Partial Discharge Fault Diagnosis of Switchgear Based on APSO-BP Algorithm

Diagnosis of partial discharge fault types of switchgear is the focus of the early warning of switchgear faults at this stage, which is of great significance to ensure the normal operation of the switchgear. Aiming at this problem, an adaptive particle swarm optimization (APSO) optimized BP neural network partial discharge fault diagnosis algorithm is proposed. By optimizing the inertia weight formula in the standard particle swarm, at the same time introducing genetic factors, mutation factors and time factors to accelerate the convergence speed of the particle swarm algorithm, so as to improve the performance of finding the optimal threshold and weight. First, the partial discharge signal is denoised, and then the signal features are extracted, and the dimensionality is reduced to 3-dimensional features through the principal component analysis algorithm, and finally the fault diagnosis is performed through the algorithm. Comparing the diagnosis results of different algorithms, it can be seen that the fault recognition rate of the APSO optimized BP neural network algorithm is about $5 \sim 15$% higher than other algorithms, and the convergence speed and convergence accuracy are both improved, which proves the proposed APSO-BP algorithm