Probabilistic Pca-Support Vector Machine Based Fault Diagnosis Of Single Phase 5-Level Cascaded H-Bridge Mli

In present era, Multilevel Inverters (MLIs) are very popular in many industrial and renewable energy applications. The fast and accurate fault diagnosis is very important for improving the reliability. The present study proposes a novel fault diagnosis method based on the Probabilistic Principle Component Analysis (PPCA) and Support Vector Machine (SVM) for controlled switches in single phase Cascaded H-Bridge Multilevel Inverter (CHMLI). The output voltage signals under different fault conditions of the CHMLI are taken as fault features by using Phase Shift PWM technique. PPCA is used to optimize the data and reduce dimension of fault features. Finally, SVM classifier is used to diagnose the different fault modes. An experimental setup of CHMLI has been designed to validate the proposed fault diagnosis method. The simulation and experimental results show that by using PPCA-SVM, we can improve the accuracy of the fault location and reduce the time taken to diagnosis the fault in CHMLI.