A Shadow Fault Diagnosis Method Based on the Quantitative Analysis of Photovoltaic Output Prediction Error

Solar energy plays an increasingly important role in new energy sources, the stable operation of photovoltaic (PV) generation system for the entire energy supply system is gradually highlighted. In this article, the fault types of PV modules are classified into temporary and permanent shadow faults. A fault feature based on the prediction error of PV output and a novel intelligent quantization fault diagnosis method are proposed. First, the PV output sequence is processed by empirical mode decomposition and fine-to-coarse to remove the second-minute disturbance. Then, the clockwork recurrent neural network is used to predict the processed PV output to construct fault features. Finally, the support vector machine is used to identify the fault, so as to realize the diagnosis of shadow fault. The experimental results prove the effectiveness of the proposed diagnosis method, providing a new idea for the related research of PV system fault diagnosis, and further ensure the stable operation of the system.