Intelligent Fault Detection and Self-Diagnosis Module Design for Pole-Mounted Switches Based on Multi-Signal Fusion

This paper proposes an intelligent fault diagnosis system for pole-mounted switches using multi-signal fusion to address limitations like single-signal dependence and high false-alarm rates. The system integrates current, voltage, temperature, and vibration sensors with edge computing for real-time signal processing and state recognition. Wavelet denoising, PCA, and SVM are employed for feature extraction and health assessment. Supporting LoRa, RS485, and 4G communications, the system achieves 96.4% accuracy with <3% false alarms in field tests, demonstrating strong practicality for smart grid applications.