NWPsolarNet: A Scalable Deep Learning Framework for Medium-Term Solar Forecasting across Europe

The accelerating global adoption of solar photovoltaic power necessitates sophisticated forecasting methodologies to ensure grid stability, yet scaling these models from local pilots to continental domains reveals critical data fidelity hurdles. Standard approaches relying on rated peak capacity for normalization are often compromised by noisy large-scale datasets containing anonymized coordinates, incorrect capacity ratings, or sensor malfunctions. To address these limitations, this study presents an advanced iteration of the NWPsolarNet Deep Neural Network