Presentation Information

[8p-P01-52]Explainable AI for utility-scale solar PV Site Selection: random forest-SHAP Analysis of land suitability factors for ground-mounted solar installations in Japan

〇(DC)Sangay Gyeltshen1 (1.Nagoya University)

Keywords:

Solar photovoltaic,Renewable Energy,explainable AI

Japan's carbon neutrality commitment by 2050 requires significant renewable energy expansion, particularly utility-scale solar photovoltaic (PV) installations. However, Japan's complex topography, limited land availability, and high population density create unique challenges for optimal solar farm site selection. This study proposes an explainable artificial intelligence framework combining Random Forest (RF) modeling with SHapley Additive exPlanations (SHAP) analysis to identify and prioritize suitable locations for ground-mounted solar PV installations across Japan.The methodology integrates multiple geospatial datasets encompassing resource factors, environmental variables, anthropogenic elements, orographic characteristics, and climatic parameters. Critically, the framework incorporates policy-related restriction layers ensuring regulatory compliance and avoiding development conflicts, including protected zones, hazard-prone areas, and legally restricted territories under Japanese environmental and land use regulations.The model's decision-making process becomes transparent through SHAP analysis, providing global feature importance rankings across Japan and local explanations for regional variations. Global SHAP analysis identifies the most influential factors for national-scale solar PV site selection, while local SHAP values explain regional variations and site-specific decisions, accounting for Japan's diverse geographical and climatic conditions. This explainable AI approach addresses the critical need for transparent decision-making tools in renewable energy planning. Results are expected to provide actionable insights for policymakers, energy planners, and investors, facilitating evidence-based decision-making for Japan's solar energy expansion.