Remaining Life Prediction of High-Voltage Circuit Breakers Based on Optimized Particle Swarm-Bidirectional Long Short-Term Memory Neural Network

Magnetic control mechanism high-voltage circuit breakers play a crucial role in power systems, but their long-term operation may be affected by various factors, leading to degradation and failure. To better predict the remaining service life of circuit breakers, this study proposes a novel method that integrates signal processing and deep learning techniques. First, the collected current signals are smoothed using the Gaussian filtering method to reduce waveform interference, allowing for more accurate extraction of current degradation features. The time series of the health index is then reconstructed using a sliding time window. Next, the particle swarm optimization (PSO) algorithm is used to train a bidirectional long short-term memory neural network (BiLSTM), establishing a model for predicting the lifespan of magnetic control mechanism high-voltage circuit breakers. The PSO-BiLSTM model can effectively predict the degradation trend of the circuit breakers and determine their remaining service life based on a set failure threshold, providing important reference data for machine maintenance and management. Experimental results show that, compared to traditional prediction models, the proposed method offers higher prediction accuracy and reliability, providing an effective means for ensuring the stable operation and safety of power systems.