Sensitivity Analysis of Weather Factors Affecting PV Module Output Power Based on Artificial Neural Network and Sobol Algorithm

Abstract

There are many factors in weather conditions affecting the output power of photovoltaic (PV) modules in the PV system. For better modelling the output of PV modules, it is necessary to analyze the input factors of the model, finding out the key factors or combinations that affect the output greatly. In this paper, sensitivity analysis was accomplished to analyze the input weather factors of PV modules precisely by combing Sobol algorithm and artificial neural network (ANN). First, the modeling of output power of PV modules was realized by ANN self-learning method with the help of experimental data. Then, a sensitivity analysis with Sobol algorithm was built to study the sensitivity of the change of output power for PV modules with the change of three input weather factors, including solar irradiance, module temperature, and air relative humidity, and obtain sensitivity coefficients of these weather factors. Finally, the welltrained ANN model with high fitting accuracy was utilized for Sobol sensitivity analysis, giving a reliable analyzing result of the weather factors.