Research On Fast Calculation Method In Hybrid Test And Numerical Simulation Of High-Speed Railway System

In a real-time hybrid test of a high-speed railway system,a train can be used as a test substructure to be loaded with a shaking table,and a bridge can be calculated numerically as a numerical substructure.The control of train test substructures requires accurate calculations,so sufficiently refined vehicle and bridge models need to be used.This results in a model that is very large and computationally extremely time consuming.On the other hand,the real-time hybrid test requires realtime performance.Each step of the numerical calculation,data exchange,time delay compensation,and actuator loading needs to be completed in a very short time(such as a few milliseconds).High requirements are imposed on the speed of numerical calculations.This paper presents an fast calculation method for real-time hybrid test of a high-speed railway car-rail-bridge coupling system,named based on the moving load integral metho.Store data required for bridge numerical substructure calculations before performing real-time hybrid tests,in the actual calculation,only the data corresponding to each time step needs to be extracted and calculated,and the displacement response of the bridge numerical substructure at the current time step can be obtained.This method has high calculation efficiency while ensuring the calculation accuracy and can well meet the requirements for the calculation accuracy and speed of the numerical substructure part of the real-time hybrid test of the high-speed railway vehicle-rail-bridge coupling system.Accurate simulation of wheel-rail rolling contact behavior is of great significance for fully studying the contact behavior of the wheel-rail coupling system under various operating conditions and analyzing the factors affecting the wheel-rail contact force.In this paper,the wheel-rail contact force solution method based on Non-Hertz contact theory is programmed in MATLAB and added to the existing OpenSees 3D wheel-rail coupling element,so that the dynamic response of the space wheel-rail coupling system can be calculated and analyzed based on the non-Hertz contact theory in the OpenSees platform.With reference to the existing literature,a space wheelset-rail model was established in OpenSees,and the accuracy of the method in this paper was verified by comparing the calculation results using the method of this paper with those of existing literature.This paper selects two different wheel-rail tread combinations,and uses Hertz contact theory and non-Hertz contact theory to calculate wheel-rail contact spots and normal forces under different conditions,and he calculated results are compared and analyzed.With reference to the existing literature,a three dimensional train-track space coupling model was established in OpenSees,and the vehicle system’s dynamic response was calculated based on the method in this paper.The stability and safety of the train were analyzed,the critical speed of the train and the derailment coefficient at different speeds were studied.The geometrical and mechanical problems of wheel-rail rolling contact are often more complicated,so the exact solution of the wheel-rail contact problem requires a large amount of calculation and the calculation is time-consuming.This paper proposes a fast calculation method for normal force of wheel-rail contact based on Back Propagation Neural Network.A BP neural network model is established to learn the nonlinear relationship between the normal force of wheel-rail contact and wheel-rail displacement.The trained neural network model is added to the existing OpenSees 3D wheel-rail coupling element.Use the trained BP neural network model to calculate the wheel-rail contact normal force at a given wheel-rail displacement,thus replacing the original wheel-rail contact normal force calculation function in OpenSees 3D wheel-rail coupling element.By comparing the calculation accuracy and calculation speed with the traditional method,the reliability and calculation efficiency of the method in this paper are verified.