Misalignment Characteristic Analysis Based on Kernel Principal Component Analysis

摘要：

A new method based kernel principal component analysis (KPCA) is used to extract interesting misalignment features from a dynamical system. In this method, the projections (PCs) of the image of a test point with misalignment onto the nonlinear principal components in normal condition in featured space F are computed to represent the misalignment characteristics. It is shown in this work that the exploitation of the projections combination can improve the detection results. Even the varying trends of misalignment fault could be identified by use of this detection method. The method is illustrated on an experimental example of an auxiliary magnetic bearing rotor system.

关键词： Angular misalignment Fault diagnose Kernel PCA

作者: Li Huimin Ma Xiaojian Wang Yanbing Lawrence A. Bergman

作者单位: College of Mechanical Engineering Donghua University Shanghai, 201620, China Department of Aerospace Engineering University of Illinois at Urbana-Champaign Urbana,IL,61801,USA

会议类型: 国际会议

会议名称: The Fourth International Joint Conference on Computational Science and Optimization(第四届计算科学与优化国际大会 CSO 2011)

会议地点: 昆明、丽江

会议语种:英文

页码: 293-296

在线出版日期: 2011-04-15（万方平台首次上网日期，不代表论文的发表时间）