Modeling, Optimal Design And Simulation Analysis Of Semi-active Suspension System For High-speed Railway Vehicle

With the development of high-speed railway technology at home and abroad, the running speed of passenger train is in constant improvement. The velocity of Speed-raising passenger train of existing lines exceeds200km/h, and of newly-built high-speed train300km/h. With the increasing of train speed, the interaction between wheel and track becomes more serious, which threatens the running stability, riding quality and safety of vehicle. It is an important research subject how to effectively inhibit the vibration of car body, reduce the amplitude of hunting, improve the ride comfort and increase safety allowance under high-speed running environment.The purpose of this research is to reduce the lateral vibration of car body, to increase nonlinear critical velocity and improve safety allowance. From the perspective of building high-speed railway vehicle model, optimizing suspension system, building MR damper model and semi-active control, this paper analyzes the effects of semi-active control system with MR damper on the dynamic performance of vehicles under high-speed running conditions. The main work is as follows:(1) In order to find the optimal suspension matching and guarantee the running stability, riding quality and curve negotiation ability of railway vehicles, we built a parametric model of suspension system for high-speed railway vehicle, using ADAMS-Matlab method. We optimized the suspension system by adopting multi-objective genetic algorithm, and found optimal suspension parameters from multi-group Pareto solutions by utilizing average value screening method and tolerance design method. The optimized suspension system could ensure good running stability of the vehicle and significantly improve the riding quality and curve negotiation ability of model.(2) On the basis of mechanical test data of MR damper, we successfully identified the parameters of Bouc-Wen model by using genetic algorithm and pattern search method, and made generalized inspections on the identified results. Since it was difficult for mechanical models to simulate the reverse characteristics of MR damper, the nonlinear mapping relationships between running state of MR damper and corresponding currents strength were simulated by utilizing BP neural network technology and the weight and threshold value were optimized by using genetic algorithm. At last, the forward and inverse MR damper model were successfully applied to semi-active suspension system of quarter car and satisfactory control was obtained.(3) To avoid high frequency flutter caused by on/off semi-active control, the continuous semi-active control strategies of sky-hook(SH) damping control, acceleration drive damping(ADD) control, Rakheja-Sankar control and SH-ADD mix control were built respectively, and the simulations of which under high-speed running state were conducted by using quarter and whole vehicle model respectively. Results showed that the effect of sky-hook control was good in low frequency area and the effect of acceleration drive damping control was better in middle and high frequency area, while SH-ADD mixed control have both the advantage of sky-hook damping control and acceleration drive damping control. Compared with ordinary semi-active control, continuous control could significantly reduce the jumps when damper converts states, so the switch between different states was smoother and the high frequency flutter was reduced effectively.(4) A MR damper-based semi-active suspension system of high-speed railway vehicle was built, which is closer to the actual conditions. The effects of MR damper and semi-active control strategies on the running stability, riding quality and curve negotiation ability of vehicle system were analyzed. The simulation results showed that effects of continuous SH-ADD mixed control were the best.(5) The effects of time delay in different signal channel on dynamic performance were simulated. It can be seen that effects of time delay on the dynamic performance was bigger only in velocity channel, especially in frame velocity channel, while in others signal channels the effects were very small. With the increasing of time delay, the dynamic performance presented a regular wavelike change rather than a monotonous change. The effect of time delay on the dynamic performance under continuous semi-active control was far less than the others control strategies.