Research On Operation Condition Monitoring And Diagnosis Of Photovoltaic System Based On Deep Learning

In recent years,with the rapid development of economy,the demand for energy is increasing day by day.Solar energy,as a clean and renewable energy,has developed rapidly.As the smallest power generation unit in the photovoltaic system,photovoltaic panels often work in the outdoor environment and are vulnerable to the erosion of rain and high temperature,coupled with the aging of their own materials,resulting in frequent failures in the photovoltaic array.Simple monitoring and fault diagnosis cannot accurately identify photovoltaic faults,which increases the costs of operation and maintenance,and affects the economic benefits of photovoltaic power generation.Therefore,this paper proposes a research strategy based on deep learning to monitor and diagnose photovoltaic faults.Firstly,combined with the theory of material physics and electronics,the mathematical model of photovoltaic panel is established,and the simulation model of photovoltaic power generation system is built,using the maximum power point tracking(MPPT)algorithm of disturbance observation and the closed-loop control strategy based on grid voltage.The typical faults of photovoltaic array can be simulated in Simulink,and the output parameters of photovoltaic system can be collected as training samples.Secondly,the faults that may occur in the operation process of photovoltaic panels are analyzed,the output data of photovoltaic panels under different fault conditions are collected based on the photovoltaic experimental box,and then the characteristic parameters of photovoltaic faults are obtained.Long-term and short-term memory(LSTM)and back propagation(BP)neural network are selected to diagnose photovoltaic fault,and the faults diagnosis process based on two kinds of neural network is described in detail.The Particle Swarm Optimization(PSO)algorithm is introduced to optimize the parameters of LSTM neural network.Then the operation status monitoring platform of photovoltaic power generation system is designed to collect the output parameters of photovoltaic power generation system in outdoor environment.The photovoltaic power generation system mainly includes three parts: hardware circuit,system program and software interface.The hardware circuit needs to complete the circuit diagram design of power supply module,detection circuit and boost circuit,as well as the selection of components and circuit board welding.The program design of the system is divided into two parts: the lower computer and the upper computer,the lower computer needs to complete the program design of photovoltaic panel data acquisition,MPPT control,temperature acquisition and light intensity acquisition.On this basis,the wireless communication module is integrated to complete the wireless transmission program design.The program design of the upper computer needs to complete the Wi Fi data receiving,data prediction processing,and control command issuing and so on.The software interface is based on the Pycharm development environment,and uses Python language to design the startup interface,real-time curve interface,fault detection interface,and data query interface.Finally,the photovoltaic system data under different fault types are collected by using the photovoltaic system simulation model and operating condition monitoring platform,LSTM and BP neural networks are used to identify and locate photovoltaic faults,and PSO is used to optimize the parameters of LSTM model,so as to improve the performance of fault diagnosis.Through simulation and experiments,the accuracy of the fault monitoring and diagnosis strategy based on PSO-LSTM model and the feasibility of the scheme are verified.