混沌免疫优化神经网络的变压器故障诊断  

Chaos Immune Optimization Neural Network Transformer Fault Diagnosis

作　　者：贾花萍[1] 李尧龙[1] 何小虎[1] 

机构地区：[1]渭南师范学院网络安全与信息化学院,陕西渭南714099

出　　处：《渭南师范学院学报》2017年第16期35-40,共6页Journal of Weinan Normal University

基　　金：陕西省自然科学基础研究计划项目:基于支持向量机和改进粒子群算法相结合的网络安全态势预测方法(2017JM6110);陕西省军民融合研究基金项目:SAR图像边缘检测技术在关中城市群变化检测中的应用研究(17JMR28);渭南市科技创新扶持资金项目:BCI-FES对脑卒中运动功能恢复的应用研究(2016KYS-3-3);渭南师范学院自然科学重点研究项目:(L;M)拟阵的模糊化与模糊拟阵的算法研究(17YKF01)

摘　　要：对变压器进行故障诊断,在提高电力系统运行的安全性和可靠性方面具有重要意义。传统的BP算法是一种局部搜索的优化方法,在变压器故障诊断的过程中很有可能陷入局部极值,使收敛速度较慢,算法效率较低。文章提出采用混沌免疫优化神经网络方法进行变压器故障诊断,在网络权值调整过程中引入免疫算法的思想,采用免疫算法对网络权值进行全局优化,最后用该算法对变压器进行故障诊断,并与传统BP神经网络诊断方法进行比较。结果表明,将该方法用在变压器故障诊断中,其正确率提高了5%,训练速度减少了182次,收敛速度更快,可靠性更高。The fault diagnosis of transformer is of decisive significance to improve the safety and reliability of power system op-eration. As a result of the BP algorithm is a kind of local search optimization method,in the process of transformer fault diagnosis, the algorithm is likely to fall into the local extreme,making the convergence speed slow and the efficiency low. The fault diagnosis method of chaos immune optimization of neural network introduces the immune algorithm in the process of network weights of thought,using the immune algorithm of network weights for global optimization. Finally the algorithm for fault diagnosis of transform-er with the traditional BP neural network diagnosis method is compared. The results show that the method is used in transformer fault diagnosis,which makes the accuracy rate increased by 5%,182 times of the training speed reduced,and faster convergence speed and higher reliability achieved.

关 键 词：混沌免疫 神经网络 变压器 故障诊断 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]