The Research On Fault Diagnosis Method Of Rolling Bearing Based On VMD And Manifold Learning

Rolling bearing is one of the indispensable parts in rotating machinery.It is widely used in many important fields.Once the rolling bearing is damaged,it may cause the whole equipment to malfunction,and even cause economic property loss and casualties.Therefore,it is of great practical significance to diagnose the rolling bearmg.The main steps of fault diagnosis are:vibration signal acquisition,fault feature extraction and patterm recognition.In this paper,the fault vibration signal of rolling bearing is taken as the research object,and a fault diagnosis research method based on VMD and manifold learning is proposed.The research focus of rolling bearing fault diagnosis is to extract sensitive fault features in nonlinear and non-stationary fault vibration signals.Therefore,in order to effectively extract the fault characteristics of rolling bearings,a new adaptive signal decomposition method—variational mode decomposition(VMD)is selected.Firstly,the basic principle of VMD algorithm is introduced in detail.For the problem that VMD algorithl parameter selection has great influence on fault diagnosis accuracy,the feature extraction method combining VMD and time series analysis is proposed to select the optimal parameter K value.Through experimental analysis of simulation signals and experimental data signals,and comparative analysis with empirical mode decomposition(EMD),the experimental results show that the method can effectively and accurately extract fault features.The fault sample data has the characteristics of high dimensional complexity,so the manifold learning algorithm is used to perform dimensional reduction and classification identification of fault data.Firstly,five manifold learning algorithms are introduced and their advantages and disadvantages are analyzed.The standard dataset and actual experimental data are analyzed and venfied,and compared with the traditional linear dimensionality reduction algorithm PCA.The experimental results show that the manifold learning algorithm can completely preserve the structural characteristics of high-dimensional data and achieve feature compression.It has strong clustering ability,can realize fault classification and identification,and improve fault diagnosis accuracy.Finally,the experimental data of rolling bearing fault simulation is taken as the research object,and the validity and feasibility of the fault diagnosis model based on VMD and manifold learning algorithm are verified.The fault diagnosis accuracy is calculated by SVM.The experimental results show that the fault recognition accuracy of the manifold learning algorithm is significantly higher than that of the traditional PCA algorithm,which can realize the fault classification and identification of rolling bearings.The LE algorithm has the highest diagnostic accuracy and the best effect.