An Ensemble Adaptive Deep Learning Method for High-Voltage Circuit Breaker Mechanical Fault Diagnosis

In recent years, data-driven intelligent diagnosis methods have made rapid progress in the field of high-voltage circuit breaker mechanical fault diagnosis. However, most of the existing fault diagnosis methods are based on a single signal, which cannot make full use of the state information of high-voltage circuit breakers. To address this problem, this paper proposes a novel ensemble one-dimensional convolutional neural network (1DECNN) for the intelligent mechanical faults diagnosis of high-voltage circuit breakers. Firstly, multiple sensors are used to obtain the vibration signal, breaking coil current signal and moving contact travel signal form high-voltage circuit breaker. Then, a multi-resolution CNN is proposed to realize multi-sensor information fusion diagnosis. The proposed method extracts the fault characteristics of the characterized high-voltage circuit breaker with a 1D CNN, and uses ensemble learning to realize comprehensive mechanical fault diagnosis with multiple data. The experimental results show that the 1DECNN can achieve high-precision robust mechanical faults diagnosis of high-voltage circuit breakers and effectively fuse the different information compared with the traditional method.