基于支持向量机的机械系统多故障分类方法

Application of Support Vector Machine in Multi-fault Classification of Mechanical System

提出了一种利用支持向量机 (SVM)对机械系统故障进行分类的新方法 ;以二值分类为基础 ,开发了基于支持向量机的多值分类器。并以齿轮的多种故障分类为例 ,进行了实际应用验证。结果表明 ,该方法具有很好的分类能力和较高的计算效率 ,不需要对原始数据进行预处理就可达到满意的效果 ,可以满足在线诊断的要求 ,适合于机械故障诊断中的多故障分类。该方法的应用 ,为故障诊断技术向智能化方向发展提供了新的途径。

A new method of fault classification for mechanical system by means of support vector machine (SVM) is proposed and a multi-class SVM classifier based on binary classification was developed. Experiments of fault gears classification by their vibration signals was conducted. The results from the experiments prove that the SVM method has a good classification ability and high efficiency for multi-fault classification in mechanical systems, even for the cases without preprocessing of the original signal.