水力发电学报2014，Vol.33Issue(3)：299-305,7.
基于多重分形谱和改进BP神经网络的水电机组振动故障诊断研究
Vibration fault diagnosis of hydropower unit based on multi-fractal spectrum and improved BP neural network
摘要
Abstract
A combined vibration fault diagnosis model of hydropower unit has been developed based on wavelet analysis,multi-fractal spectrum,and improved BP neural network in this study.To eliminate noise pollution in measured vibration signals,this model adopts a wavelet decomposition algorithm of signal purification.Multi-fractal spectrum parameters are used to characterize vibration signal and its features are identified with an improved BP neural network,so that different operating states of the unit rotating parts can be reflected and distinguished efficiently.The results show that the application of multi-fractal spectrum to fault diagnosis of hydropower unit is feasible and it could improve intellectualization and humanization of the diagnosis work and enhance human-computer interaction.关键词
水电机组/振动故障/多重分形谱/小波系数/改进BP神经网络Key words
hydropower unit/ vibration fault/ multi-fractal spectrum/ wavelet coefficient/ improved BP neural network引用本文复制引用
郭鹏程,孙龙刚,李辉,罗兴锜..基于多重分形谱和改进BP神经网络的水电机组振动故障诊断研究[J].水力发电学报,2014,33(3):299-305,7.基金项目
国家自然科学基金项目(51209172) （51209172）
教育部高等学校博士点基金(20126118130002) （20126118130002）
陕西省自然科学基础研究计划项目(2012JM7005) （2012JM7005）
