摘要
为了提高变压器复合故障诊断精度,提出了一种基于改进哈里斯鹰(Improved Harris Hawk Optimization, IHHO)算法优化概率神经网络(Probabilistic Neural Network, PNN)的变压器复合故障诊断方法。采用Tent映射、非线性调整逃逸能量和小孔成像学习策略对哈里斯鹰优化(Harris Hawk Optimization, HHO)算法进行改进,以增强IHHO算法的优化性能,避免算法陷入局部最优。采用IHHO算法对PNN的平滑因子进行优化,建立了基于IHHO-PNN的变压器故障诊断模型。利用实际运行的变压器故障数据进行仿真分析。结果表明,所提出的IHHO-PNN模型在进行变压器故障诊断时出现错误诊断的次数更少,诊断精度更高,变压器故障诊断效果好于其他几种对比模型,验证了该变压器复合故障诊断方法的实用性和有效性。
In order to improve the accuracy of transformer composite fault diagnosis,a method based on improved Harris hawk optimization(IHHO)algorithm probabilistic neural network(PNN)is proposed,in which Harris hawk optimization(HHO)algorithm is improved by using Tent mapping,non-linear adjustment of escape energy,and small hole imaging learning strategy to enhance the optimization performance of the IHHO algorithm and avoid falling into local optima.IHHO algorithm is applied to optimize the smoothing factor of PNN and a transformer fault diagnosis model is established based on IHHO-PNN.Actual transformer fault data are used for simulation analysis and the results show that the proposed IHHO-PNN model has fewer misdiagnosis occurrences,higher diagnostic accuracy,and better transformer fault diagnosis performance than other comparative models,verifying the practicality and effectiveness of this transformer composite fault diagnosis method.
作者
杨威
万文欣
陈柏寒
李巧玲
YANG Wei;WAN Wenxin;CHEN Baihan;LI Qiaoling(State Grid Hubei Electric Power Co.,Ltd.Jingmen Power Supply Company,Jingmen 448000,China)
出处
《安徽电气工程职业技术学院学报》
2024年第2期28-36,共9页
Journal of Anhui Electrical Engineering Professional Technique College
关键词
变压器
复合故障
改进哈里斯鹰算法
概率神经网络
transformer
composite fault
improved Harris hawk optimization algorithm
probabilistic neural network
