In view of the frequent faults of the existing photovoltaic array

simple monitoring and fault diagnosis can not accurately identify and locate the fault

resulting in an increase in operation and maintenance costs and a reduction in the economic benefits of photovoltaic power generation. To solve this problem

this paper proposes a photovoltaic system fault identification strategy based on long-short term memory(LSTM) neural network

establishes the mathematical model of photovoltaic array

and builds the photovoltaic system simulation model in MATLAB/Simulink environment

by using the maximum power point tracking(MPPT) algorithm based on disturbance observation and the closed-loop control strategy based on grid voltage. The fault causes of photovoltaic module such as open circuit and occlusion are analyzed

and the fault state of photovoltaic array is simulated in MATLAB

the influence of different fault states on the output characteristics of photovoltaic system is studied

and the fault characteristic parameters are obtained. The LSTM deep learning fault diagnosis model is established

and the characteristic parameters of photovoltaic array under different fault conditions are collected as training samples to train the model. It shows the LSTM neural network model has higher accuracy than the BP neural network. The photovoltaic experimental platform is applied to simulate different photovoltaic array faults

and the fault characteristic parameters are input into the LSTM neural network fault diagnosis model for diagnosis. The results show that

the LSTM depth learning fault diagnosis model can accurately diagnose and locate the PV array faults.