摘要： 针对重庆轻轨铸钢支座系统故障诊断中缺乏故障样本的问题，提出一种基于经验模态分解(EMD)和支持向量数据描述(SVDD)的故障诊断方法。对采集到的振动脉冲响应信号进行EMD分解，提取第一、第二模态的能量和平均值作为特征输入到SVDD分类器进行训练和分类。实验结果表明，采用EMD分解后提取的特征能有效地浓缩故障信息，使SVDD分类器具有分类效果好、计算效率高等优点。

关键词: 支持向量数据描述, 铸钢支座系统, 经验模态分解, 故障诊断

Abstract: A method of malfunction diagnosis based on Empirical Mode Decomposition(EMD) and Support Vector Data Description(SVDD) is proposed to solve the problem of lacking malfunction smaples in Chongqing light-rail’s cast steel pedestal system diagnosis. The vibration impulse response signal is captured and decomposed by EMD, and the first mode’s and the second mode’s energy and average are extracted as features to input the SVDD classifier for training and classifying. Experimental result indicates that, extracting the features after EMD decomposition can concentrate the malfunction information effectively, which makes the SVDD classifier have well represention and high efficiency.

Key words: Support Vector Data Description(SVDD), cast steel pedestal system, empirical mode decomposition, malfunction diagnosis

中图分类号: 

• TP39