Intelligent Monitoring Technology For High Speed Train Automatic Production Line-Research And Realization Of Rolling Bearing Fault Diagnosis Method

The rapid progress of Chinese high-speed railway transportation has attracted worldwide attention.It has become a pillar industry in boosting the economy of our country.The high-speed train automatic production line plays an important role in ensuring the production quantity and quality of high-speed trains.Operational failure of the key equipment in the production line will directly affect the efficiency of the production line and even make it shut down.Therefore,the company hopes that the automatic production line can run under trouble-free conditions.Even if there is a fault,they can quickly identify the reason.How to ensure a healthy and stable operation of the production line and reduce the frequency of failures of the production line has become an important issue for the automatic production line of high-speed trains.Rolling bearing is a key component of the production line.Its operating status often affects the entire production line.There are many reasons for bearing failure,and the traditional equipment maintenance methods are difficult to meet the requirements of stable and healthy operation of the equipment.This paper has studied the achievements made in the field of fault diagnosis both at home and abroad.After analyzing the advantages and disadvantages of each diagnosis technique,this paper proposes a new fault diagnosis model based on improved genetic neural network.Fault diagnosis of rolling bearings needs to address two key issues.First,how to select the appropriate characteristics to reflect the health status of rolling bearing;Second,how to establish a suitable model for effective fault diagnosis.Aiming at these two key issues,the following study is carried out in this paper.(1)To extract effective features of the vibration signal.There is a connection between the bearing fault status and the statistical feature of the vibration signal.The vibration signal of bearing with flake in the outer ring,bearing with flake in the inner ring,bearing with flake in the rolling body and the vibration signal in the normal state is analyzed.This paper extracts 14 features that can characterize its running status from the two aspects of time domain and frequency domain.The principal component analysis is used to reduce the feature set due to the correlation and redundancy between these features.(2)It is a feasible method to carry on the fault diagnosis by using genetic neural network to classify the fault state of bearing effectively.The learning mechanism of traditional genetic neural network model is flawed,making it difficult to obtain the global optimal solution or better suboptimal solution.Therefore,this paper improves the crossover operation and mutation operation to explore the space of gene combination more extensive,and to improve the fitness function to avoid the inhibition of the optimal part of individuals.These methods can increase the potential of optimization of the population.They can also get better solution in solution space,reduce the network output error and improve classification accuracy.(3)A fault diagnosis model based on improved genetic neural network is established.It shows that the algorithm model proposed in this paper can effectively improve the accuracy of fault diagnosis,and has strong practical value After verification of experimental data set.