Future cost projections for photovoltaic module manufacturing using a bottom-up cost and uncertainty model

Abstract

Photovoltaic module prices have typically decreased faster than projections. There are two methods usually used for these projections; cumulative market shipment experience curves or detailed bottom-up cost calculations for specific technologies. The former suffers from a lack of specificity in terms of technology or changing market situations, and the latter from a lack of quantifiable uncertainty. We present an alternative bottom-up future cost model for a new vertically integrated c-Si PV factory, from poly silicon to module, incorporating input ranges and uncertainty via a Monte Carlo analysis. Neglecting profit margins, the majority of projected scenarios for global 2025 module manufacturing cost fall between 0.10 US$/W and 0.18 US$/W, due mostly to reductions in raw materials costs for module fabrication. The lowest 10th percentile projections, below 0.10 US$/W, would be realised by the largest scale manufacturers with access to very low materials costs and low operational costs. The model projects production cost learning rates between 29% and 43% compared to a long-term historical average module selling price learning rate of 24%. Analysis of the competitive position of silicon heterojunction cell technology in combination with multi-wire module technology is performed. Access to the multi-wire technology improves competitiveness of 2025 manufacturing in high labour rate countries if silver prices remain high and the cost of the multi-wire material is reduced in line with other materials, motivating continued development of this technology. Finally, continued reduction in labour intensity, through automation and throughput will enable more competitive manufacturing in higher labour cost locations.