Estimation of time-varying power quality indices with an adaptive window-based fast generalised S-transform

The S-transform (ST) finds widespread application in non-stationary signal analysis. However, the relatively high computational complexity of the ST remains as a challenge. Further, the optimum choice of the window function and the discretisation side effects of the ST need to be addressed for accurate time-frequency localisation. This study proposes a fast adaptive discrete generalised ST (FDGST) algorithm based on a new frequency scaling named selective frequency scaling, window cropping and an adaptive window function. The proposed algorithm optimises the shape of the window function for each analysis frequency to improve the energy concentration of the time-frequency distribution, and applies folded window functions to minimise aliasing affect owing to discretisation. Further, the algorithm is applied for analysis and measurement of parameters in various types of power quality waveforms. Standard transient and steady-state indices calculation from the FDGST analysis is also illustrated. The improved performance of the proposed algorithm is supported by simulations using synthetic as well as practical signals.