Seeking a better path for the circular economy of solar panels: Global sensitivity analysis focused on socioeconomic and physical factors

Abstract

As more solar photovoltaic panels are expected to be introduced globally to promote the stabilization and decarbonization of electricity supply, concerns have been raised about the associated resource constraints and the environmental impacts. To promote circular economy initiatives, this study aims to provide insight into the variables and the conditions that play a key role in the expression of environmental impacts and find the better path for a circular economy of solar panels, which requires analyzing impacts on multiple indicators as a result of various factors and their cross-interactions. To achieve this goal, we constructed a resource circulation simulator that considers factors related to the business environment as well as physical factors. The simulation model was then applied to the largest administrative region in Japan, and the parameter space was explored by combining Regional Sensitivity Analysis, a global sensitivity analysis method, and Random Forest to consider the interactions among these variables. As a result of 30,000 Monte Carlo simulations performed on 26 input variables, it was found that the key parameters in final disposal and resource consumption potential differ between the two purposes; avoiding catastrophic loads and achieving ambitious reductions. In contrast, for greenhouse gas emissions, it is important to maintain or increase the efficiency of photovoltaic panel power generation for both purposes. Additionally, in the case where the key variables were optimized in terms of final disposal/resource consumption potential, it will effectively reduce the two impacts to 0.7 % and 12.6 %, respectively, with a minimal negative impact on greenhouse gas emissions. In contrast, in the case when optimized in terms of greenhouse gas emissions, it has only a small impact on greenhouse gas emissions and large negative impacts on final disposal and resource consumption potential. Further refinement of the parameter space and thresholds in the analysis is a topic for future research.