摘要
为实现水电机组轴系运行常见故障的快速实时诊断,提出了一种基于支持向量机的故障诊断及预测方法。该方法应用支持向量机分类的基本原理,提取机组振动信号的频谱能量作为学习样本,通过训练建立基于水电机组轴系运行常见故障的分类模型,进行故障类型识别。同时,结合状态监测系统的实时采集数据,应用时间加权因子和支持向量机回归模型,实现特征数据的实时预测。经实验分析验证,该诊断方法具有较高的准确性,其回归预测方法有效可行,能满足实时故障诊断的要求。
In order to realize the real-time fault diagnosis for shaft system of hydropower unit,a fault diag-nosis and prediction method based on Support Vector Machine(SVM)has been put forward. Applying thebasic principle of classification,frequency energy from vibration signal data of shaft was collected,featurevectors build with energy were extracted as learning samples,and then used in training and establishingclassification models for fault type identification. In addition,real-time acquisition data from condition moni-toring system and time weighting factor were used for constructing regression models,which realized the re-al-time prediction of trend data. The experiment shows that the diagnosis and prediction method has a highaccuracy,which is suitable for online fault diagnosis of hydropower unit.
出处
《水利学报》
EI
CSCD
北大核心
2013年第S1期111-115,共5页
Journal of Hydraulic Engineering
关键词
水电机组
轴系
故障诊断
支持向量机
小波包分解
hydropower units
shaft system
fault diagnosis
least squares support vector machine
wavelet packet decomposition
