Fuel cell fever favours distributed generation20 July 1998
When symposia proliferate, conferences and demonstrations abound, and new specialist bodies emerge all round the world, it usually means that new technology is ripe for commercial market penetration. If this is still true in the new age of commercially competitive power supply, then fuel cells should be about to burst upon the scene. In practice, much of the hype may still be premature, but a status review and a comparison of competing technologies may now be timely.
It is by no means certain whether more progress is being made in fuel cell development for transport applications or power generation, but in both fields of application the battle for supremacy between rival technologies is far from settled.
Generally, it seems that the newest and least developed fuel cell types show the most promise of commercial market penetration. This could indicate a degree of technological immaturity, but most contenders are now promoting allegedly commercial products for market penetration in the first year or two of the next millennium.
In the power generation scene, it may well be that the approach that offers the most kW for the lowest costs and lowest cell degradation rate and least maintenance requirements will come out on top. In both fields, however, it is clear that there is a great deal more work to be done on materials and manufacturing technology selection, process optimisation and design development before a mature commercial product can be introduced to the market.
As to which of the currently developed technologiesAlkaline Aluminium–air Direct methanol MCFC (molten carbonate fuel cell) PAFC (phosphoric acid fuel cell) PEM (solid polymer or proton exchange membrane fuel cell) SOFC (solid oxide fuel cell)
is at the most mature state of development or closest to commercial viability, will probably depend as much as anything on the successful development of mass production processes that can minimise unit costs.
With the grand inauguration of the Westervoort SOFC demonstration 200 kW prototype on 14 May, 1998, and the hype of the European Fuel Cell Group's excellent workshop in Arnhem the following day, it would be easy to assume that the tubular SOFC had a major lead on the rest of the field, particularly when Westinghouse projected commercial sales of ten advanced 250 kW units in 2001. But we should still make some comparisons. The Westervoort prototype was described in some detail in the May 1998 issue of MPS.
The phosphoric acid fuel cell, which operates at a temperature of 200°C and generates electricity in small packages at 40 per cent efficiency with no noise or vibration, few moving parts and negligible emissions, was the first and most commercially developed variety of fuel cell power plant. They are being used as cogeneration plants in hospitals, hotels, schools, utility power plants and even one airport terminal. Early 11 kW field test units have been operating for some years in countries all round the world. This has contributed greatly to establishing the technical viability of fuel cell power systems, but their ultimate commercial viability is still open to question.
Initial market penetration owes much to the decision by the US Department of Defence to purchase 22 fuel cell cogeneration units from the ONSI Corporation for military bases across the country, in February 1996. The US military had already purchased 11 of these early units. At the same time, Kaiser Permamente, the largest health maintenance organisation in the USA, aimed to install as many as 160 similar 200 kW units at sites all round the country, to be owned by Enron Corporation from whom Kaiser would purchase the power.
Since power generation efficiency in fuel cells generally increases with operating temperature, the solid oxide fuel cell, which can reach operating temperatures of nearly 1000°C, vies with the molten carbonate design, which typically operates at about 650°C, as the best prospect for utility distributed power supply or remote-from-network local power generation applications. Working at temperatures as high as this precludes the need for external reformers to separate the hydrogen from the hydrocarbon fuel – internal reforming occurs naturally at these temperatures, but it also makes the electrodes more susceptible to carbon monoxide attack.
In the SOFC, the internal reforming reaction tends to be excessively productive to the extent that control can become unstable unless some way of buffering the reaction is applied.
Since the solid oxide system uses a hard ceramic material instead of a liquid electrolyte, the SOFC tends to be regarded as the more sophisticated and basically simple concept with hardly any moving parts and none of the plumbing associated with MCFC modules.
Two approaches are being pursued: the Westinghouse tubular electrode configuration, which is the more developed concept; and the flat plate approach favoured by Siemens and Sulzer. Since Siemens is taking over the Power Systems division of Westinghouse, there may well be some rationalisation of the development work, but the combined development efforts of the two major companies should accelerate the commercial introduction of this concept.
Two small 25 kW units have already been on line in Japan for some time, and the 100 kW utility CHP prototype at Westervoort has been working remarkably well since November 1997 in spite of changes to its original operating concept. A Mark 2 version of this prototype with an increased output of 150 kW and a cell deterioration rate of less than 0.1 per cent per 1000 hours is soon to be introduced to the commercial market by Westinghouse Electric. A longer term target is 1 to 5 MW modules using the same cells.
Siemens, which has already tested a 20 kW two-stack prototype, was also planning a 100 kW pilot plant project. It has been working with 5 cm x 5 cm rectangular flat plate cells, but it is now into a second phase using 10 cm x 10 cm plates with a mechanical integrity capable of withstanding loads of 300 to 350 MPa, which is double that of the earlier cells. It is using the latest electrode material alloy of vacuum plasma sprayed LaSrCrO3 with lanthium–strontium–manganate on the cathode side. Wet powder spraying is used for the latter, but screen printing is now being tried. Degradation is still fairly high on these cells at the moment, at 2 to 3 per cent per 1000 hours.
Sulzer in Winterthur also uses flat plate cells in the form of compressed flat discs, which look something like the end of a tin can, for their HEXIS mini-CHP programme which has some Japanese participation. A first field test unit is understood to have been tested in Dortmund where the mini-CHP concept was originated in 1997.
All of these initial projects use stand alone, unpressurized fuel cell modules with generating efficiencies of less than 50 per cent. In theory, efficiencies as high as 60 per cent should be attainable, but higher performances can be obtained using SOFC stacks in conjunction with gas turbine systems.
Integrating the fuel cell with a gas turbine essentially involves replacing the combustor with the fuel cell and using the turbomachinery as a turbo-charger. Efficiencies of over 60 per cent can be gained in such systems. A more ambitious option is to add a recuperator and waste heat recovery system. Efficiencies as high as 72 per cent are forecast for such systems.
ECN (Energie Centrum Nederland) is working on a recuperated SOFC/GT system using the Dutch Heron miniature gas turbine to produce some 6 MW of power at an efficiency of 60 to 70 per cent. They are working with both Siemens and Sulzer on component development and new materials for operation at higher temperatures than the present stacks.
Following the the successful operation of the Westervoort Westinghouse SOFC, a 250 kW integrated gas turbine SOFC plant operating at 3 bar pressure is being designed and built under the sponsorship of Southern California Edison to operate later in 1998.
According to Westinghouse, the development, design, fabrication, and test of both a 1 MW 63 per cent efficient system, and a multi MW 70 per cent efficient two-stage system are part of the recently signed five year cooperative agreement with the US DOE. Initial operation of the 1MW SOFC/GT system is scheduled for the year 2000, with the 70 per cent efficient unit an upgrade scheduled for late 2001.
By 2001, Westinghouse expects to be taking orders for commercial systems of from 1 to 20 MW output for the distributed power generation market at costs of between $1200 and $1500 per kW when the manufacturing facilities reach full production rates.
Mass production manufacturing facilities that will reduce unit costs to below $1000/kW are now being planned, but at present the fuel cells constitute a smaller proportion of the capital costs than the balance of plant. The power conditioning system appears to have less potential for cost reduction than the generating plant.
While SOFC developers are working on 100 kW and 200 kW prototypes, MCFC developer Energy Research Corporation had a nominal 2 MW demonstration plant in Santa Clara, California (SCDP) running in April 1996. It had a few plumbing problems and voltage fluctuations. Testing continued until March 1997, after which "funding limitations and the realisation that the demonstration had accomplished most of its objectives led to a decision to terminate the test phase." The plant was rated at 1.8 MW, but for a time the plant was run at 1.93 MW with extremely uniform performance. It ran for some 4100 hours in grid connected mode, somewhat short of the 10 000 hours planned. Over 1700 MWh of electricity was supplied to local consumers.
This programme began in 1990 when ERC gathered support from a group of electricity utilities as a result of an APPA/EPRI fuel cell commercialisation initiative. These utilities subsequently formed the Fuel Cell Commercialisation Group (FCCG) a group of municipal, rural, and investor owned utilities in the USA and Canada, which gained a royalty-based opportunity to purchase initial "direct fuel cell power plants".
ERC has now formed two wholly owned subsidiaries to support the commmercialisation plans:
FCE (Fuel Cell Engineering Corporation) – The general contractor for the SCDP, FCE is also responsible for product definition and the marketing of commercial power plants
FCMC (Fuel Cell Manufacturing Corporation) – A manufacturing subsidiary which, after having produced the stacks for the SCDP, is now manufacturing the larger components that will be the basis for FCE's commercial power plant contracts.
The next generation commercial power plant takes the form of a 2.85 MW packaged unit that is expected to be available as early as 2000. The extra 1 MW output is achieved by increasing the size of the cells in the stacks as well as the height of the stacks while reducing the footprint of the entire plant by some 90 per cent. Efficiency as high as 55 per cent is claimed.
More recently, a 250 kW MCFC CHP plant manufactured by M-C Power commenced operation in February 1997 at the Miramar Naval Air Station in California. Including the use of the fuel cell's waste heat the efficiency is said to be more than 70 per cent.
While the MCFC does not require quite such sophisticated materials as the SOFC, unit costs are generally expected to be higher than the SOFC option. Nonetheless, MCFC technology is being very actively pursued in Japan, Germany, Italy and the Netherlands.
In Japan, Mitsubishi will build a 200 kW MCFC for demonstration and testing by the Kansai Electric Power Company, which is a member of a Japanese consortium for developing MCFC power systems. The project is supported by NEDO (New Energy and Industrial Technology Development Organisation).
At the European Fuel Cell Group's 1998 Spring Workshop, MTU described a highly impressive truck-mounted 400 kWe MCFC integrated power generation plant assembled in a single standard transport container. This has been well tested, with a few interesting misadventures, but they are now also ready to launch a fully commercial product at very competitive prices.
Designated the "hot module", all the hot components are housed in one cylindrical vessel, which benefits from bearing the weight of the entire assembly. This arrangement results in low temperature differentials and minimum flow resistance. The hot module and system demonstration tests started in August 1997 at Ruhrgas's facilities, but the results had not been fully analysed at the Arnhem meeting. An interesting aspect of this MCFC is that it allegedly operates at higher temperatures than most SOFC systems at 950°C. The system uses ERC cells assembled by MTU and Ruhrgas.
Ansaldo MCFC work, including its 100 kW prototype at Molcare, is sponsored by a powerful team, including ENEL, CISE, ENEA, and FN in Italy as well as Spanish utilities Iberdrola and Endesa and Babcock & Wilcox Espanola.
ECN (Energie Centrum Nederland) is at an advanced stage in developing a 400 kW direct MCFC cogeneration unit with a thermal output of some 260 kW from steam at 150°C. This again is a packaged assembly mounted in a single standard transportation container. It uses an advanced version of the now familiar "three bunnies" multiple cell stack network known as the "Smart" system. This programme is funded by the EU and Dutch Government as well as NOVEM, and supported by an international group of participants including British Gas of the UK, Gaz de France and Sydkraft of Sweden as well as BCN, ECN and Stork of the Netherlands who contribute to the funding.
Proton exchange membrane (PEM)
Like the alkaline fuel cells, the PEM or "solid polymer" fuel cells were originally developed mainly for space vehicle power and military applications. The big difference is that whereas the alkaline fuel cell will always be far too costly for public transport or distributed power generation applications, the PEM cells could well compete with SOFC or MCFC in all of these markets in spite of its more modest efficiency.
Like the PAFC, they operate at relatively low temperatures, but they also share the advantages of the SOFC's solid electrolyte, avoiding the complexities of the pumps plumbing of the liquid electrolyte systems.
Siemens developed the technology for use in the Solar – Wasserstoff – Bayern programme in which energy from tracking solar dishes was used to crack water to produce hydrogen which could be transported over pipelines to be consumed in fuel cell power plants to generate electricity or to be used in vehicle propulsion. Siemens is now concentrating on SOFC technology for power generation and PEM for vehicle propulsion.
Daimler-Benz in Germany, which is making large investments in the development of PEM propulsion units, recently decided to pool its experience in both the mass production processes and the design of fuel cell vehicle engines with Ballard Power Systems of Canada in a C$450 million co-operation.
The one high cost material which might have threatened uncompetitive unit costs is the need for a platinum catalyst, but recently BASF have developed a non-precious metal catalyst for the Daimler-Benz methanol-fuelled fuel cell vehicle engine.
The Canadian concern now seems to have taken over the stationary generating plant applications of PEM fuel cells, more recently in collaboration with Alstom. A Ballard Power Systems subsidiary – Ballard Generation Systems – has been formed as a joint venture between Ballard, New Jersey Utility group GPU, and Alstom in Paris. GPU received two of the 250 kW pre-production units for field test operation.
The core of the PEM fuel cell consists of two electrodes separated by a polymer membrane electrolyte. Both anode and cathode are coated with a thin layer of platinum catalyst. At the anode, hydrogen fuel catalytically disassociates into free electrons and protons. The free electrons are conducted in the form of usable electric current through an external circuit. The protons migrate through the membrane electrolyte to the cathode and combine with oxygen from the air and electrons from the external circuit to form water and heat. The electrochemical process takes place at a temperature of around 85°C.
Ballard's initial fuel cell power plant development activities included the demonstration of a 30 kW plant operating on hydrogen and a 10 kW plant operating on natural gas. Current activities are focussed on developing a 250 kW package fuelled by natural gas ready for commercial introduction at selected sites in 1999 and 2000.
In August 1997 a 250 kW pre-production model successfully generated electricity after five years of development activity.
In April 1997, Plug Power LLC was formed in the USA as a joint venture between Edison Development Corp, a subsidiary of Detroit Edison, and Mechanical Technology Inc to develop and market PEM fuel cell units capable of providing the electricity needs of a complete household.
At an even smaller scale, miniature PEM fuel cell units are being developed to power lap top computers.
TablesComparison of fuel cell technologies