Adaptive Control For Braking Process Of High-Speed Train Based On Multilevel Switching

Braking system is a key part of high-speed train(HST)as well as a guarantee for safety operation of train.With the rapid development of C RH,higher requirements for braking system of HST is also put forward.C urrently,braking operation is mainly accomplished manually through braking handle,which is lack of high degree of automation.Therefore,how to build a precise braking model and how to achieve high-precision tracing control for target speed curve(a velocity-displacement curve that meets the demand of safety,comfortable and accurate of HST)are of great significance for automatic driving and parking automation.In consideration of the feature of the braking process of HST,this paper puts forward a multilevel switching method for braking process based on relevant braking principle,and realize the tracking control for target speed curve with model reference adaptive control strategy on the basis of RBF combined with the ac tual operation situation.Specific studies are as follows:1?According to the braking characteristics of CRH380 AL,the braking model of HST consists of static and dynamic characteristics,in which the static characteristics is a nonlinear relationship between middle variable of model(braking power)and input variable of model(braking level)on current speed,in order to facilitate the analysis process,this nonlinear relationship is disposed with multiple linear models in this paper,and the static linear functions is identified with least square method to acquire the part of static characteristics;moreover,dynamic characteristics is the effect of middle variable of model(braking power)on the model output(velocity)under the effect of the braking devices,which can be described with dynamic transfer function,the parameters of dynamic performance equation can be achieved on the basis of improved velocity variation particle swarm optimization algorithm.Finally,barking model that described with multiple linear model can be obtained through combined the two parts.2?Aiming at the nonlinear characteristics of the braking process,the traditional control method is relatively complicated and large computation is needed as design the algorithms.In this paper,a RBF based model reference adaptive control strategy is adop ted,which can avoid the complex mathematical analysis in traditional control methods,and the target optimizing curves to be tracked can be described with a stability reference model,which can greatly enhance the tracking accuracy in combination of learning and memory function of neural network.3?The proposed method in this paper is simulated and verified on the base of braking characteristics curve data and actual operation data of CRH380 AL,and simulation results show that this braking model consistent with the actual braking requirements,besides,model reference adaptive control based on RBF networks can realize high precision control for target optimization curve.