摘要
为有效进行柴油机故障诊断,利用柴油机振动信号,将振动信号处理和机器学习相结合,提出一种基于优化的变模态分解(VMD)和支持向量机(SVM)的柴油机故障诊断算法。首先,针对VMD中分解层数对分解结果的影响,给出自适应分解层数的优化VMD算法方案,利用该优化VMD算法处理柴油机振动信号并获取IMF矩阵;然后,提取IMF矩阵的奇异值、原始信号的有效值和峭度作为SVM输入的特征向量,并设计多分类SVM模型。最后,利用台架实测样本数据(包含正常、曲轴故障、连杆故障和惰轮故障)验证该故障诊断算法的有效性,选取80%的样本数据对分类模型进行训练得到分类模型,20%样本数据进行预测验证。结果表明,模型10折交叉验证值为0.0067,算法的故障诊断正确率为98%以上。
In order to effectively diagnose diesel engine faults,a diesel engine fault diagnosis algorithm based on optimized variable mode decomposition(VMD)and support vector machine(SVM)was proposed by using the vibration signal of diesel engine,combining vibration signal processing and machine learning.Firstly,in view of the influence of the number of decomposition layers in VMD on the decomposition results,an optimized VMD algorithm scheme for adaptive decomposition layers was given,and the optimized VMD algorithm was used to process the vibration signal of diesel engine and obtain the IMF matrix,then,extracted the singular values of the IMF matrix,the effective value and kurtosis of the original signal were used as the feature vector of the SVM input,and a multi-class SVM model was designed.Finally,the validity of the fault diagnosis algorithm was verified by using the measured sample data of the bench(including normal,crankshaft fault,connecting rod fault and idler fault),and 80%of the sample data were selected to train the classification model to obtain the classification model,and 20%of the sample data were used for prediction validation.The results show that the model's 10-fold cross-validation value is 0.0067,and the algorithm's fault diagnosis accuracy rate is over 98%.
作者
严孝强
张振京
宋业栋
薛雷
张衡
Yan Xiaoqiang;Zhang Zhenjing;Song Yedong;Xue Lei;Zhang Heng(Weichai Power Co.,Ltd.,Weifang,Shandong 261061,China)
出处
《机电工程技术》
2022年第10期279-283,共5页
Mechanical & Electrical Engineering Technology
