This is effectively a total reflu of tin with elemental sulphur production as a worthwhile by-product. Accordingly, it is ciaiTned that this direct production of elemental sulphur has many advantages over the petroleum industry-based technology currently associated with SCGP.

Calculations similar to those outlined in Example I can be carried out, and accordingly it can be shown that a single item of turbo-machinery, say a rnodiffed series H GE gas turbine, can be anticipated to generate more net electrical power than that currently produced at the commercial installation of SCGP at Buggenum in Holland employing IGCC.

By using massive steam quenching of a fuel gas at a very high temperature in advance of the gas turbine combustors in accord with the teachings of this invention, single digit ppm levels of NOX based on 15% oxygen can be anticipated confidently with the proposed technology.

Also because of its volatility, mercury can be recovered in the elemental state by controlled cooling of the exit strip gas as it is compressed from the reduced pressure level back to atmospheric pressure by the vacuum pumping system. Other air toxic elements are also prevented frown discharging into the environment, but rather are absorbed into the molten tin, necessitating periodic partial removal by conventional pyrometallurgical refining techniques, either in-situ, acrossthe-fence or further afield by a metallurgical processor in return for a treatment charge arrangement for partial replacement of contaminated tin on a periodic basis with refined high purity tin.

Massive steam addition to fuel gas in advance of the gas turbine combustors implies a requirement for relatively simple water recovery measures. For the SCGP embodiment of the invention, a power plant generating a net output of about 400MW of electricity would require about 140kgls of superheated steam at 460 C to quench the partially oxidised fuel gas equilibrated at say 1735 C down to a temperature of 835 C before rapid transmission to the combustors of the gas turbine. For a typical imported steaming coal, say El Cerrejon from Columbia, the turbine exit gas will contain a total of about l 59k0s of water vapour produced by combustion of the steam quenched highly preheated fuel initially derived from gasification of El Cerrejon coal in the oxygen/blown gasifies, and for this case, combustion with 10% excess oxygen above the stoichiometric requirement from air compressed by the air compressor on a single shaft turf machine, ensures that the turbine entrance temperature is at about 1500 C. To condense 99% of this water vapour in a direct contact randomly packed tower produces more than enough demineralized water for the massive steam addition involved. In certain arid areas this excess fresh water may be a positive attribute.

Downstream of the heat recovery steam generator (HRSG), the system is designed such that the gas phase pressure drops are minimal so the effects on cycle efficiency are relatively insignificant without the need for induced draft fans. In the present case, the plant requirements to reach the target figure of 99% condensation of inlet water vapour can be estimated using the procedures outlined by Ralph F. Stfigle, Jr, in his text "Packed Tower Design and Applications" 2n edition 19g4. In the preface to this book, it is stated that the procedures described therein have been used for over a decade to design towers as large as 46ft in diameter in the petroleum industry and it is noted that flue gas desulphurisation packed towers exist 40ft in diameter. For the present case, a single 55ft diameter tower (16. 7m) is indicated, packed to a height of 27.6ft (8.4m) with 3.5 inch plastic Pall Rings. The total pressure drop over the packing works out at just over 4 inches water gauge. If the water fed to this unit is at a temperature of 25 C and the exit water temperature at 48 C, for gases leaving the HRSC} at 120 C with a water vapour content equal to 0.55 mole fraction and a dew point of about 87 C, cooling the water back to 25 C to irrigate the packed tower condenser is a major consideration and without access to readily available cool water, this is an energy intensive operation, requiring, for example, somewhere in the region of at least 20MW installed fan capacity for a water to air cooler, based on extrapolation of infonnation given by De Paepe and Dick (Water recovery in steam injected gas turbine: A technological and economical analysis, European Journal of Mechanical and Environmental Engineering, v 44, n 4, Winter 1999, p 195-204). Clearly, location of the power plant at a coastal location would be highly desirable and then all that would be required to cool the recycled water for the condenser is a conventional water to water tubular heat exchanger system. In this context and for the present example, it must be borne in mind that approaching 430MW of thermal energy would have to be dissipated and the environmental consequences of this would need to be carefully considered.

In this discussion it has been assumed that a visible plume of water vapour emanating from a conventional cooling tower in the future will not normally be acceptable both from the aesthetic viewpoint as well as the wastage of fresh water involved. Accordingly, the above scheme will need to reheat the flue gas after condensation of most of its water vapour content. Again bearing in mind the need to minimise gas phase pressure drops downsteam of the HRSG to preserve cycle efficiency, this flue gas reheating can be accomplished advantageously using direct contact between the condensation tower discharge gas and an extended surface area of the circulating liquid metal in what the inventor has elsewhere coined the phrase a "swimming pool" reactor in the context of continuous steeknaking (PCT GB2003/003069), analogous to a float glass bad, which also uses tin but in this case with induced melt circulation as opposed to the static tin bath in glass technology.

At low temperatures the reaction between liquid tin and oxygen in air comes to a halt once a thin coherent layer of oxide is formed and this is referred to as passivation. In the metallurgical industry, as a matter of general principle, molten meml handling systems should be designed to eliminate, where possible, splashing and surface turbulence. For example, launder transfer from aluminium holding furnaces to tundishes is standard practice in direct chill casting of aluminium. Also in the zinc industry, where molten metal may have to be transferred from a zinc blast furnace to a refluxer, zinc crossing can occur in the tapping of a holding bath to fill a ladle and subsequently on pouring from the ladle to the feed bath of a refluxer for refining. The alternative to this practice is to transfer the molten zinc via a rather long launder on a continuous basis. It is therefore essential to be able to design a launder system, which will ensure smooth flow and minimum crossing in operation. Similar considerations also apply in the condenser cooling launders of an Imperial Smelting Furnace (ISF). Considerable progress towards rational launder design was made by R Healey at Imperial Smelting Corporation, when he developed a theoretical model of the conditions required to stabilize a very thin oxide layer or flhm on the surface of metal transfer launders associated with an ISF. A stabilized dross layer is maintained by its adherence to the side of the launder.

The strength of the adhesive bond must be sufficient to withstand the drag force resulting from fluid friction. The same considerations apply in the present case.

In the present example, the cooled flue gas leaving the condensation tower is at about 40 C and to avoid any possibility of plume visibility from the stack emission, this is reheated to say 100 C, for a conservative design estimate. This operation involves flowing the flue gas above a molten tin surface with its passive oxide film in tact, countercurrently to the return tin flow from the fuel gas reheater on its way back to the LMQ in a "swimming pool reactor" - sized contactor. A first estimate of the requirement indicates that for the present example some 71 5m2 of molten metal surface area is needed, so the "swimming poor' would need to be about 10 meters wide and 71.5 meters in length. In this connection, it is relevant to note that a modified float glass chamber has been proposed by Serth and co-workers in the USA (Recovery of waste heat from industrial slags via modified float glass process. Energy Communications, 1981, 7(2), 167-188.) for recovering energy from industrial molten slags. These workers give the dimensions of a typical float glass molten tin bath as 7.3m wide by 55m in length. Assuming a tin depth of 5cm is adequate, the interest lost on the capital tied up in the meml associated with a swinunmg pool of the size estimated and with an interest rate of 5 percent p.a., the extra cost of electricity works out at less than 0.00lp/kWh.

The associated gas phase pressure loss in t'ne "swimming pool,' reactor is estimated to be between 4.4 inches water gauge and 2.7 inches water gauge, covering Fanning friction factors extending from completely smooth surfaces through to maximum roughness, and taking an arithmetic means this yields an estimated pressure loss less than 4 inches water gauge.

In addition to the other relatively small frictional losses already identified, the ejection of the reheated flue gas, according Desideri et al. (Water recovery from Hat cycle exhaust gas: A possible solution for reducing stack temperature problems, International Journal of Energy Research, vol. 21, p 809-822, 1997), ideally requires a gas velocity of 1. 2 to 1.5 times the maximum windspeed in the geographical area, so this adds a further 3 or 4 inches water gauge velocity head loss to the pressure drop in the overall exhaust system from the HRSG to the point of discharge into the atmosphere.

Provision needs to be made also to keep the molten tin bath above the melting point in case of temporary interruption of the plant or indeed to assist in bringing the plant on-line for the first time. This can be achieved conveniently, for example, by use of electrical conductive heating in a manner analogous to that used in pilot-scale smelting trials at the University of Birmingham (Warner, N.A. Trans.Instn.Min Metall. (Section C: Mill Process Extra. Metall.) 2003, 112, No. 3 December in press).

Example 3

A further embodiment of this invention is its potential application in association with the extensively researched but not yet commercially proven Air Blown Gasification Cycle (ABGC) proposed initially by British Coal and referred to formerly as the British Coal Topping Cycle. It is now suggested that the basic ABGC arrangement could be made more competitive with conventional pulverised power plant and commercially more attractive, if the development were focused on the spouted fluidised bed gasification technology that is the central component of the hybrid ABGC system. Without the separate atmospheric pressure circulating fluidised bed char combustor and without the steam turbine, it is believed that more than one third of the total capital cost could be saved. Rather than using a dedicated char combustor system, a need exists for exploring the benefits of co-firing the char with coal in a pulverized fuel (p.f.) boiler system, probably generating supercritical steam for advanced steam turbine power generation involving double reheat and elevated temperature operation moving towards the EC's Thermie Project, which may ultimately lead to efficiencies in the region of 53 to 5S%. However, in no way should the p. system be directly integrated with the spouted bed technology referred to above. This is now owned by Mitsui- Babcock Energy Ltd. in the U.K. and should be developed in a stand-alone mode. This air blown technology is ideally suited to enhancement using the liquid metal systems of the present invention and its char product could be stockpiled or fed continuously to an associated p. plant. By these means some existing power stations could be retrofitted win the new technology to substantially increase He overall power output at a cost less than installing a full new ABGC plant. This provides a powerful combination of air-based technology rather than relying on the more expensive option of air separation for oxygen blowing and perhaps more importantly can provide a solution to the inherent problem of air toxic emissions, including mercury, that increasingly can be anticipated to shape further development and to affect planning consents for the installation of new coal-firod plant in the future. Also a single advanced gas turbine such as the General Electric Series H could be modified to receive steam quenched highly preheated partially oxidised fuel gas equilibrated at say 1100 C or above before massive steam quenching with superheated steam generated in the HRSG and also in the radiative LMSG interposed before the fuel gas reheater and immediately after the vacuum desorber for elimination of sulphw and air toxic elements from the closed loop circulating liquid tin circuit. This liquid tin circuit also services the LMQ, fuel gas reheat and flue gas reheat systems as in Example 2.

Again single digit ppm NOX levels (corrected to 15% oxygen on a dry basis) can be anticipated and no flue gas desulphwisation will be needed to meet the strictest emission regulations foresocable in the future.

Filtration of the fuel gas free of particulates is conducted at a relatively low temperature using well-proven commercially available filtration technology, coupled with fuel gas reheating via direct liquid metal contacting, prior to partial oxidation and ammonia reduction by equilibration at say 1200 C using market- grade carbonyl nickel pellets, for example, prior to quenching to above 800 C before transmission ofthe highly preheated fuel gas to the gas twbine combustors as in Example 2.

For this example, a dry feed of U.K, Daw Mill coal as specified in Table 2 of Example 1, but now at the rate of 34. Ikgls (d,a.f.) was evaluated for the present example. The steam quench in the liquid metal associated system worked out at 11 3kgIs of superheated steam at 600 C, generated at the relatively low pressure of 28 bar in both the radiant LMSG and the HRSG, with superheating in the gas turbine's closed loop steam cooling system with the balance in the HRSG. In this case, the steam quenched fuel gas is transmitted to the gas turbine combustors in stateof-the-art heat resistant alloy piping at 870 C and the gases are designed to reach a temperature level of 1430 C in the can-annular combustors for entry into the gas turbine, the present maximum temperature level believed to be acceptable for a Series H - GE Gas Turbine. The gross power generated by the gas turbine is calculated to be 352.3MW. Combustion in air in this case, is assumed to take place close to the stoichiometric oxygen requirement. Work by Hasegawa et al.(Effect of pressure on emission characteristics in LBG-fueled I 500 C<lass gas turbine, Trans. ASME, J. Eng Gas Turbines and Power, 120, 1998, p 481) indicates that high temperature/high pressure combustion just above the stoichiometric requunent is all that is needed to ensure low carbon monoxide emission levels under similar operating conditions of high temperature and high pressure.

In this particular case, it may be advantageous to use oxygen rather than air to effect the partial combustion needed to raise the reheated fuel gas from say 750 C ex the direct liquid metal contactor to the 1200 C level used in this case for equilibration to reduce the ammonia level prior to steam quenching. Oxygen is used in recognition of the fact that the gas exiting the Mitsui-Babcocl Gasifier is expected to be at about 965 C, which is not hot enough to generate the higher temperature levels needed for effective stripping under reduced pressure of stannous sulphide from the molten tin exiting Me LMQ. Accordingly, oxygen was also added for dais initial pre-combustion to raise the raw fuel gas temperature from 965 C to 1190 C in advance of the LMQ by adding 3.2kg/s oxygen. The second oxygen addition to raise the reheated fuel gas temperature from 7SO to 1194 C requires addition of 6.2kg/s oxygen. However, it is by no means certain that oxygen must necessarily be used and a more extensive evaluation may indicate that preheated air is all that is really needed for both partial pre-combustions.

The evaporative steam duty for provision of the steam necessary for quenching in this case is derived from three sources. Firstly, the HRSG, second the radiant LMSG and finally the char cooler. Superheating is conducted principally via the closed loop cooling system of the gas turbine and the balance in the HRSG.

Table I Performance parameters for reference plant Turbines Rotation speed (rpm) 3600 Bearing loss (% shaft power) 0.60 Adiabatic efficiency (%) 93.0 Exhaust losses neglect Steam lealcs (% of inlet flow) 2.0 Shaft leakages neglect Electric Power Generator efficiency (%) 99.2 Pumps Adiabatic efficiency (%) 85 Motor efficiency (%) 98 Mechanical efficiency (%) 98 Table 2 Solids analysis (Daw Mill Coal) Coal proximate analysis, wtYo Daf Coal 76 Ash 1 3 Moisure 1 1 Coal ultimate analysis, wt%(dal) Carbon 8 1.3 Hydrogen 4.9 Oxygen 10.7 Nitrogen 1.25 Sulphur 1.62 Chlorine 0.25 Coal HHV, MJ/ig 32.9 Table 3 Typical calculated performance parameters Fuel Gas at I I OC Equilibrium Flowrate (kmoUs) Composition (mol traction) CO 1.578 0.1825 CO2 0,631 0.0730 H2 1.090 0.1261 H2O 0. 90g 0.1051 N2 4.386 0.5073 Ar 0.052 0.0060 LHV = 3.668 MJ/m3 Air plus steam for top blow Flowrate (kmoUs) Composition (mol fraction) O2 1.166 (227 C) 0.1714 N2 4.386 (227 C) 0.6446 Ar 0.052 (227 C) 0.0076 H2O 1.200 (257 C) 0.1764 TopbIow jet parameters Nozzle diameter 0.030m Distance between nozzle exit and solid surface 0.225m Wall jet radius on solid surface 0.390m Nozzle velocity 206m/s Number of jets 230 Modelling of top- blown gasification zone Non-reaceing jet (dilution by Equivalent hlb reacted jet entrainmeat but no reaction) (equilibrated entrainment) Total Pressure 28 bar 28 bar Mean Temperature Jet 1400 C 2130 C Reynolds Number (Re) 6.5 x 105 6.6 x 105 Schmidt Number (Sc) 0.803 0.803 Holger Martin (Heat and mass transfer between impinging gas jets and solid surfaces.

Advances in Heat 7ransfer, 13, 1977, 31-60) Sh /Sc 0-42 (extrapolated) 720 750 Sherwood Number sh) 657 684 Mass Transfer Coefficient 0.266 rnls 0.571 m/s Fixed Carbon Oxidation Flux 0.411kg/sm2 0.373 kg/sm2 Mean Carbon Oxidation Flux = 0.392 kg/sm2 Estimated bath dimensions for toblow zone: 3m wide x l9m long Table 4 Overall mass balance Inputs to Gasifler Coal (daf) 32.612 'XtYggenn 31,7023.7O8 Argon 21.600 Ttl 221.987 Outputsirom GasiSer plus nitrogen from coal 2o.l$o387 Total Fuel Gas 215.77 S 0.082 Total 221.966 Table 5 Typical heat balances for air-blown gasification with fuel gas re- heat to 730 C for gas turbine combustor 1. pleat balance on gasification zone/gas quench Heat In MW LHV coal 1026. 0 Sensible heat coal (240 C) 9.4 Airlsteam at entry to top blow (1098 C) 239.6 Liquid metal entering gas quench (240 C) l476244 Heat out Liquid metal leaving top blow (1450 C) 675.6 LHV fuel gas 716.6 Fuel gas leaving liquid-metal quench (250 C) 59.0 Molten slag (1 S00 C) 4622 4 2. Overall heat balance on gassier Air (227 )/steam (270 C) 42.0 Liquid metal from gas quench (1169 C) 5102650 Sensible heat of coal (240 C) 196439.9 Liquid metal to fuel gas reheater (750 C) 393.8 LHV fuel gas 716.6 Sensible heat fuel gas to quench (1550 C) 4S6.2 Solidified slag (600 C) 4.5 Allowable heat lo" 68.8 (6.7% of coal LHV) 1639.9 Table 6 Additional calculated perfonnance parameters Tin circulation rate 9.508 kmoUs Tin exit temperature from gasifier 775 C Tin exit temperature from fuel gas re-heater 322 C Number of thermal transfer units for fuel gas re-heater 11.25 Fuel gas temperature after re-beating 750 C Total thermal load for dryinglpreheating coal 16.74MW Low pressure steam generation via tin cooling to 240 C 0.33 knolls Required approach temperature for HRSG 30 C Booster compression for gasified air extraction 22 26.S bar Booster temperature range 156 C 18S C Booster for gasifier air extraction 5.45MW Boiler feedwater (BFW) pumps 30 bar BFW power consumption 0.69MW Pump for liquid tin circulation 2.38MW De-aerator thermal requirement 12.24MW Heat available from air-cooler for gasified air extraction before booster compressor 60.00MW IP steam generation for 15 C pinch point 1.43 kmolls Heat available to steam from closed-loop gas turbine (GT) cooling 63.05MW Isentropic expansion from 1448 C for GT pressure ratio of 23: Entbalpy change at 100% isentropic efficiency without steam cooling 800.06MW Enthalpy change for 93% isentropic efficiency 744.06MW Actual enthalpy after steam cooling 498.28MW Gas turbine exit gas temperature 626 C Air temperature to gas turbine 15 C Air temperature ex GT compressor 504 C Maximum tin oxide gas (Snob) in re-heated Mel gas before pre-combustion 2. 7 ppb(v) Maximum tin oxide (SnO2) carry-over to gas turbine from combustors (mass concentration) S parts per billion Annual maximutn metallic tin discharged to environment 75 kg Tabb 7 Sin key 1 __ _ no. _ o ' of -on go. o _ So -to- _ ___ _ _ _ g In, : U :- to to _4 18 o _ S. -hi. o no, _ -. __ _ _ _ _ _ _ m O O Oo, Y! _ 0:0 r 0 S 80 -; ii) 1 Table 8 Combustion calculations for 20% excess O2 above stoichiometric requirement Quenchedfuel gas at 870 C 69om 1350 C) kmoUs y CO 1.588 0.08767 CO2 0.622 0.03434 H2 0.875 0.04831 H2 O 10. 184* 0.5622 *(9.10 + 1.084) N2 4.788 0.2643 Ar 0.057 3.147 x 103 Total 18. 114 1QQ Stoichiometric O2 = '/2(1.588 + 0.875) = 1.2315 kmoUs Toml 0xygen = 1.478 kmoUs Associated N2 = 5.558 kmoVs After combustion, H2O = (10.184 + 0.875) = 11.059 kmol/s N2 = 4.788 + 5.558 kmoUs = 10.346 kmoVs Ar = 0.123 kmoUs Combustion Gases: kmoUs yi N2 = 10.346 0.43 tO 02 = 0.246 0.0103 C02 = 2.209 0.0921 H2O = 11.059 0.4610 Ar = 0.123 5. 129x 103 Tot = 23.983 0!9995 Comhstion air: (Pressure ratio 23) Exit temperature = 504 C O2 = 1.478 kmoUs N2 = 5.558 kmoUs Ar = 0.066 kmoUs Total = 7.lo2lOVs Combustor Heat Balance

MW

Sensible heat fel gas at 870 C 541.698 Fuel gas LHV rate 661.105 Combustion air at 504 C 102.791 1305.382 Output Combustion gases at 1448 C 1305 81Q Table 9 Ested cycle efficiency Gas Torblae 20Yo Excess O2 above stoichiometric combustionisteam addition = 9.10 kmol/s quenchfrom 1350 C Fuel Gas 750 C 1350 C partial combustion then equilibrationisteam quenchirom 1350 C 87C C H. = 1305.383MW H2 = 498.277MW (H2- H) = 807.106MW Q = 63.050MW Total W = 807.106 - 63.050 = 744.056 Beanag loss (0.6%) = 4.464MW Actual W = (744.056 - 4.464) = 739.592MW For 77p = 0.88 Total Air to be compressed = 13.202 hnoVs Actual Wc = 508.32 W/kg = 14741.28W/hnol Wc actual = (14741.28) 13.202 = 194614/s = 194. 614MW W available for power generation = 739.592 - 194.614 = 544.978MW Gms eleebic power generated at 99.2% generator efficiency = 540.618MW AnciUarv Power

MW

BFW pumps 0.686 Sn pump 2.380 Au booster compressor 5.45 Solids handling (half ABGC) 1.00 Miscellaneow (as per ABGC) 1.16_ 10.68MW NET ELECTRIC POWER = 529.942MW COAL LHV RATE (See Table S) = 1026MW CYCLE EFFICENCY ( /O) = (529.942)100/1026 = 51.65 Table 10 Fuel gas flowates and eqmlibrium concentrations A. Partial pre-combustedfuel gas B. Gasirom A after steam quenching at 1350 C to 87C Flowrate Mole Fraction FlowrateMole Fraction moVs) QcmoUs) N2 4.188 0.S31 H2O 9.059 0. 499 CO 1.588 0.176 N2 4.787 0.264 H2O 1.084 0.120 H2 1.994 0.110 H2 0. 874 0.097 CO2 1.789 0.099 CO2 0.622 0.069 CO 0.460 0.025 Ar 0.0S7 0. 006 Ar 0.057 0.003 NH3 2.591 x 10- 2.874 x 10'5 NH3 2.535 x 10 3 1.397 x 10" H 3.552 x 10'5 3941 x 10 CH 1.378 x 103 7.594 x 10-5 HCN 1.526 x 10- 5 1.693 x 10 HCN 9.837 x 10-7 5.420 x 10 1.1S8 x 10-5 1.285 x 10 H 7.128 x 10 3.928 x 10-9 To 9,Q14 1.000 Total 11. 154 I,Q00 Table 11 NOX in combustion products Re-heated fuel gas flow = 7. 558 kmoVs Re-heated fuel gas temperature = 750 C Partially combusted fuel gas flow = 9.014 kmol/s Partially combusted fuel gas temperature = t350 C Equilibrium temperature = 1350 C Equilibrium NH3 at 1350 C = 28.74 ppmv Combustion gas flow to turbine = 23.983 lcrnolls NH3 concentration by dilution first with injected steam quenching to 870 C and then combustion gases win 20% excess O2 in air over and above the stoichiometric requirement, hypothetically in the absence of NH3 reaction is 28.74 x 9.014/23.983 = 10.80 ppmv Say, conversion ratio of NH3 to NOX is equal to 0.45 as reported for a low calorific gas (Hoppesteyn et al. (Combustion of biomass-derived low calorific value fuel gas, Combustion and Emissions Control III, Inst. Energy, 1997, 293-303), then the expected NOX in the combustion gases going to the gas turbine and indeed exiting the heat recovery steam generator (EDISG)) prior to any condensation of the associated water vapour content is 10.80 x 0.45 = 4.86 gp,V

Claims

1. A gas turbine-based power system enhanced by the provision of hot gas cleanup and thermal energy recovery means afforded by closed-loop liquid metal transport facilities servicing the whole integrated power generation system comprising some or all of the following:- (i) massive steam injection into effectively sulphur and particulate-free high temperature fuel gas or cleaned partly combusted fuel gas to facilitate quenching of the gas after it has had sufficient residence time at temperatures in excess of 1 1 00 C so that the ammonia content of the gas is decomposed heterogeneously in a bed of relatively inexpensive commodity-market product such as carbonyl nickel pellets, (ii) rapid transmission of the quenched fuel gas to the gas turbine combustors presently limited to a temperature of about 870 C but with the further development of oxide dispersion strengthened (ODS) alloy materials at somewhat higher temperatures in the future, so that the gas phase equilibria are not given sufficient time to re-establish but rather remain frozen at the 1 1 00 C or higher temperature level involved in the thermal decomposition of ammonia in order to ensure that the gas turbine combustors are fed with highly preheated fuel gas with a very low ammonia content, (iii) water recovery means commensurate with massive steam injection in advance of the gas turbine combustors, downstream of the heat recovery steam generator (HRSG), from the cooled gases emanating from the water recovery system prior to them being reheated to avoid generation of a visible plume on admission to the atmosphere.

2. A gas turbine-based system as claimed in claim l, in which the raw high temperature fuel gas iirom a commercially proven gasification technology such as the Shell Coal Gasification Process (SCGP) is directly quenched from typically around 1 500 C in a liquid meta} quench (LMQ) as the first step in a closed-loop liquid metal circuit servicing a range of requirements associated with the overall power generation plant.

3. A gas turbine-based system as claimed in claim 1, in which the recently proposed but as yet commercially unproven air-based liquid metal transport gasifies system is adapted in a modified form more compatible with an ultimate zero gas emission scenario than that already the subject of PCT Application GB2003/00355, and thus is oxygen rather than air-based, incorporating the following new features:- (i) improved utilisation of the inherent anti-oxidation shielding properties of the liquid slag sheet by controlling the melt circulation rate so that the liquid slag sheet extends over almost the whole width of the liquid metal bath, closely approaching the refractory walls on both sides without actually coming into contact with them, (ii) provision of anti-oxidation refractory shielding and inert gas purging to prevent oxidation of the liquid metal surface not protected by the slag cover and thus exposed on the small clearance between the circulating liquid metal and the refractory walls as implied in (i) and in addition, making sure that the liquid metal bath level is maintained immediately just below the refractory shielding by appropriate monitoring means and flow-control procedures, (iii) for liquid metal systems employing tin as the preferred choice of liquid metal, limitation of the amount of volatile stannous oxide introduced into highly preheated fuel gas with potential problems associated with oxide accretions on cooler surfaces such as the gas turbine blades, by improving the initial gasification step so that a very low oxygen potential (low CO2 content) gas is produced in the first instance, a direct result of a more favourable heat balance being achievable in oxygen blown gasification once the punitive thermal demand associated with nitrogen is removed from the system.

4. A gas turbine-based system as claimed in claim 1, in which the extensively researched but not yet commercial proven Air Blown Gasification Cycle (ABGC) is re- arranged with further development focused on the spouted fluidised bed component, whilst replacing the dedicated char combustion circulating fluidised bed with a system based on co-firing the char with coal in a pulverised fuel (pf) boiler, probably generating supercritical steam for advanced steam turbine advanced generation and accordingly the spouted bed gasification technology preferably operated in a stand-alone mode, and ideally enhanced with liquid metal systems for controlling emissions of particulates, sulphur, nitrogen oxides and other air taxies associated with coal Quelling and thus provided with a powerful combination of air- based technologies Other then relying the more expensive option of oxygen- blowing, for feeding low ammonia highly preheated fuel gas to an advanced gas turbine in order to reduce nitrogen oxide emissions and achieve increased cycle efficiencies.

- l

- L - _ _

at, _

_ h;. L. r (.

L rat 11 I

1,, _, i _^ _ ..

it_ _ 1'' I , J r V _ 1! U US I 1_ aR Up \ 1