Research On Digital Twin Modeling And Anomaly Detection Method Of Resonance Bending Fatigue Testing Machine

Resonance bending fatigue testing machine is used to measure the limit life of pipeline under high stress state and reasonably predict the service life of pipeline,which is of great significance for the protection of marine environment and saving economic cost.However,at present,the resonance bending fatigue testing machine is vulnerable to its own constraints and external environment interference in the vibration process,resulting in unstable experimental measurement.Therefore,the resonance bending fatigue testing machine needs a complete prediction and monitoring system to ensure the safety in the vibration process.Based on the machine learning method,this paper establishes the digital twin model of the resonance bending fatigue testing machine,an anomaly detection method for testing machine is proposed,which can be applied to various complex occasions.The main contents of this paper are as follows:Firstly,the differential equation of the pipeline under external load is derived according to the transverse vibration model of the beam.Combined with the boundary conditions,the first-order modal shape function and frequency equation of the pipeline with additional mass are obtained,and the natural frequency and vibration fulcrum position of the pipeline are solved;According to the deflection curve equation of the pipeline under external load,the maximum stress and amplitude of the pipeline are solved;The mechanical model of air spring and roller bearing is established,and the stiffness of exciter and support mechanism is analyzed;The calculation method of stress amplitude and curvature is proposed,which lays a foundation for the subsequent establishment of ROM model and acquisition and control system.Secondly,the whole machine is modeled and simulated.Through the modal analysis and harmonic response analysis of the whole machine,the first-order modal frequency of the pipeline and the frequency stress displacement response curves under different load states are obtained;The reliability of the resonant bending fatigue testing machine is verified by analyzing the modal frequencies of key parts.Then the digital twin model of resonance bending fatigue testing machine is established based on SVM method;By sorting out the input and output parameters of resonance bending fatigue testing machine,different training data sets are obtained,and the SVM program training model is imported to obtain the ROM model,which is the digital twin model.Finally,the reliability of the model is verified by using the prediction set.Then,according to the engineering test requirements,the acquisition and anomaly detection system of resonance bending fatigue testing machine is established,including the connection of sensor and acquisition instrument,data processing and the development of data acquisition control interface,the digital twin model is used to predict the experimental state,set the allowable error to determine the abnormal state,and finally build the control system of the testing machine.Finally,through the vibration experiment of the pipeline under two working conditions,the stress time curve and error curve of abnormal state and normal state are collected and analyzed to verify the reliability of the model and the feasibility of anomaly monitoring.