Abstract：To achieve accurate fault location of distribution lines，a fault location method based on adaptive

time-frequency and energy feature fusion was proposed，addressing the problem of inconsistent features in multibranch distribution lines during cross-feeding and the low precision of traveling wavehead extraction in traditional Fourier methods. This method employed wavelet packet decomposition combined with energy ratio and

information entropy features for cross-feed line normalization. It also utilized the energy change rate to trigger an

adaptive windowed short-time Fourier transform （STFT），allowing for the acquisition of time-frequency

spectrograms and the identification of wave crest moments. By weightedly integrating the time difference of arrival

（TDOA）and the energy decay algorithm（EDA），the localization results were optimized. Research conducted via

Matlab simulations demonstrates that the proposed localization method effectively achieves fault localization under

various signal-to-noise ratios and grounding resistances，thereby enhancing supply reliability.

        Key words：wave propagation positioning；adaptive time-frequency；feature fusion；time difference of arrival

（TDOA）；energy decay algorithm（EDA）

     Format Citation:王志，张志刚，李舒桢,等. 自适应时频与能量特征融合的故障定位方法[J].电气传动,2026,56(09):81-90. WANG Zhi，ZHANG Zhigang，LI Shuzhen,et al. Fault location method based on fusion of adaptive time-frequency and energy features [J].Electric Drive, 2026,56(09):81-90.