Extraction of Voiceprint Features for Mechanical Abnormal Operating Conditions of High-Voltage Circuit Breakers Based on Generalized S-Transform and Time-Frequency Spectrum Filtering

Aiming at the problem of low fault detection rate and low detection efficiency caused by multiple factors such as complex structure and environmental factors in traditional high-voltage circuit breaker detection methods, which leads to reduced reliability of the power system, a voiceprint feature extraction method based on generalized S-transform and time-frequency spectrum filtering of 10 kV circuit breaker voiceprint features is designed. By using the generalized S-transform to perform time-frequency analysis on the voiceprint signals generated during the operation of circuit breakers, the time-varying frequency characteristics of the signals were obtained, achieving effective capture of subtle states of circuit breakers under different fault conditions. At the same time, by designing specific time-frequency spectrum filters, the results of the generalized S-transform were filtered to enhance fault characteristics while effectively suppressing noise interference. To verify the effectiveness and correctness of the detection method designed in this article, an experimental platform was built for experimental verification, and compared horizontally with other methods. The experimental results showed that this method not only effectively extracts the voiceprint features of high-voltage circuit breakers under different fault states, providing favorable technical support for fault diagnosis, but also improves the accuracy of circuit breaker fault diagnosis, providing guarantee for the safe and stable operation of the power system.