Fuel cells are electrochemical devices for generating electricity. Like batteries they use the energy released by a chemical reaction – in the case of the fuel cell the reaction between hydrogen and carbon monoxide obtained from a fuel, and oxygen from an air supply – to generate electrical energy. But unlike a battery, whose life is dictated by the limited store of chemical reactants it carries with it, the fuel cell will operate continuously provided it is supplied with an external source of air and fuel.

There are four primary types of fuel cell being developed for power generation applications: the phosphoric acid fuel cell (see MPS June 1997); the molten carbonate fuel cell; the polymer (or proton exchange membrane) fuel cell (see MPS December 1997); and the solid oxide fuel cell (SOFC).

Both the phosphoric acid and the polymer fuel cells operate at below 200°C. The molten carbonate cell runs at the much higher temperature of 650°C. This allows the reaction between hydrogen and oxygen to proceed more efficiently. It also makes the cell more tolerant of carbon monoxide (CO).

Potentially the most promising fuel cell of all is the solid oxide fuel cell. Unlike any other type, this cell is entirely solid-state with no liquid components. (The polymer fuel cell is sometimes described as a solid state cell but it contains a high proportion of water.) As a consequence it has to operate at an extremely high temperature, 1000°C in the case of the Westinghouse unit deployed in the Netherlands.

This elevated temperature is needed to achieve the necessary level of conductivity in the cell’s solid electrolyte to enable it to operate efficiently. The all-solid construction eliminates many potential maintenance problems, but the high temperature places great demands on the materials.

The attraction of the SOFC is its potential efficiency. Simple cells should be able to operate at efficiencies of 50 per cent or higher, while integrating the fuel cell with a gas turbine could provide a natural gas fired power plant with an electrical energy conversion efficiency of up to 70 per cent, the most efficient fossil fuel fired plant yet conceived.

Strong support

The history of the solid oxide fuel cell dates back to the early 1970s when engineers first developed the concept of a ceramic high temperature cell. This concept was backed by the US government with development grants. Government sponsorship has supported development ever since.

The Westinghouse development programme is now entering its final stage, and should reach fruition soon after the turn of the century. In August 1997, the US Department of Energy and Westinghouse Electric Corp. signed a joint research and development agreement costing $202 million over five years. This will fund the final stages of SOFC development, cost reductions, the development of a prototype for a commercial megawatt class SOFC plant and the setting up of a commercial manufacturing facility.

While several companies around the world are developing solid oxide fuel cells today, most are based on flat plate or planar electrodes. Westinghouse has worked on a completely different configuration based on a tubular design offering several advantages in construction of a complete power plant module from individual tubular component cells.

The tubular ceramic fuel cell comprises two electrodes, one on the inside of the tube that acts as the oxygen electrode and one on the outside of the electrode which acts as the hydrogen electrode. These are separated by a solid electrolyte which, when heated to 1000°C, will conduct oxygen from one electrode to the other. The tube is sealed at one end to prevent the two gaseous reactants, oxygen and hydrogen, coming into physical contact, and to avoid the need for seals in the cell module.

Originally the whole cell was supported on a porous support. But Westinghouse has been able to eliminate this, using the oxygen electrode to provide support instead. This has simplified construction, and has also boosted power output by 35 per cent.

Much of the development work on the tubular cell involved 50 cm long cells. These were judged too small for commercial use. Commercial cells, such as those used in the Netherlands plant, are 150 cm long and 22 mm in diameter. A single tubular cell produces 210 W.

The fuel cell requires two gaseous fuel components, oxygen and hydrogen. Oxygen can be supplied directly from air but hydrogen must be obtained by converting natural gas into a mixture of hydrogen (H2) and carbon monoxide (CO) in a process called reforming. This mixture is then used by the fuel cell to generate electricity.

Reforming is an energy absorbing process. Conveniently, as a result of the elevated operating temperature of the solid oxide fuel cell, the reaction can be carried out internally in a fuel cell module using heat generated by the cells.

Fuel cell electrodes are extremely sensitive to the presence of sulphur, and so the natural gas must be desulphurized before it is used. As a result, the sulphur emissions are virtually zero. Its high efficiency also means that it generates relatively less carbon dioxide than conventional power plants. In addition, the reaction at the fuel electrode takes place in the absence of atmospheric nitrogen so negligible amounts of nitrogen oxide are produced, making the power plant extremely clean to operate.

The Westervoort project

The road to commercialization of the solid oxide fuel cell has involved a number of pilot scale projects. The largest, and most significant to date, is the 100 kW plant built by Westinghouse for a consortium of Dutch and Danish energy companies. The group of companies comprises the umbrella organisation EnergieNed, five Netherlands energy distribution companies: EDON; NUON; PNEM; Megalimburg; and Enerco, and Elsam, representing six Danish energy companies. The project received the financial backing of the Netherlands Ministry of Economic Affairs.

The interest of these companies in the solid oxide fuel cell was stimulated by a search to reduce the negative environmental effects of power generation and distribution. The companies hope the demonstration project will help to make the technology competitive by 2002-2005.

Feasibility studies for the scheme took place between 1992 and 1994. A contract was signed in 1994, after which Westinghouse proceeded to design and manufacture the fuel cell module. Installation took place in November 1997 and operation began in December, with an official inauguration ceremony planned for 14 May 1998.

The Westervoort plant was conceived as a European project and much of the balance-of-plant equipment was procured in Europe. This included a desulphurizer, a UPS unit and the district heat export system from the Netherlands, and an inverter from Switzerland.

The SOFC generator module for the 100 kWe plant comprises 1152 tubular cells of the type intended for use in commercial power plants. These modules are combined to form cell bundles, each containing 24 cells in a 3 x 8 array. Four bundles are connected in series, to form a bundle-row and 12 rows, arranged side by side, make up the complete module.

Reformers are located between each bundle row. These reformers are driven by radiant heat from the adjacent fuel cells. The module is fuelled with natural gas which is available at the site. The fuel supply system includes a supply of purge gas for use during startup and shutdown, a source of steam to be used during startup only, and the natural gas desulphurizer.

The main air supply for the module is provided by two blowers. From these the air passes through a low temperature and then a high temperature recuperator where heat contained in the exhaust from the fuel cell is used to preheat the incoming feed. During startup and low power operation, an air heater is neccessary to preheat incoming air.

The exhaust gas from the recuperators is fed into a heat recovery system where it is used to heat water in the return feed of the Duiven-Westervoort district heating system. Pressurized water, entering at 45 to 60°C, is pumped through the finned-tube heat exchanger where the water temperature is boosted.

On the electrical side, a power conditioning module converts the dc output of the fuel cell module to a 400 V, three phase, ac supply suitable for delivery to the local grid.

The Westervoort unit is a packaged design. The complete unit, which is installed indoors in the NUON auxiliary heat plant, has been fabricated as four subsystem skids. These are the solid oxide fuel cell generator module, the thermal management module including an instrumentation and control module, a fuel supply module and the power conditioning and heat export module.

The SOFC field unit, excluding the inverter, is 3.6 m high, 2.8 m wide and is 8.5 m long. It operates unattended, with its control system linked to the control room of NUON, one of the project partners.

Monitoring performance

The 100 kWe power plant will be operated for two years, during which time it will be extensively monitored and exercised to test and demonstrate the feasibility of the technology in a ‘real’ CHP application. A schedule of tests will be preprogrammed in to the control software.

The main parameters to be monitored include maximum and minimum capacity limits, maintenance requirements, mean time between failures, durability, availability and reliability. Fuel and air requirements and component degradation will also be followed.

The plant has a nominal output of 100 kWe of electricity and 54 kW of heat. The net electrical efficiency at this operating point is estimated to be 47 per cent. Such an efficiency is expected to be maintained over a broad range of power outputs about the nominal operating point. The maximum output of the unit is 145 kWe and 130 kWth. At this output level, electrical efficiency falls to 40 per cent. The predicted performance parameters for the plant.

As of 23 April, 1998, the unit had run continuously for 2175 h at, or above, its nominal capacity. The output voltage level was stable at 256.5 V dc. Performance data for the first 1600 h of operation. The unit had consumed an average of 20 kg/h of fuel. It has so far operated at the maximum rate of 108 kW of electricity and 85 kW of heat. Electrical efficiency was 42 per cent. This is believed to be lower than the best achievable; the unit has been operating at 45°C below its normal operating temperature. Overall efficiency has been 78 per cent.

Future development

The Westervoort project represents one of the final stages in the development of the Westinghouse solid oxide fuel cell power plant for commercial operation, testing its operation in a CHP environment. Westinghouse believe that under optimum conditions, electrical efficiency in this mode can reach 48 to 50 per cent at atmospheric pressure.

For larger installations primarily intended for power generation, the company is developing the concept of integrating the fuel cell unit with a gas turbine. To achieve this the fuel cell must be run at elevated pressure.

Tests have shown that cell voltage increases at any given current as the pressure is increased. Westinghouse engineers calculate that the optimum pressures would be five to ten atmospheres in order to match the pressure at the turbine inlet of small gas turbines.

Integration of the fuel cell with a gas turbine effectively replaces the gas turbine combustor with the fuel cell, and uses the gas turbine as a turbocharger. Air from the turbine compressor is used to feed the fuel cell. The hot, pressurized exhaust gas from the fuel cell then powers the generator, providing an additional power output on top of that generated in the fuel cell module itself.

This arrangement could have an electricity generation efficiency of around 60 per cent in systems as small as 250 kW. Even higher efficiency could be obtained in larger systems starting at 2 MW by further refinement of the cycle to include a gas turbine reheat stage. This could boost efficiency as high as 72 per cent.

A 250 kW integrated gas turbine solid oxide fuel cell power plant operating at a pressure of 3 atmospheres is currently being designed and built with sponsorship from Southern California Edison, the US Department of Energy and Westinghouse. This unit is expected to start operating next year.

A 1 MW plant of similar design is scheduled to be built for startup in 2001. This will later be adapted for the incorporation of a reheat turbine stage. Predicted electrical efficiency is 70 per cent.

Westinghouse hopes to start taking orders for commercial units within two years. The first products are likely to be pressurized units integrated with small gas turbines. Plant size will be 1 to 20 MW. These will be aimed at the distributed generation market. The plants are expected to cost $1200/kW to $1500/kW when the factory has reached full production. Efficiencies of these plants are expected to fall in the range 62 to 72 per cent.

Table 1. 100 kW solid oxide fuel cell performance estimates