Research on Photovoltaic System Faults Detection based on Deep Learning

Photovoltaic system is an essential component for electrical network and generate the power for society. However, the photovoltaic cell may occur the faults due to aging of facilities and cause the loss of electrical power system. Therefore, faults detection methods are essential for the stability and continues supply of electric power. Existing detection methods are based on the measurable values or utilizing a remember network to investigate faults points, which causes numerous workloads and may contain several mistake detection. In this paper, we establish a deep learning algorithm through training enough amount training process. The model aims to achieve automatically identify the faults in the photovoltaic system, which can predict the fault occurrence probability based on the history data input and assist the system to accomplish self-detection while executing the photovoltaic system. From our extensive experimental results and comparison performances, we can conclude that our proposed mechanism can achieve the system faults detection with more than 80% accuracy with reasonable training costs.