基于局部均值分解样本熵与支持向量机的2000型压裂车液力端故障诊断

Hydraulic failure diagnosis of type 2000 fracturing vehicle based on LMD sample entropy and SVM

压裂车液力端工作时一般需要承受很高的载荷、介质腐蚀及冲蚀作用。针对这特殊的工作环境,提出了一种基于局部均值分解(LMD)样本熵与支持向量机(SVM)的2000型压裂车液力端故障诊断方法。LMD能够将采集到的信号自适应的分解成若干个乘积函数(PF)分量,通过PF与原信号的相关性系数,选取其中有代表故障状态主要信息的PF分量;然后用得到的主要PF分量进行样本熵的计算,得到特征值;最后,使用常规SVM分类器进行分类,诊断是否故障以及故障类型。实验证明,基于LMD样本熵与SVM的方法对2000型压裂车液力端故障诊断有效。

When the hydraulic of a fracturing vehicle is working,it generally needs to bear high loads,media corrosion and erosion.For the work environment of the hydraulic of type 2000 fracturing truck is complex,this paper proposed an effective fault diagnosis method based on LMD(Local Mean Decomposition)sample entropy and SVM(Support Vector Machine).LMD can be adaptive to decompose the collected signals into several PF(Product Function)components.By calculating the relative coefficients between the PF components and the original signal,the PF components containing the main information were selected.Then the sample entropy of these selected PF components was calculated in order to get feature vectors.Finally,conventional SVM was used to sort out the fault types and judge whether the equipment failed and the fault type.The experimental results prove that the fault diagnosis method about the hydraulic of type 2000 fracturing truck based on LMD sample entropy and SVM is effective.