Abstract
The power supply line of the coal shearer cutting unit contains stable high-order harmonic components and high-frequency noise, which are generated by its non-linear load and power electronic drive device. When the power supply line experiences a fault, faint transient changes mix significantly with the existing high-order harmonics and high-frequency noise. This presents significant challenges for waveform detection and fault location. A fault-induced detection method based on Pigeon Inspired Optimization-Variational Mode Decomposition-Envelope Derivative Operator (PIO-VMD-EDO) is proposed. By employing fuzzy entropy (FE) as the optimization objective for PIO, the PIO-optimized VMD algorithm enables adaptive decomposition of power line traveling waves. Meanwhile, the enhanced EDO mechanism facilitates wavefront calibration and signal amplification. The proposed method was tested and validated based on actual noise signal characteristics and a model of the long-distance power supply system for coal shearer. Results indicate that compared with the discrete wavelet transform (DWT), continuous wavelet transform (CWT), hilbert-huang transform (HHT), and empirical mode decomposition (EMD), the proposed method achieves higher fault location accuracy, with a maximum relative error not exceeding 2.5%. Additionally, it remains unaffected by factors such as fault conditions, noise interference, and sampling rates. When applied to experimental platforms for cable fault location, this method achieves precise fault pinpointing, promising to significantly enhance the reliability of fault location in coal mine power grids. It is particularly well-suited for scenarios involving nonlinear loads and power electronic equipment.