Investigation of learning and experience curves

The applicability of learning and experience curves for predicting future costs of solar technologies is assessed, and the major test case is the production economics of heliostats. Alternative methods for estimating cost reductions in systems manufacture are discussed, and procedures for using learning and experience curves to predict costs are outlined. Because adequate production data often do not exist, production histories of analogous products/processes are analyzed and learning and aggregated cost curves for these surrogates estimated. If the surrogate learning curves apply, they can be used to estimate solar technology costs. The steps involved in generating these cost estimates are given. Second-generation glass-steel and inflated-bubble heliostat design concepts, developed by MDAC and GE, respectively, are described; a costing scenario for 25,000 units/yr is detailed; surrogates for cost analysis are chosen; learning and aggregate cost curves are estimated; and aggregate cost curves for the GE and MDAC designs are estimated. However, an approach that combines a neoclassical production function with a learning-by-doing hypothesis is needed to yield a cost relation compatible with the historical learning curve and the traditional cost function of economic theory.