Adaptive Dynamic Q-Factor Wavelet Transform For Natural Frequency Extraction In Fault-Originated Transient Signals

This paper presents an adaptive dynamic Q-factor continuous wavelet transform (ADQ-CWT) algorithm for accurate natural frequency identification in power system fault transient analysis. The proposed approach dynamically adjusts the Q-factor of the analysing wavelet by modifying its centre frequency and time-decay parameter, thereby optimising the trade-off between time and frequency resolution across multiple scales. To further enhance temporal localisation and suppress redundant low-frequency components, multi-resolution analysis (MRA) filtering is integrated into the framework. The algorithm is validated through electromagnetic transient simulations conducted in EMTP-RV using the IEEE 34-bus distribution network. The evaluation considers three representative aspects: variations in network topology, fault types, and fault locations. Results demonstrate that the proposed algorithm consistently identifies natural frequencies with high precision, yielding deviations below 0.1% from the theoretical values. Moreover, the algorithm exhibits strong robustness to changes in system configuration and fault conditions.