LF-PSO优化神经网络的高压断路器机械故障诊断

目前对高压断路器的故障诊断方法较多，其中采用神经网络方法居多。提出一种基于莱维飞行粒子群算法（LF-PSO）优化PNN神经网络的故障诊断技术。PNN结构简单，收敛速度快，但其中平滑因子σ对网络输出结果正确性影响较大，采用改进的粒子群算法对σ进行寻优。在标准粒子群基础上加入LF 能有效地使粒子通过随机游走产生新的解，经历新的搜索路径和领域，从而增加了种群的多样性，提高发现更优解的概率，不易陷入局部极值，提高了搜索的速度。通过实验数据验证，LF-PSO优化的PNN算法加快了搜索的速度，提高了诊断的精度，减小了误差，分类效果明显，是一种有效的故障诊断方法。

Mechanical Fault Diagnosis of High Voltage Circuit Breakers Based on LF-PSO Optimized PNN Neural Network

At present,there are many methods of fault diagnosis for high voltage circuit breakers,among which neural network is the most common by used method.A fault diagnosis technique based on LF-PSO optimized PNN neural network is presented.PNN has simple structure and fast convergence speed,but the smoothing factor has a great influence on the correctness of network output.The improved particle swarm optimization (PSO) algorithm is used to optimize the smoothing factor.Adding LF to the standard PSO can effectively make the particles generate new solutions through random walks and experience new search paths and domains,thus increasing the diversity of the population,improving the probability of finding better solutions,not easily falling into local extremum,and improving the search.The experimental results show that the LF-PSO optimized PNN algorithm speeds up the search,improves the diagnosis accuracy,reduces the error,and has obvious classification effect.It is an effective fault diagnosis method.