Acoustic Emission Signal Prediction Of Axle Fatigue Cracks Based On BiLSTM-CRF

During the running process of the rolling stock,the contact between the wheelset and the rail produces rigid vibration,and the axle is continuously interacted with rotation and impact,and various forces cause fatigue cracks on the axle.In order to prevent the axle cracks from developing to the fracture stage and causing serious accidents,monitoring and early warning is carried out on the running status of the axles,the signals of fatigue cracks are predicted in time,the development trend of the axle fatigue cracks is grasped,and the axles are warned in advance before the dangerous situation occurs,to maintain and repair the axle in time,to avoid the axle failure,to ensure that the train completes the normal production task,it is very necessary to predict the acoustic emission signal of the fatigue crack of the axle.In order to prevent the axle crack from developing to the fracture stage and causing serious accidents and to predict the propagation waveform of the axle fatigue crack in advance,this paper proposes a prediction method of the axle fatigue crack acoustic emission signal based on Bi LSTM-CRF.First,the axle fatigue crack acoustic emission signal is divided into a training set and a test set.The training set is input into the Bi LSTM-CRF network model for training,and the parameters of the network model are determined.Then,using the test set to verify the network model.The results show that the signal predicted by the network has a smaller error compared with the actual signal.Secondly,compared with other prediction network models,the Bi LSTM-CRF network model has better prediction effect on the acoustic emission signal of axle fatigue cracks and has higher stability.Finally,another 7 million train axle fatigue crack acoustic emission signal is selected as the verification experiment,which proves that the network model has good generalization ability and applicability.The method proposed in this paper has certain guiding significance for the prediction of the acoustic emission signal of the axle fatigue crack,and lays a good foundation for the future signal prediction research.In order to realize the intelligent diagnosis of the fault information of the axle,the prediction system of the acoustic emission signal of the axle fatigue crack is initially developed in this paper.The system is designed in the Python language,and the various functional modules of the prediction system are programmed,so that the system has the function of data acquisition,display function,and the function of predicting the acoustic emission signal of axle fatigue crack using the Bi LSTM-CRF neural network algorithm.Finally,the various functions of the system are verified by lead breaking experiments.Through the cooperation of various modules,the acquisition and prediction of acoustic emission signals are initially realized.