Sound Recognition Algorithm for Power Devices Based on Substation Inspection Robots

Abstract

Transformer and high reactance are two important devices in power substations, and the running sound has the information of their running state. But inspection robots don’t have the function of sound recognition. A new sound recognition algorithm based on harmonics feature and vector quantization for power substation devices was proposed. Inspection robots is adopted to record and sample the running sound of the devices, the 27 harmonics in [0Hz,1300Hz] is adopted as the feature vector, the LBG algorithm is adopted to train the optimal codebook for VQ, and the running state of the transformer and high reactance was recognized with the high recognition ratio 99%.