Paper Title
An Improved AI Fault Detection Technique for Solar Cells and PV Modules

The photovoltaic market has been growing rapidly for several years. One of the main reasons behind use of PV module is the easy availability of PV module as well as reduction of PV production costs. The output power of PV module depends mainly on two parameters called radiation and temperature. There are a number of factors that affect the performance of the PV array, such as diode and contact loss, incompatibility loss, DC / AC compression loss, solar tracking loss, shadow loss, dirt loss, and material loss. The detection of these faults is complicated and risky by conventional means. To overcome this drawback, an automatic fault detection system need to be design to classify above mentioned faults. In this research, an improved Artificial Intelligence (AI) based PV array fault detection system is designed in MATLAB/SIMULINK. The system optimized the features of input parameters (irradiance, voltage, current), using nature inspired optimization approach named as Genetic Algorithm (GA). Based on the optimized features, Artificial Neural Network (ANN) is trained and validated. Using ANN, if for any module find power reduction above the defined value, then considered it as faulty PV cell, and then excluded that cell from the entire PV module. In this way, the performance of the system is increases and hence improves the efficiency of the PV array system. Keywords - Fault Detection, Material Faults, Photovoltaic Model, Solar Cell, Genetic Algorithm, Artificial Neural Network.