关键词: 故障定位, 相模变换, 北极海鹦优化算法, 多元变分模态分解, Teager能量算子, Transformer模型

Abstract:

To address the low accuracy and reliability of existing power cable fault-location methods, a new cable fault-location method based on APO-MVMD-Transformer is proposed in this study. The collected cable fault signal is decoupled through phase-mode transformation to obtain the fault traveling wave signal. The Arctic puffin optimization (APO) optimized multivariate variational mode decomposition (MVMD) parameters are then used to decompose this traveling wave signal. The Teager energy operator is applied to extract instantaneous energy variations in the high-frequency modal components, allowing the identification of the wavefront arrival times. The sampling points corresponding to these times are then used as feature values to construct a feature dataset, which is further employed to optimize the APO-Transformed-based model. Finally, feature dataset are input into the APO-Transformer-based location model to locate cable faults. Results showed that the proposed model exhibited a coefficient of determination as high as 0.999 91, a relative fault-location error within 1%, and a distance error within 100 m, thereby demonstrating high fault location accuracy.

Key words: fault location, phase-mode transformation, Arctic puffin optimization, multivariate variational mode decomposition, Teager energy, Transformer