论文部分内容阅读
提出了基于时间序列参数模型和核Fisher判别分析(KFDA)的滚动轴承故障诊断方法.该方法首先通过自相关算法对轴承振动信号建立自回归(AR)模型,将自回归模型的参数作为特征向量并映射到高维核空间.然后在高维核空间中进行Fisher判别分析,求出Fisher判别分析的最优投影向量以及各类状态的Fisher判别值.最后获取未知状态轴承的高维核空间特征向量,求出其在最优投影向量上的投影值,通过与判别值进行距离判别来识别轴承所处的状态.实验结果验证了所用方法的有效性.
A rolling bearing fault diagnosis method based on time series parameter model and nuclear Fisher discriminant analysis (KFDA) is proposed.The autoregressive (AR) model of bearing vibration signals is established by using the autocorrelation algorithm, and the parameters of the autoregressive model are used as eigenvectors And then the Fisher discriminant analysis is carried out in high dimensional kernel space to find the optimal projection vector of Fisher discriminant analysis and Fisher discriminant values of various states.Finally, the high dimensional kernel space eigenvectors , Find its projection value on the optimal projection vector, identify the state of the bearing through the distance discrimination with the discriminant value.The experimental results verify the effectiveness of the method.