Mechanical Fault Diagnosis of High-Voltage Circuit Breakers Based on IPSO-VMD and KFCM-SVM

Abstract

Due to that the complex mechanical faults of high-voltage circuit breakers and the difficulty in extracting fault features, a fault diagnosis method combining Improved Particle Swarm Optimization enhanced Variational Mode Decomposition (IPSO-VMD) with Kernel Fuzzy C-Means and Support Vector Machine (KFCM-SVM) is proposed. Initially, the vibration signals are decomposed using IPSO-VMD to obtain a series of Intrinsic Mode Functions (IMFs) that reflect the mechanical state information during the circuit breaker operation. Then, Hilbert transform is performed on each IMF component to construct the corresponding Hilbert marginal spectrum, and the energy entropy is obtained as the feature vector. Finally, the KFCM is used to pre-classify the features, and the SVM is used to establish the training model to realize the mechanical state identification. Experimental results indicate that the energy entropy of the Hilbert marginal spectrum of vibration signals is sensitive to changes in the mechanical state of high-voltage circuit breakers, and KFCM-SVM can accurately identify mechanical faults during the circuit breaker tripping process. © 2025 Institute of Electrical Engineers of Japan and Wiley Periodicals LLC.