Fault Detection in Single-Phase Inverters Using Wavelet Transform-Based Feature Extraction and Classification Techniques

A fault classification and identification system for prognosis of single-phase inverters was presented in this paper. This concept involves the resolution and analysis of different technical constraints associated with the operation and integration of the inverters. The work presented is centered on fault identification for single-phase inverters by developing a suitable technique utilizing signal processing and machine learning techniques. This task is approached by observing the behavior of inverters under various faults and the factors affecting them. Thereafter, the output voltage under both normal and faulted operation was subjected to signal processing techniques to observe the features. The obtained features are trained for classification and identification of faults using support vector machines. Simulation was carried out for operation of inverter under different events of a fault. By combining output voltages and their features with wavelet transforms and feature detection, the developed algorithms are capable of capturing the key features of inverters. The results depicted how to perform numerical fault analysis calculations using MATLAB classification learner by adapting the trained data for identifying the fault.