本文已被：浏览 1955次 下载 951次
投稿时间：2015-08-27
投稿时间：2015-08-27
中文摘要: 提出了一种基于改进的邻域粗糙集与概率神经网络的水电机组振动故障诊断方法.该方法将邻域粗糙集中的近似精度与信息论观点中的条件熵结合,提出近似条件熵的属性约简算法,减少故障冗余信息,得到最优决策表,并将得到的最优决策表作为概率神经网络(PNN)的训练样本,提高了PNN的训练速度和诊断效率,通过实验证明了所述方法的可行性和有效性.
Abstract:A diagnosis method of improved neighborhood rough sets and PNN is proposed to achieve vibrant fault diagnosis for hydro-turbine generating unit.This method obtains the approximate condition entropy by uniting approximation accuracy of neighborhood rough set and condition entropy of information theory,which reduces the redundant information,acquires the optimal decision table.Then the table is the best decision as probabilistic neural network (PNN) training samples to improve the speed and efficiency of diagnosis.Finally,the experimental analysis and comparison show the feasibility and effectiveness of the method.
keywords: neighborhood rough sets approximation condition entropy attribute reduction probabilistic neural network fault diagnosis
文章编号：20160215 中图分类号： 文献标志码：
基金项目:
引用文本：
谢玲玲,雷景生,徐菲菲.基于改进的邻域粗糙集与概率神经网络的水电机组振动故障诊断[J].上海电力大学学报,2016,32(2):181-187.
XIE Lingling,LEI Jingsheng,XU Feifei.Vibrant Fault Diagnosis for Hydro-turbine Generating Unit Based on Improved Neighborhood Rough Sets and PNN[J].Journal of Shanghai University of Electric Power,2016,32(2):181-187.
谢玲玲,雷景生,徐菲菲.基于改进的邻域粗糙集与概率神经网络的水电机组振动故障诊断[J].上海电力大学学报,2016,32(2):181-187.
XIE Lingling,LEI Jingsheng,XU Feifei.Vibrant Fault Diagnosis for Hydro-turbine Generating Unit Based on Improved Neighborhood Rough Sets and PNN[J].Journal of Shanghai University of Electric Power,2016,32(2):181-187.