Research on Mechanical Fault Diagnosis Method of Circuit Breakers Based on Fusion of Multi-Source Signal Data

Propose a circuit breaker mechanical fault diagnosis method based on the fusion of multi-source signal data, integrating circuit breaker operation mechanism current, mechanical vibration, and contact travel signals. Construct an experimental platform and simulate various circuit breaker mechanical faults, measure and store experimental data from current, vibration, and travel sensors. For each fault state measured, research feature extraction methods for the three types of single-source signals, analyze the curves of each signal and multi-source feature data, and use box plots for multi-source feature selection, fusion, and restructuring to obtain optimal diagnostic feature vectors for each fault state. Establish a circuit breaker fault diagnosis model and divide it into fault diagnosis subsystems. Use machine learning algorithms to establish fault diagnosis models for subsystems, and validate them using large sample data to achieve accurate identification of circuit breaker mechanical faults. Validate the system’s fault diagnosis performance by simulating circuit breaker mechanical faults.