Fault Diagnosis Of Grid-connected Inverter In Closed Loop State

With the rapid development of the modern wind power industry,the automation of equipment continues to increase.Once the closed-loop system of the grid-connected inverter fails,it will bring unpredictable damage to the entire power generation system.Therefore,in order to reduce the impact on human life and public property,it is necessary to diagnose the closed-loop system of the grid-connected inverter to ensure the stability and safety of the wind power system.At present,many grid-connected inverter fault diagnosis studies are carried out for open-loop systems,and the influence of feedback control law on grid-connected inverter fault diagnosis is not considered.Therefore,the fault diagnosis method for grid-connected inverter open-loop system is used.It may no longer be suitable or needs to be redesigned in a closed loop system.In this paper,the closed-loop system of grid-connected inverter is taken as the research object,and the method based on analytical model and the method based on artificial intelligence are used for fault isolation and location.Firstly,this paper analyzes the research methods of fault diagnosis technology and closed-loop system fault diagnosis technology by domestic and foreign scholars.Based on the previous ideas,a fault diagnosis method based on analytical model is proposed.The model of the double closed-loop system of the grid-connected inverter was built,and the experiment was carried out on the Matlab simulation platform to realize the simulation waveform of the output current of the grid-connected inverter based on double closed-loop control under various fault conditions.The output current was found through analysis and comparison.The waveform can be used as an input to the fault diagnosis system.Secondly,this paper proposes a fault diagnosis method based on phase current.The phase current generated by the Park transformation in the closed-loop control loop is compared with the phase current of the measured system,and the arithmetic processing is performed to generate the normalized phase current average and average absolute.The value is based on the residual result and the diagnostic rule to isolate and locate the fault.The simulation proves that the diagnostic method can quickly detect and identify single-tube and multi-tube open-circuit faults for the grid-connected inverter under closed-loop control.Finally,due to the random weights and thresholds used by the Extreme Learning Machine(ELM),the diagnostic accuracy is not high.Therefore,Particle Swarm Optimization(PSO)is used to optimize the initial weights and thresholds of the extreme learning machine.The fault eigenvector of the output current is extracted by wavelet transform,and then the extracted fault eigenvector is input into the diagnostic system of Particle Swarm Optimization Extreme Learning Machine(PSO-ELM)for fault diagnosis.The simulation results show that the method has strong approximation ability,high precision and improved diagnostic reliability.