Appendix F. Technology Learning and Market Penetration

Every commercialized technology has shown a propensity to reduce costs with cumulative manufacturing experience. Cost reduction and performance improvements can occur for a wide variety of reasons, including R&D, economies of scale, technology spill-over, economy-wide advances in science and technology, and process improvement resulting from manufacturing learning. Most of these factors are virtually impossible to separate from each other because of the lack of data and the high correlation among many of the factors.

This appendix explores the implications of technological progress induced by “learning-by-doing” to assess the challenge presented by the cost reduction target of PEM fuel cells. To apply the theory of learning to a particular technology, it is necessary to establish initial unit overnight capital costs and the cumulative quantities/capacity of the PEM fuel cell technology already built at a point in time. Cumulative capacity built is a surrogate for cumulative learning in the formulation. The learning rate must be assumed, i.e., the percent cost reduction for every doubling of cumulative capacity due to experience.

Cumulative PEM Capacity and Initial Capital Cost

As noted above, according to Fuel Cell Today (March 2006), 550 FCVs were built world-wide between 2000 and 2005 and at least another 70 units of 70 to 80 kilowatts each are estimated to have been built in each of 2006 and 2007.¹¹⁶ Honda will add another 200 FCVs by 2009 and more are reasonably expected in 2010. An additional 3,000 fuel cells, similar or identical to the PEM systems used in FCVs, have been built and used for niche transport markets such as marine and auxiliary power applications, light rail, and fork lifts through year 2006,¹¹⁷ with sizes varying from 65 kilowatt to 130 kilowatt. For this illustration of the potential impacts that “learning” might have on cost reduction, the starting point for technology learning of PEM fuel cells in 2010 was assumed to be at least 250 megawatt and at costs of between $3,000 and $5,000 per kilowatt in the learning process.¹¹⁸ A cost of $3,000 per kilowatt in 2010 is assumed for this example.

PEM Technology Learning Rate

The rate of technology learning for the PEM fuel cell is critical to the success of the hydrogen FCV. To achieve PEM capital cost of $30 per kilowatt and achieve a dominant share of FCVs in the LDV market, the learning rate for both the fuel stacks and the balance of plant (BOP)¹¹⁹ must be at least a 30 percent for every doubling of cumulative capacity built. Such a learning rate has never been realized for any durable good product throughout the production life of that product.¹²⁰ Portions from the McDonald and Schrattenholzer article are provided in Table F.1. For estimated learning rates with R2 of over 80 percent, learning rates vary by region and time period and generally range between 8 and 26 percent per doubling of cumulative capacity. Most researchers use a learning rate of about 20 percent for newly-commercialized technologies in their projections for the initial phase of cost reductions.

Lipman and Sperlman (2000) at the University of California at Davis discussed the PEM technology in its infant stage and warned against assuming that high early learning rates will continue indefinitely: “For products such as PEM fuel cells that may reach high levels of accumulated production, we suggest methods [be developed] for bounding [cost] forecasts in order to guard against eventually forecasting unrealistically low costs.”¹²¹ The table by McDonald and Scharattenholzer and the learning rates for the gas turbine anecdotally support the warning. While gas turbine costs declined worldwide by 22 percent for every doubling of production capacity between 1958 and 1963, the learning rate declined to about 10 percent between 1963 and 1980.

As seen by these curves and their extension, PEM fuel cell costs would not fall enough under any of these assumptions to meet the $30 per kilowatt capital cost target by the time two million FCVs are sold. In most instances, the target fuel cell cost including the catalyst could not be achieved if 10 million FCVs were sold. Assuming a 30-percent learning rate for the complete fuel cell, the PEM fuel cell cost, i.e., $47 per kilowatt, would nearly reach the DOE target costs when 10 million vehicles are sold. Learning rates of 20 percent would yield fuel cell costs of $223 per kilowatt and would not achieve the target DOE fuel cell costs.

While R&D and engineering research could eventually succeed in solving all of the challenges that are faced in making fuel cell LDVs a cost-effective reality, the number of necessary simultaneous R&D successes that are required within the next 22 years makes large scale penetration of FCVs largely improbable in the United States without significant long-term Federal and State policies that promote FCV adoption over a 10-to-20 year period.

Notes