AbstractThis paper uses the formal concept of learning curves the analyze regular behavior of performance improvements in various energy technologies. The concept allows the estimation of a single indicator of technological progress, the learning rate, which expresses the constant percentage improvement (usually in terms of cost reductions) in a technology for each doubling of the technology's cumulative installed capacity. We present 42 energy-related learning rates, either calculated directly from available data or assembled from the literature. We elaborate briefly on eight of these to illustrate issues addressed by technology assessments to convert these raw historical learning rates into prospective learning rate distributions for use in long-term energy models. The paper includes a sensitivity analysis of policy-relevant variables with respect to learning rates, a discussion of possible extensions and limitations of the approach and an outlook on future work in the field.
McDonald A <http://pure.iiasa.ac.at/view/iiasa/1442.html> & Schrattenholzer L <http://pure.iiasa.ac.at/view/iiasa/2366.html> (2002). Learning Curves and Technology Assessment. IIASA Research Report (Reprint). IIASA, Laxenburg, Austria: RR-03-002. Reprinted from International Journal of Technology Management, 23(7/8):718-745 .