Estimation of power outputs of two different photovoltaic solar panels with different heuristic algorithms

The estimation of the power values obtained from photovoltaic (PV) systems is of critical importance for the reliable and economical use of solar energy panels. This estimation affects many processes, starting from the installation phase of solar panels to guiding electricity companies, energy management, and distribution. At the same time, it is necessary to detect the adaptations of solar panels in a timely manner and reach the optimal production capacity to provide the most efficient energy production. In this context, Artificial Neural Networks (ANN) were used to estimate the power values obtained from PV panels. In this study, heuristic algorithms such as Particle Swarm Optimization (PSO), Genetic Algorithm (GA), Clonal Selection Algorithm (CSA), Ant Colony Optimization, and Artificial Bee Colony (ABC) were used to estimate the power values obtained from monocrystalline and polycrystalline photovoltaic panels. In the verification of the estimation results, the most common statistical evaluation criteria, Mean Absolute Percentage Error (MAPE), Root Mean Square Error (RMSE), and Variance (R2) equations were used. The estimation values made with the PSO algorithm were the closest to the real values. 98.95% estimation was achieved in monocrystalline photovoltaic solar panels and 93.94% in polycrystalline photovoltaic solar panels.