Abstract: For nonlinear mapping relationship between vibration signal and state information in rolling bearing, a bearing feature vector extraction based on intrinsic mode function (IMF) feature energy in combination with support vector machine (SVM) is proposed for fault pattern recognition of bearing. The vibration signal of rolling bearing is decomposed into some IMF reflected bearing fault information by empirical mode decomposition(EMD), the energies including major information IMF are taken as eigenvectors. They are input into SVM classifier respectively for achieving fault recognition of rolling bearing. The normal state，outer race fault, inner race fault and rolling fault are simulated and tested. Results show that this method can recognize bearing fault effectively and accuracy.

Key words: vibration and wave, rolling bearing, empirical mode decomposition (EMD), feature energy, fault recognition, support vector machine (SVM)

摘要： 针对滚动轴承振动信号和状态信息非线性映射关系，提出一种基于内禀模态函数（IMF）特征能量的轴承特征向量提取方法，并与支持向量机（SVM)相结合实现轴承的故障识别。该方法对滚动轴承振动信号进行经验模态分解（EMD）得到若干能反映轴承故障信息的IMF分量，选取包含主要信息的IMF能量作为振动信号的特征向量，并将其输入到SVM分类器中实现轴承故障模式识别。对滚动轴承的正常状态、外圈故障、内圈故障和滚动体故障进行仿真试验，结果表明，该方法能够有效、准确地识别轴承故障。

关键词: 振动与波, 滚动轴承, 经验模态分解, 特征能量, 故障识别, 支持向量机