核主元分析在透平机械状态监测中的应用

讨论了核主元分析(K erne l P rinc ipa l Com ponen t A na lys is,简称KPCA)原理,提出了基于KPCA的透平机械状态监测方法。该方法在低维特征空间利用内积核函数,实现原始空间到高维空间的非线性映射以及对高维映像数据的主元分析,从而在低维空间得到原始特征的非线性主元,并根据非线性主元构建特征子空间,实现特征提取和对透平机械状态的分类识别并监测其状态变化。对仿真数据及透平机械在正常、重负荷状态下试验数据的研究表明,KPCA分类效果比主元分析好,能有效地识别出透平机械的不同状态,并能及时监测到状态发生的变化。

Application of Kernel Principal Component to Turbines Condition Monitoring

The principle of kernel principal component analysis(KPCA) is discussed,and an approach for turbines condition monitoring based on KPCA is presented,in which the integral operator of kernel function is used in low dimensional feature space to realize the nonlinear mapping from the raw feature space to high dimensional space,and perform principal component analysis(PCA) on the high dimensional image feature sets.Then the nonlinear principal components of raw feature space are obtained.In succession,these selected nonlinear principal components are used to construct the feature subspace for turbines condition monitoring.The analysis results on simulation data and experimental data of turbines working under normal and heavy load conditions indicate that,KPCA performs better in pattern classification,recognizes turbines conditions efficiently and monitors the developments of turbines conditions in time.