Research On Open-Circuit Fault Diagnosis Method Of Inverter Based On Data Drive

With the development of power electronics technology,inverters are widely used in electric vehicles,aerospace,and household appliances.However,due to the complex environment,frequent switching,high voltage,and high current impact,the inverter is prone to failure in actual operation.Once the inverter fails,it will affect system performance,fail to operate,and even cause serious safety accidents.Therefore,it is essential to diagnose the inverter fault quickly and accurately.This paper studies the open circuit fault diagnosis of power electronic devices in the inverter.The main contents are as follows:（1）Firstly,the topology and working principle of the two-level inverter are analyzed,establish the simulation model of the two-level inverter,and the fault simulation and fault type analysis are carried out.The open-circuit faults of power electronic devices are divided into four categories: single tube fault,cross tube fault,peer tube fault,and same column tube fault.Finally,the current data of IGBT fault is obtained on the MATLAB/Simulink simulation platform.It provides a basis for the subsequent research on the open circuit fault diagnosis method of inverters based on the data drive.（2）Aiming at the problem that the traditional park transform diagnosis method is easily affected by noise and load disturbance,a fault diagnosis method of Park-FFT-K-means is proposed in this paper.Firstly,the current data of inverter under normal and fault conditions are obtained,then the vector graphics and the Fourier graphics of switch angle are obtained through park transform to extract practical fault eigenvalues,and then the fault feature classification is completed by K-means clustering algorithm.Finally,the fault diagnosis method of Park-FFT-K-means proposed in this paper is verified on the platform of MATLAB/Simulink.The experimental results show that this method has high diagnostic accuracy and fast diagnosis speed.（3）Aiming at the problem that the traditional fast Fourier transform can only represent the frequency domain information,but can not represent the local information of the signal,a wavelet SVM fault diagnosis method is proposed in this paper.Firstly,the current data of the inverter under normal and fault conditions are obtained,and then Park dimensionality reduction is carried out.Then the energy values of each frequency band are obtained by db3 wavelet transform,and the fault feature vector is obtained by normalization.Then the fault feature classification is completed by the gray wolf optimization algorithm（GWO）-SVM classification method.Finally,the feasibility of this method is verified in the MATLAB/Simulink platform.The experimental results show that compared with the fault diagnosis method of Park-FFT-K-means,the fault diagnosis method based on wavelet SVM has higher diagnosis accuracy.（4）Finally,the physical verification platform of the motor drive system with inverter is built,the overall design,hardware design,and software design of the platform are completed,and the current data of the inverter under normal and fault conditions are obtained.On this basis,the two methods proposed in this paper are used for fault feature extraction and fault feature classification.The experimental results verify the feasibility and effectiveness of the two diagnosis methods.