Abstract：Low-frequency oscillation transient is a common type of power quality disturbance in power system, directly affecting the safe and stable operation of power system. A low-frequency oscillation transient detection method based on an improved S-transform using an optimized window function is proposed. Firstly, the changing characteristics of the window function with detection frequency in time-frequency analysis algorithms are analyzed, identifying an optimized window function suitable for detecting low-frequency oscillation transients. This optimized window function is then utilized to construct an improved S-transform algorithm. Secondly, through the application of the convolution theorem, Fourier transform, and its inverse transform, a computational expression for the fast implementation of the improved S-transform algorithm is derived, resulting in a two-dimensional time-frequency matrix containing signal amplitude and phase information. The implementation process of the improved algorithm is provided. Finally, the improved algorithm is tested using power grid signals containing low-frequency oscillation transients. The algorithm demonstrates relatively optimal time-frequency energy concentration performance for low-frequency oscillation transients. In practical experiment, the detected disturbance center frequency is 600 Hz, consistent with the generation frequency of the disturbance, verifying the feasibility and effectiveness of the improved algorithm.
2025, Vol. 46 
