Research On Fault Diagnosis Methods Of Micro-grid Multi Inverter Parallel

Micro-grid is the natural extension of distributed power generation development,but also the latest and most effective form of distributed generation.With the large number of micrograph applications,people have a high demand for the maintainability of micro-grid system.In the micro-grid,the inverter fault problem is more and more prominent,and because of its failure and the impact is enormous,so the study of micro-grid environment,the inverter fault detection and diagnosis of great significance.Based on the analysis of the most common fault feature extraction and diagnosis algorithm of the inverter in recent years,the fault characteristics and fault performance of the inverter in a single inverter and micro-grid system are analyzed in detail.This paper presents a fault diagnosis mechanism under the micro-grid platform.On this basis,wavelet transform and sparse coding algorithm are used in the fault feature extraction.Finally,the BP neural network classifier is used to complete the fault of the inverter in the micro-grid system diagnosis.Based on the structure and working principle of micro-grid system,the simulation model of micro-grid system is established under MATLAB / Simulink,and the correctness of platform is verified,which lays the foundation for subsequent fault analysis and diagnosis.Then,the fault characteristics and performance of the single inverter are analyzed and the simulation waveforms of the typical faults are given before the fault analysis of the inverter under the micro-grid multi-inverter parallel system.The fault performance and fault type and The time of the fault occurs;based on the further analysis of the inverter under the micro-grid,it is concluded that the inverter module in the micro-grid system in the micro-grid mode of independence and off-grid The relevance of the model illustrates the complexity of the fault diagnosis.Therefore,this paper proposes that each inverter in the micro-grid should have the mechanism of fault self-test,and this mechanism should be a kind of fault diagnosis mechanism which can be compatible with both modes and exclusive fault capability.In order to obtain the essential characteristics of fault performance,the fault feature extraction algorithm is needed.In this paper,the wavelet transform and sparse coding algorithm are used to extract the characteristic of the inverter under the micro net,and the respective eigenvectors are obtained to obtain the total sample set.Then,two BP neural network classifiers were trained and trained with the same number of test samples.The results show that based on the two algorithms,the BP neural network is trained by the training set of the same number.The accuracy of the fault diagnosis mechanism is verified,and the fault feature vector obtained by the sparse coding shows a higher degree of discrimination.The accuracy of the diagnosis result is also higher,In the fault diagnosis of the inverter,the sparse coding algorithm is more reasonable and desirable.