Performance and cost modelling taking into account the uncertainties and sensitivities of current and next-generation PEM water electrolysis technology

Abstract

This paper presents a bottom-up approach to the assessment of model performance and costs of a proton-exchange-membrane electrolysis considering cell, stack and process levels. The cell voltage is modelled dependent on current density and detailed models for stack, investment and hydrogen costs are developed. Taking into account current research on PEM electrolysis, such as the use of thinner membranes or low precious metal loading on the electrodes, allows the prediction of next generation's efficiency and costs. By comparison of a current and next-generation PEM electrolysis, the effectiveness of individual development steps was assessed and remaining space for efficiency and cost improvement was identified. This can help to prioritize and to focus on development steps which are most effective.In the next generation, efficiency will be increased even at higher current density operation. Thus, specific stack costs will drop to less than half of present day costs which is decisive to achieve lower hydrogen production costs in the next generation. Specific installed costs and hydrogen production costs of the current and next generation are calculated for plant sizes up to 100 MWDC and reveal significant cost decrease for plant capacities up to 25 MWDC while only changing slightly for capacities larger than this.Costs are always subject to uncertainties due to model assumptions and boundary conditions that need to be defined. Uncertainties and the sensitivities of the model are estimated and assessed to provide an indication of the actual cost range. Main cost model uncertainties are identified to arise from membrane electrode and stack assembly costs, civil engineering and construction surcharge as well as the electrical system. Hydrogen costs are dominated by operating costs and therefore are highly sensitive to the annual operating hours and the electricity price, which have a greater impact on the hydrogen costs than the model assumptions for capital costs.