A Multi-model Generalized Predictive Control Method For High-speed Trains Based On Event Triggering Mechanism

The running process of high-speed train has strong nonlinear characteristics and contains a lot of random modal information,which will have a great influence on the modeling of highspeed train running process.Predictive control is widely used because of its compatibility to system modeling and strong stability.However,when the high-speed train runs in steady-state condition,the traditional predictive control strategy still needs periodic frequent sampling,which will lead to a large number of redundant calculations of the controller to delay the response speed of the controller and thus lead to the occurrence of accidents.In order to overcome the above defects of traditional predictive control and improve the computational power of the controller;Based on the event trigger mechanism,this paper combines the multi-model idea with the event trigger generalized predictive controller.The experimental results show that the controller can improve the response speed of the controller on the premise of ensuring the control accuracy.The specific research is as follows:1.The dynamic characteristics of high-speed trains are analyzed in detail and the mechanism model is established to provide theoretical support for data-driven modeling and event-triggered generalized predictive controller control.2.In order to solve the problem that generalized predictive control in high-speed train operation will increase the calculation amount,this paper proposes a generalized predictive control method for high-speed train based on event triggering mechanism.In this method,the stability of the system is derived according to the state error,and then the event triggering condition is determined according to the upper limit of the error.With the help of event triggering mechanism,the system will only carry out identification and rolling optimization when the triggering conditions are not met,so as to reduce redundant optimization operations and improve the dynamic response speed of the system.The simulation results based on real data of CRH380 B high-speed train verify the correctness of the theoretical analysis.3.In this paper,the optimal number of submodels is determined by subtraction clustering algorithm based on the real-time data of train operation.The least square method with forgetting factor was used to linearize the nonlinear model near the working equilibrium point of the submodel.Then,according to the multi-model idea,the model switching strategy with minimum error is set up.Finally,the multi-model idea is combined with the event-triggered generalized prediction controller,and the multi-model generalized prediction controller based on the event-triggered mechanism is designed.The simulation results based on real data of CRH380 B high-speed train verify the correctness of the theoretical analysis.