Multi-Data Source Feature Extraction and Spatio-Temporal Data-Driven Optimization Decision for Photovoltaic Power Station Site Selection

This study presents a novel method for photovoltaic power station site selection via multi-data source feature extraction and spatio-temporal with data-driven optimization. By integrating high-resolution digital elevation models, we combine satellite data, and ground-based meteorological sensors measurements to estalish a dynamic solar irradiance estimation algorithm. Under a two-stage robust uncertainty optimization, a practice framework is employed to coordinate demand response, energy storage systems, and grid interactions. The experimental results shows that the operational costs have been reduced by 12 to 18% and the fault prediction reached 92%. The apporach provides an actionable policy insights for enhancing demand response participation and advancing adaptive decision-making tools for polar-volatile operating selection environments.