基于混沌理论与SVM的内燃机振动信号趋势预测  

Trend Prediction of Engine Vibration Signals Using Chaotic Theory and Support Vector Machine

作　　者：冯广斌 吴震宇 袁惠群[3] 

机构地区：[1]军械技术研究所,石家庄050000 [2]武汉船用机械有限责任公司,武汉430084 [3]东北大学机械工程与自动化学院,沈阳110004

出　　处：《振动．测试与诊断》2011年第1期64-69,129,共6页Journal of Vibration,Measurement & Diagnosis

基　　金：国家高技术研究发展计划("八六三"计划)资助项目(编号:2006AA04Z408)

摘　　要：针对内燃机振动信号信噪比低且呈非线性、非平稳的特性,提出将经验模态分解(empirical mode decomposition,简称EMD)相空间重构理论与支持向量机(support vector machine,简称SVM)相结合,实现内燃机振动监测数据的建模及预测。首先,将含噪声的振动信号经验模式分解,去掉主要干扰因素所对应的固有模态函数(intrinsicmode function,简称IMF)分量,再将剩余IMF分量进行重构,得到去噪声后振动信号时间序列;然后应用混沌理论,选择合适的嵌入维数和时间延迟对去噪后的振动信号时间序列进行相空间重构;最后采用SVM对其进行建模预测,并与径向基函数(radial basis function,简称RBF)神经网络的预测结果进行比较。试验数据表明,该方法能够预测内燃机振动信号的变化趋势,性能优于传统的分析方法,具有一定的工程实用性。For the low signal to noise ratio,non-linear and non-stationary characteristics of engine vibration signals,a modeling and forecasting technique is put forward by integrating the empirical mode decomposition(EMD) de-noising,the theory of phase space reconstruction and support vector machine(SVM).Firstly, the vibration signals were decomposed by using the EMD method and the intrinsic mode function(IMF) components related to main interference factors were eliminated.Then,the true vibration signal was obtained by reconstructing the remaining IMF components,and the time-series embedding space about the denoised vibration signals were rebuilt after the embedding dimension and time delay had been determined based on the theory of chaos.Finally,the modeling prediction was carried out by SVM and compared with radial-basis-function(RBF) neural network.The experimental results show that the proposed method can predict the trend of engine vibration signals and its performance surpasses the traditional techniques and has better project practicability.

关 键 词：内燃机 经验模态分解 相空间重构 支持向量机 趋势预测 

分 类 号：TK421[动力工程及工程热物理—动力机械及工程] TH165.3[机械工程—机械制造及自动化]