Intelligent predictive model of electrical power in photovoltaic systems through solar radiation and temperature on site

A Neuro-fuzzy ANFIS model is presented that allows predicting the generation of electrical power in a photovoltaic system considering the solar radiation and environmental temperature of the geographical area. This model enables fuzzy modeling to learn from the data set and, as a result, compute the most appropriate membership function parameters. It also makes use of neural networks' abilities to classify data and find patterns. The algorithm is developed using MATLAB® software and is trained with the acquired data weather station located in Hermosillo, Sonora City, and the electrical power output of the system at the site. The algorithm is evaluated using rigorous evaluation factors which shows how much the predicted value differs from the actual value, resulting in the following predicted values: RSME: 295.2686, NRMSE: 0.09496, RMSPE: 9.4306, MAE: 218.6996, and MAPE: 6.985, respectively.