摘要
柴油机发电机组结构复杂,故障类型多样,其动力传动部件兼具往复机械与旋转机械的振动特性。传统的频谱分析主要通过利用傅里叶变换将在时域内难于分辩的信号映射到频域内进行分析,这对于具有平稳特点的原始信号比较有效,但是对于柴油发电机组而言,频谱分析难以提取其频率分量,因此难以实现故障诊断。通过总体平均经验模式分解(EEMD)的方法获得其本征模式函数的近似熵,将该近似熵作为特征向量结合支持向量机(SVM)进行分类,从而实现柴油发电机组的故障识别。通过实验仿真和某柴油发电机组振动异常问题的实测试验表明,该方法可以准确有效的提取其故障信息和频率,为柴油发电机组传动机构故障诊断提供支持。
The structure of the diesel generator set is complex and fault type is various, their power drive components both have the vibration chavacteristics of reciprocating machinery and rotating machinery. Traditional frequency analysis mainly mapped the signal in time domain to the frequeney domain by the use of fourier transform which is difficult to distinguis/~ This is effective for the original signal which is stably, but for spectrum analysis, this is hard to extract the frequeney components from the diesel generating set, so it is difficult to realize fault diagnosis. The vibration signal of fault is decomposed into some IMFs by ensemble empirical mode decomposition, and then get the approximate entropies. Takes approximate entropies as eigenvectors to input support vector machine （SVM） for classification, then it realizes the fault identification of diesel generator set. The test of simulation experiment and vibration problem of a diesel generator set shows that this method can extract the fault information and frequency effectively, and provides support for fault diagnosis of the transmission mechanism of a diesel generator set.
出处
《机械设计与制造》
北大核心
2015年第3期24-27,31,共5页
Machinery Design & Manufacture
基金
国家自然科学基金资助项目(51175051)
关键词
振动与波
故障诊断
总体平均经验模式分解
支持向量机
传动部件
Vibration and Wave
Fault Diagnosis
Ensemble Empirical Mode Decomposition
Support Vector Machine
Drive Parts
作者简介
别锋锋，（1979-），男，湖北仙桃人，博士，副教授，主要研究方向：机械状态检测与故障诊断
