Modeling of the solar radiation output in the Democratic Republic of Congo

Abstract

The Democratic Republic of Congo is not only a geological contrast but also an energy contrast, as it possesses a great hydroelectric potential capable of satisfying its own needs and those of many African countries. However, according to the most optimistic studies, the electrification rate in the Democratic Republic of Congo is less than 19%. Photovoltaic solar energy therefore represents a viable solution to help meet national energy needs. Accurate knowledge of solar irradiance is essential for proper sizing and optimization of photovoltaic systems. This study aims to establish general statistical models to predict solar irradiance (kWh/m²/day) as a function of sunshine duration (h/day) for any province of the Democratic Republic of Congo. Global solar irradiance data were obtained from RETScreen through the Renewable Energy Atlas of the DRC, while sunshine duration data were collected from the European Centre for Medium-Range Weather Forecasts. Five models (linear, quadratic, cubic, logarithmic, and exponential) were developed using the least squares method and validated using normalized statistical errors (NMAE, NMBE, NRMSE), the Nash–Sutcliffe efficiency criterion (NS), and the coefficient of determination (R²). Two scenarios were analyzed: scenario 1: excluding the provinces of Nord and Sud Kivu, and scenario 2: including these provinces. By excluding these provinces, the quadratic and cubic models explained 88% and 93% of the relationship between annual average sunshine duration and solar irradiance, respectively. Including them reduced performance, with R² ≈ 0.70, indicating that additional climatic variables may be required. A national sunshine duration heat map identifying four main exposure zones was also produced, supporting photovoltaic planning across the Democratic Republic of Congo.