基于KSLPP与RWKNN的旋转机械故障诊断  

Rotating machinery fault diagnosis based on KSLPP and RWKNN

作　　者：王雪冬[1] 赵荣珍[1] 邓林峰[1] 

机构地区：[1]兰州理工大学机电工程学院,兰州730050

出　　处：《振动与冲击》2016年第8期219-223,共5页Journal of Vibration and Shock

基　　金：高等学校博士学科点专项科研基金(20136201110004)

摘　　要：针对旋转机械高维故障特征集识别精度低的问题,提出基于核监督局部保留投影(Kernel Supervised Locality Preserving Projection,KSLPP)与Relief F特征加权的K近邻(Relief F Weighted K-Nearest Neighbor,RWKNN)分类器相结合的维数约简故障诊断方法。该方法首先应用KSLPP提取故障特征集中的非线性信息,同时在降维投影过程中充分利用类别信息,使降维后最小化类内散度,最大化类间分离度;随后,将降维后得到的低维敏感特征集输入RWKNN进行模式识别,RWKNN能够突出不同特征对分类的贡献率,强化敏感特征,弱化不相关特征,提升了分类精度和鲁棒性。最后,通过典型转子实验台的故障特征集验证了该方法的有效性。Aiming at the questions of high dimension and low precision of the recognition of rotating machinery fault diagnosis, an intelligent fault diagnosis method based on kernel-supervised locality preserving projection and K-nearest neighbor weighted by feature selection ReliefF algorithm （RWKNN） was proposed. KSLPP can effectively extract nonlinear information in an original feature data set and at the same time make full use of class information in dimension-reduction projection, and make the sample minimize the within-class dispersion and maximize the separation between classes. Then, the sensitive low-dimension feature data set is fed into the K-nearest neighbor weighted by feature selection ReliefF algorithm to recognize the fault type. RWKNN can highlight the contribution rate of different features for classification, strengthen sensitive characteristics, weaken irrelevant features and improve the classification accuracy and robustness. Finally, the validity of the proposed method was verified by the typicalrotor fault vibration signal.

关 键 词：故障诊断 核监督局部保留投影 RELIEF F特征选择 加权K近邻分类器 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TH165[自动化与计算机技术—控制科学与工程]