Economies of Scale in PEMEC and SOEC Manufacturing based on a Bottom-up Model

Under the scope of net zero emission, a large-scale deployment of renewable technologies, especially clean hydrogen technologies are required. Hydrogen is a clean fuel and an ideal energy carrier that can be used to store, move, and deliver energy. Nowadays, most hydrogen is produced by steam methane reforming or gasification of coal, with only less than 5% of hydrogen produced by electrolysis. Among the dominant electrolysis technologies, PEM electrolysis and SOE are favored of high hydrogen production rates and high energy efficiencies, but the cost hindered their further application. However, it is foreseeable that the cost of green hydrogen will decrease with the global shipment due to the learning effect. This effect is attributed to the advancement of manufacturing, learning-by-doing (LBD), economies of scale (EoS), etc. Considering the surged demand of green hydrogen, EoS could be a breakthrough point to reduce its costs. In this study, a comprehensive bottom-up cost estimation model was built and validated by the public database and specific literature. This model evaluated the cost of PEMEC and SOEC from several different perspectives, such as material cost and capital expenditure. Especially, the relationship between these components and the production volume was addressed. It was found that when the volume is relative small, the costs shrink rapidly with capacity growth and an exponential correlation is almost fulfilled. When the throughput of the factory is between 500 MW/y and 10000 MW/y, the continuing cost reduction is mainly brought by higher utilization rate of equipment, labor and buildings. In addition, when the plant scales up, costs are nearly constant, approaching minimum values which are decided by the material costs. Thus at the same capacity, PEMEC costs more due to its consumption of noble metal. We also explained six principles of EoS in detail with statistic evidences. Moreover, by associating the cost model with the learning curve and applying different parameters, this model has the potential to be applied to green hydrogen price forecasting and various industries.