Nonstationary waveform detection in power quality signals using multisynchrosqueezing transform

Abstract

Time-frequency (TF) analysis (TFA) is widely used as a reliable signalprocessing tool for the detection and identification of time-varying waveforms in power quality (PQ) signals. However, the effectiveness of a TFA method is limited by Heisenberg uncertainty principle. The presence of cross terms is also a significant issue in TF analysis of real-time signals. In PQ signal analysis, it is highly necessary to apply a suitable TFA technique with higher clarity and fewer cross-term problems. This paper introduces a newly developed TFA method known as multisynchrosqueezing transform (MSST) to detect the presence of nonstationary waveforms in PQ disturbance signals. The MSST is a high-resolution TFA method, which is similar to synchrosqueezing transform (SST) but has an iterative reassignment process to enhance the energy concentration of the TF representation. We apply the MSST to various types of single and combined PQ disturbance signals and show its potential application to PQ disturbance signal analysis. Results show that MSST provides higher TF resolution compared to other contemporary methods.