Research on Peak Detection Method of Keyphasor Signal

This research presents an improved peak detection method tailored for the keyphasor signal of steam turbines. By employing first-order differencing and specialized treatments on the signal, the method effectively highlights waveform features while addressing issues related to consecutive values and boundaries. Subsequently, peaks are searched and screened using set thresholds and minimum peak distances, eliminating false peaks and enhancing detection accuracy. Additionally, the study introduces an adaptive parameter calculation approach to optimize the process. Comparative experiments demonstrate that this method performs comparably to professional methods, offering high accuracy in peak detection and satisfying real-time peak monitoring requirements in terms of algorithm time complexity.