Assessment of Land and Renewable Energy Resource Potential for Regional Power System Integration with ML Spatio-Temporal Clustering

Toward the planning and development of future electric power systems with extremely large penetration of renewable resources (DERs), detailed assessment of power and energy potential is needed considering both spatial and temporal data per region. Within this paper, a methodology for spatio-temporal DER capacity potential considering land cover types and weather variation are presented using spatio-temporal data. Additionally, an application of empirical orthogonal functions (EOFs) and max-p unsupervised learning techniques is proposed for DER generation to identify zones of similar output power in space and time. A detailed case study for the example region of Kentucky, USA is completed with state-of-the-art utility scale solar photovoltaic (PV) panels, wind turbines, and publicly available data from the National Aeronautics and Space Administration (NASA) Earthdata resource and the National Land Cover Database (NLCD). Annual estimates of wind and solar PV power for the example region are found to meet the state's public annual energy demand, even in the low land usage case. Further efforts to decarbonize energy generation and build additional renewable energy capacity are supported through the methodology and case study.