Research on Local Mean Decomposition and Extreme Learning Machine based Circuit Breaker Fault Diagnosis Method

Abstract

Aiming at the problem that it is difficult to diagnose mechanical faults of high-voltage circuit breakers using sound signals, a circuit breaker fault diagnosis method based on Local Mean Decomposition (LMD) and Extreme Learning Machine (ELM) is proposed. Use the laboratory 10kV indoor vacuum high voltage circuit breaker to perform sound closing data when the closing operation is normal, the screws are loose, the transmission mechanism is jammed, and the closing spring has insufficient energy storage, and the collected data is wavelet denoised and the LMD decomposition of the denoised signal is reconstructed according to the correlation coefficient, and the multi-scale entropy of the reconstructed signal is calculated as the feature vector. The feature vector set is divided into a training set and a test set and input to a limit learning machine for model training and fault classification. The experiments show that the LMD-ELM based high voltage circuit breaker mechanical fault diagnosis method can effectively identify circuit breaker mechanical faults in different states, and provides a new idea for circuit breaker mechanical fault diagnosis.