Research On Diagnosis Technology Of Rotor Rubbing Fault Degree

The rotor system,as the core component of rotating machinery such as steam turbines,aero engines,water turbines,fans and centrifugal compressors,plays a very important role in the high-speed rotation of the equipment.As the requirements for the speed and efficiency of mechanical equipment become higher and higher,the gap between the rotor and the stator also becomes smaller and smaller,so that the probability of rubbing faults continues to increase.Rotor system rubbing faults may affect the life of the equipment in a small way,and cause serious safety accidents in a serious way.In order to avoid various hidden dangers caused by rubbing faults,this paper takes different degrees of rubbing fault vibration signals as research objects,and proposes a rotor rubbing fault based on the combination of optimized variational modal decomposition and support vector machine model.Degree of diagnosis.Rotor rubbing degree diagnosis research focuses on extracting sensitive fault features from non-linear and non-stationary fault vibration signals.Therefore,a new adaptive signal decomposition method,Variational Modal Decomposition(VMD),is used to perform fault signal analysis.Feature extraction.First,the basic principles and steps of the VMD algorithm are introduced in detail.For the two important parameters of the VMD algorithm modal number K and the penalty factor,which have a large impact on the signal decomposition and need to be set manually,a particle swarm optimization algorithm(PSO)The optimization is performed and the envelope entropy is used as the fitness function to determine the optimal parameters of the VMD.The VMD method is used to process and analyze the experimental data,and compared with the empirical mode decomposition(EEMD),the results show that the VMD method can decompose signals more effectively and accurately.The sample data of the rubbing fault belongs to a small sample,so a support vector machine(SVM)is used as a classification model for pattern recognition.For small sample data,the SVM classification model with the radial basis function as the kernel function has higher accuracy,but the selection of the penalty factor C and the parameters of the kernel function in the kernel function is optimized by the PSO optimization algorithm.Determine the optimal parameters to improve the accuracy of fault diagnosis and identification.Finally,three sets of rotor rubbing experiments under different conditions are designed.One is to produce different rubbing under the condition of the oil film force;the other is to change the different eccentric mass to produce different rubbing;the third is there are different levels of rubbing under changing conditions.By analyzing the vibration signals in three states,the validity and feasibility of the fault diagnosis recognition model based on PSO-optimized VMD and SVM parameters are verified,and the diagnosis accuracy is high.