Spatial and temporal variability of crystalline silicon solar cell performance in Ethiopia: A hybrid approach

Estimating solar irradiance and evaluating solar cell performance is crucial for successful photovoltaic (PV) system deployment. This study developed a stacked ensemble machine learning model to generate high-resolution solar radiation data for Ethiopia in 2022. Utilizing observational data from 16 synoptic weather stations, provided by the National Meteorological Institute, the model effectively predicted solar radiation and converted to Plane-of-Array (POA) irradiance. Subsequently, employing single-diode physical models, the performance of crystalline silicon (c-Si) solar cells was assessed across the country. Our findings demonstrate substantial spatial and temporal variability in solar cell performance using a hybrid approach for better device parameter estimations. The highest performance (23.34–25.96 %) was observed in the northern and eastern/southwest regions during the spring season, primarily attributed to optimal sunshine duration and high irradiance levels. In contrast, the western and southern regions exhibited the lowest performance (9.88–12.95 %) during the summer, where it is related to the relatively less favorable solar conditions including lower irradiance levels. The annual average device efficiency across Ethiopia ranged from 13.02 % to 22.35 %. This study provides valuable inputs for a scientific site selection, comparison, seasonal/annual device performance forecast, and design of a PV system in Ethiopia.