Research On Photovoltaic Module Fault Diagnosis Based On Data Driven

Because of the pollution resulted from the deterioration of the global climate issues,the evolution of clearer energy is impending.China’s solar radiation energy reserves are huge,and its development potential is extremely broad.As the core part of utilization of solar radiation energy,photovoltaic power generation has the advantages of cleanness and low cost.Due to the high technical difficulty in manufacturing pv modules and the complex and variable working environment,failures of photovoltaic plant are unavoidable.This not only influences the safe working of photovoltaic plant,but also leads to safety accidents.Accurate diagnosis of pv module failures is the most important measure to reduce the losses of photovoltaic plant.The existing fault diagnosis methods mainly focus on the diagnosis of single failures of pv modules.To solve the diagnosis problem of pv module composite faults,this paper proposes a data driven fault diagnosis ways for pv modules,and establishes a pv module fault diagnosis model according to adaptive sparrow search algorithm to optimize the parameters of hybrid kernel limit learning machines,which can accurate diagnosis of single or composite failures of pv modules.The main research contents of this paper are as follows:Firstly,by comparing and analyzing the accuracy of two mathematical models of pv cells,the selection of pv modules was realized.The model with low error was selected and packaged as pv modules in MATLAB/Simulink.A pv array simulation model was built.Based on the export characteristic curves under normal,single fault,and composite fault conditions,the fault feature variables of pv modules under distinct factors were extracted,and the pv module fault symptom set was built.Secondly,A fault diagnosis way for pv modules according to Hybrid Kernel Extreme Learning Machine(HKELM)was proposed for single or composite failures of pv modules.The hybrid kernel extreme learning machine is constructed using polynomial and RBF kernel functions as the kernel functions of the extreme learning machine,and the sparrow search algorithm is upgraded by introducing an adaptive policy,using the upgraded Adaptive Sparrow Search Algorithm(ASSA)to optimize the parameters of HKELM,a pv module fault diagnosis model based on ASSA-HKELM was built.Compared with the fault diagnosis results of pv modules using the improved algorithm before and after,the effectiveness of the model was verified through simulation analysis of actual operation data of a photovoltaic plant.Simulation results indicate that this method can accurately diagnose single and composite faults of pv modules.Finally,a pv module fault diagnosis GUI system was established through the MATLAB GUI,which basically covers all the functions of pv module fault diagnosis.The system is easy to operate and well visualized.Users can freely choose fault diagnosis algorithms based on their own needs,and can manage the data in the database.The system was tested through an example,verifying the application value of the system.