基于数学形态学和模糊聚类的旋转机械故障诊断

提出了一种数学形态学与GG (Gath-Geva)模糊聚类相结合的旋转机械故障诊断方法,通过对滚动轴承信号的多尺度形态运算得到信号的形态谱,定量反映了信号在不同尺度下的形态变化特征.为进一步对滚动轴承信号进行故障识别,提取出基于形态学操作的分形维数和描述不同信号形态特征的指标即形态谱墒,并把这2个参数作为GG聚类的故障特征向量,进行聚类分析,同时对GG聚类与FCM(fuzzy center means)聚类和GK (Gustafaon-Kessel)聚类进行了比较.实验证明了基于数学形态学与GG聚类相结合的机械故障诊断方法的有效性,且证明了GG聚类更适合对不同形状、大小和密度的空间故障数据模糊聚类,聚类效果更好.

Rotating machinery fault diagnosis based on mathematical morphology and fuzzy clustering

A new method for rotating machinery fault diagnosis based on mathematical morphology and Gath-Geva(GG) clustering algorithm is introduced.The mathematical morphological spectrum curves are created using multi-scale morphological opening algorithm with varying flat structure elements,which could show different fault characteristics quantitatively.In order to recognize fault pattern of rolling bearings further,fractal dimension based on morphological operation and morphology spectrum entropy describing morphological characteristics of different signals are extracted;and the two parameters are used as the fault feature vectors of GG clustering algorithm.And the GG clustering is compared with fuzzy center means(FCM) clustering and Gustafaon-Kessel(GK) clustering.Experiment result proves that the machinery fault diagnosis algorithm based on mathematical morphology and Gath-Geva clustering is effective,and GG clustering algorithm is more suitable for the datasets with different shapes,sizes and densities,and has better clustering effect.