基于EMD和VPMCD算法的燃油系统故障诊断方法

A Fault Diagnosis Approach of Fuel System Based on VPMCD and EMD Algorithm

提出了一种基于经验模式分解(empirical mode decomposition,简称EMD)算法和变量预测模型(Variable predictive model based class discriminate,简称VPMCD)的燃油系统故障诊断方法。利用EMD分解获得信号的本征模式函数(intrinsic mode function,简称IMF),选取前5个IFM的能量作为特征向量,提取信号的特征信息。利用VPMCD模式识别方法得到各故障特征值的预测模型,并利用预测模型对待诊断样本的故障类型进行了分类和识别。分析结果表明,该方法能够准确识别出燃油系统故障。

A fault diagnosis approach of fuel system based on empirical mode decomposition （EMD） and vari- able predictive model based class discrimination （VPMCD） algorithm is proposed. The EMD algorithm can decompose a non-stationary signal into the sum of many intrinsic mode functions （IMF）. In this paper, the energy of the front five IMFs which include the most fault information were extracted to serve as feature param- eters. Finally, variable predictive models of different feature parameters were established by VPMCD, which were used to classify and recognize different fault types of samples to be diagnosed. The diagnostic results demonstrate the effectiveness of the proposed method.