Cost Projections For Planar Solid Oxide Fuel Cell Systems

Dr. Kevin Krist Gas Research Institute 8600 West Bryn Mawr Chicago, IL 60631

John D. Wright, Dr. Cecily Romero TDA Research 12345 W. 52nd Avenue Wheat Ridge, CO 80033

Dr. Tan Ping Chen Bechtel Group, Inc. 50 Beake St. San Francisco, CA 94119-3965

The Gas Research Institute (GRI) is funding fundamental research on solid oxide fuel cells (SOFCs) that operate at reduced temperature. As part of this effort, we have carried out engineering analysis to determine what areas of research can have the greatest effect on the commercialization of SOFCs. Previous papers have evaluated the markets for SOFCs and the amount which a customer will be willing to pay for fuel cell systems or stacks in these markets, the contribution of materials costs to the total stack cost, and the benefits and design requirements associated with reduced temperature operation (1,2). In this paper, we describe the cost of fabricating SOFC stacks by different methods. The complete analysis is available in report form (3).

In summary, we found little difference in the cost of stacks manufactured by traditional ceramic forming techniques (such as ball milling, tape casting, calendering and screen printing and sintering) atmospheric plasma spraying. The cost of putting down a layer of material of a given thickness is much higher in a process that must be carried out under vacuum (sputtering and vacuum plasma spraying). Even though vacuum processes are capable of making thin gas tight layers, their use still significantly increases the cost of the stack. Thus, vacuum processes are desirable only if their use can significantly increase the performance. In general, when a low cost metal interconnect is used (an interconnect stamped from a sheet of high alloy metal), a wide variety of ceramic and atmospheric plasma spraying processes can be used to make fuel cell stacks at a cost of $500-600/m2 ($250-300/kW at a power density of 2 kW/m2). In contrast, the addition of a single vacuum step to produce a 5pm thick electrolyte layer adds $100/m2 (sputtering) to $150/m2 (for vacuum plasma spraying). The use of a thick ceramic interconnect (2 mm thick with gas passages for the fiiel and air) instead of a metal interconnect adds approximately $500/m2 to the cost of the stack.

To calculate the cost of manufacturing planar stacks, we defined complete manufacturing process flow sheets. The analysis included all the steps needed to fabricate a stack, starting with powders and ending with stacks that have been assembled and tested. We assumed the production of 200 MW of capacity per year. This is large enough that all of the equipment in the manufacturing facility can be used 24 hours per day; this facility is larger than the market entry facility (10 to 20 MW capacity per year), but costs could be reduced even further at greater production rates.

To calculate the SOFC stack manufacturing costs, we identified all of the steps needed to produce a stack. We specified the equipment and the number of units needed. The cost and throughput of the equipment was based on manufacturers specifications. The total fixed capital investment (which includes the cost of equipment, installation, land, site development, service facilities, and indirect costs such as engineering, construction overhead, contingency and contractors) was taken as 3.9 times the FOB equipment cost. Working capital was 20% of the fixed capital investment.

The annual direct operating costs include the cost of raw materials, utilities, labor and maintenance. Raw materials costs were taken from suppliers estimates, and generally account for 40% for the total stack cost (slightly less when expensive vacuum deposition processes are used). The major raw materials costs are those for the lanthanum strontium manganite cathode (LSM) at $40/kg, yttria stabilized zirconia (YSZ) electrolyte at $45/kg, YSZ stabilized nickel cermet anode ($24/kg) and the metal ($10/kg) or lanthanum, strontium chromite (LSC) interconnect at $60/kg. The costs for these materials were manufacturers estimates of the price of these materials supplied to a large scale SOFC manufacturing facility, and are in general a factor of 2 to 10 less than the current cost of these materials in small scale production. The cost of solvents, binder and gases-were also calculated explicitly, but have little effect on the results. Electricity was valued at 7.0?/kWhr.

The direct costs also include utilities (which were calculated for each unit operation), maintenance (4% of fixed capital) general supplies (5% of fixed capital). However, the largest of the direct operating costs after raw materials is the labor needed to run the plant., Here we assume a workforce of 200 at an annual salary of $35,000/person. On average, direct labor accounts for 15% pf the cost of manufacturing a fuel cell.

Indirect expenses are proportional to the direct expenses. Overhead (70% of labor) and administrative costs (25% of overhead) essentially double the effect of the labor costs (labor and expenses proportional to labor account for 30% of the stack cost). Other costs are proportional to the fixed capital investment; depreciation (10%), local taxes (2%), insurance (0.6% of fixed capital). Distribution (10%) and marketing and R&D (5%) are proportional to total expenses. Profit is calculated to provide a 20% rate of return, ie: (profit + depreciation)/(fixed + working capital) = 20%. In general, the direct cost of production account for 50% to 60% of the total manufacturing cost and indirect expenses account for 40% to 50%.

The cost of the unit operations needed to make a fuel cell are shown in Table I. The costs are given in terms of $/m2. Each of these cost is the sum of three monthly costs (capital equipment * depreciation + maintenance + operation) divided by the number of m2 of material the unit can process per month. In general, the traditional ceramic processing steps (ball milling, spray drying, screen printing, tape casting and calendering and laminating) and atmospheric plasma spraying are relatively inexpensive. The exception is the sintering steps, which cost several times as much as the forming steps. The cost of atmospheric plasma spraying is similar to that of the more traditional ceramic processes. The vacuum processes are much more expensive because of their high capital cost and low throughput.

Table I Processing characteristics and costs of the manufacturing steps.


Cap. Cost ($K)

Throughput m2/d

Monthly costs (SK) capital + operating

Processing Cost S/m2

Spray drying (130 pm layer)

0 0

Post a comment