Abnormal Detection in Photovoltaic Array Based on Artificial Neural Network

Abstract

Abnormal condition on photovoltaic array (PV array) such as partially closed surface of the PV module caused by accumulation of dust, bird droppings, or objects falling on it then resulting in non-uniform irradiation on the surface of the PV module. In this paper, abnormal conditions detection algorithm based on Artificial Neural Network (ANN) is proposed. ANN is used to predict the condition of PV array. ANN architecture is designed to consist of one hidden layer by combining the number of neurons. The parameters used are nine power operating points as ANN inputs and logic as ANN outputs. The proposed algorithm is tested by simulating a PV array system on PSIM with a simplified C Block to implement the abnormal detection algorithm. The proposed algorithm is proven to be effective in detecting PV array conditions with rapidly changes of irradiation and successfully distinguishes between abnormal and normal conditions.