基于鲸鱼算法优化PNN的变压器故障诊断

Transformer Fault Diagnosis Based on Whale Algorithm Optimized PNN

为了增强变压器故障诊断中概率神经网络(probabilistic neural network,PNN)的诊断能力,提出了一种鲸鱼优化算法(whale optimization algorithm,WOA)与PNN相结合的算法。采用WOA对PNN神经网络的平滑因子参数进行优化,并将优化得到的参数值代入到PNN诊断模型中,通过网络训练得到变压器故障诊断的最佳网络模型。试验结果表明,WOA-PNN算法对变压器的故障诊断能力较强,准确率达到了93.88%,网络训练能力约为97%,此方法具有快速收敛、避免陷入局部极小的能力。与传统PNN相比具有明显的优势,为变压器稳定运行提供强大的技术支撑。

In order to enhance the diagnostic ability of the probabilistic neural network(PNN)in transformer fault diagnosis,a combination of whale algorithm(WOA)and PNN was proposed.The whale algorithm was used to optimize the smoothing factor parameters of the PNN,and the optimized parameter values were substituted into the PNN diagnosis model,and the best network model for transformer fault diagnosis was obtained through network training.The experimental results show that the WOA-PNN algorithm has a strong ability to diagnose transformer faults,with an accuracy rate of 93.88%and a network training ability of about 97%.This method can not only converge quickly,but also greatly improve the ability to avoid falling into local minima.Compared with the traditional PNN,it has obvious advantages and provides strong technical support for the stable transformer operation.