Research On Fault Diagnosis Of Rolling Bearing Of Subway Train Based On VMD

Rolling bearings are the key rotating parts of the subway train bogie,and are closely related to the safe operation of the subway train.Once the rolling bearing of the bogie fails,it will lead to huge economic losses.At present,the regular manual inspection can no longer meet the fault diagnosis needs of the subway train bogie bearings.At the same time,the working environment of the subway train bogie is bad,and the bogie axle box bearing is under the working condition of high speed and heavy load for a long time,and its vibration signal is seriously disturbed by noise.It is difficult to use traditional methods such as EMD to realize the fault diagnosis of the rolling bearing of the subway train bogie.Compared with the normal bearing vibration signals,the vibration signals of the faulty bearing of the subway train are rarer,and the available fault samples are not abundant.Therefore,taking the rolling bearing of the subway train bogie as the object,a fault diagnosis method of rolling bearing based on the improved particle swarm optimization algorithm optimization variational modal decomposition and the improved gray wolf algorithm optimization support vector machine is proposed.The main contents are as follows:Aiming at the problems of modal aliasing and poor anti-noise performance when the traditional empirical mode decomposition(EMD)is used to process the vibration acceleration signal of rolling bearings,which makes it difficult to extract fault features,a variational mode decomposition(VMD)based on parameter optimization is proposed.The rolling bearing signal processing method.Combined with improved particle swarm optimization(IPSO)and improved envelope entropy,the number of VMD decomposition layers and the combination of penalty parameters are optimized,and the components with high correlation coefficient with the original signal are screened out,and the fault features are extracted by combining with the sample entropy.The IPSO-VMD method is used to decompose the simulated bearing fault signal and the actual rolling bearing fault signal with strong noise,which proves that the IPSO-VMD method can effectively decompose the vibration signal of the rolling bearing with strong noise pollution and extract the fault features.Aiming at the low classification accuracy of the support vector machine with fixed parameters,and the lack of population diversity in the gray wolf algorithm,the initial population has a great influence on the optimization effect,etc.,an improved gray wolf algorithm based on multi-strategy improvement(IGWO)was proposed and tested.The advantages of the improved algorithm compared with the original algorithm and other algorithms are verified functionally,and the improved gray wolf algorithm is combined with the support vector machine to obtain a rolling bearing fault diagnosis model based on the support vector machine(IGWO-SVM)optimized by the improved gray wolf algorithm.Improve the correct rate of fault diagnosis.Under laboratory conditions,combined with the actual situation of subway trains,a rolling bearing fault diagnosis test bench was built,and the vibration acceleration signals of rolling bearings in different states were collected,and the above method was used for fault diagnosis.The highest fault diagnosis accuracy rate was 100%.After many experiments,the average fault diagnosis The accuracy rate reaches 98.875%,which can meet the fault diagnosis needs of rolling bearings of subway train bogies.