摘要
针对船用柴油机故障诊断时振动信号的非平稳特性和难以获取大量样本的实际情况,提出一种总体经验模态分解EEMD和最小二乘支持向量机LSSVM相结合的诊断方法。运用EEMD方法对特定时段的振动信号进行分析,计算各内禀模态函数IMF,并求其包含时间信息的能量熵,以之作为特征向量输入到LSSVM分类器来判断柴油机的故障类型。经实例验证,该方法能在保持信号完整性的前提下有效提取故障特征,在小样本情况下具有较高的诊断精度、较快的诊断速度和较强的泛化能力,能有效应用于同类型机械的故障诊断。
With reference to the unsteady characteristic of vibration signal and the difficulty of obtaining mass sample when making a fault diagnosis of ship diesel engine, a diagnostic method combining EEMD and LSSVM is put forward. EEMD is used to analyze vibration signals at special time, calculate IMF and energy entropy including time information that will be used as eigenvector of LSSVM classifier, so as to judge the type of faults. An experimental example is given to validate the method, and it shows that this method can abstract characteristics of fault effectively and keep the integrality of signals. The method can also be applied to diagnose similar mechanical faults effectively because of its higher precision, faster diagnostic speed and stronger flexibility.
出处
《柴油机》
2012年第6期10-13,23,共5页
Diesel Engine
关键词
柴油机
故障诊断
总体经验模态分解
最小二乘支持向量机
特征向量
diesel engine
fault diagnosis
ensemble empirical mode decomposition
least square supportvector machine
eigenvector