A Radial Basis Function Neural Network Based Approach to Mitigate Soiling from PV Module

Abstract

Solar power is available in abundance and is free of cost but despite this fact photovoltaic (PV) systems are not reliable as the efficiency of PV module is still not up to the mark. Amount of sun light absorbed by the module is the fuel for the PV modules, there are many factors which play a vital role such as front surface reflectivity, solar beam incidence angle and the most important is the transmissivity of the front surface. Objective of this paper is to examine the function of a neural network based model of a module to mitigate the soiling and improve the accuracy of dust prediction over the module and enhance the efficiency which was degraded by soiling. One of the techniques in artificial intelligence named, radial basis function neural networks (RBFNN) is utilized here to predict the dust data by reading the data of solar irradiance, module temperature, PV voltage, current and power and generates the control signal to the windshield wiper motor that drives the wiper to wipe the dust from the panel. A considerable measure of accessible exploratory information was utilized for the preparation of the RBFNN, and a back propagation algorithm was utilized.