关键词: 局部特征尺度分解(LCD), 核最近邻凸包(KNNCH)分类算法, 能量, 齿轮, 故障诊断

Abstract:

A gear fault diagnosis approach was proposed based on LCD and KNNCH classification algorithm. Firstly, LCD method was applied to an original gear bearing vibration signals and adaptively decomposed the signals into a series of intrinsic scale components (ISC). Secondly, the energies of the ISCs which contain main information were extracted and regarded as the fault feature vector. Finally, KNNCH classifier accepted the fault feature vector as the inputs, and then the working condition of gear could be identified by the outputs of the classifier. The experimental analysis results show that the proposed approach can effectively extract the fault feature information and accurately classify the working conditions of gear even in the case of small samples. Additionally, the comparative analysis results demonstrate that KNNCH classification algorithm can gain a considerable or better classification rate compared to SVM(support vector machine) algorithm.

Key words: local characteristic-scale decomposition(LCD), kernel nearest neighbor convex hull(KNNCH) classification algorithm, energy, gear, fault diagnosis