All-coefficient Adaptive Control Of High-speed Train Based On Characteristic Model

The development of high-speed railway is of great significance to China's economic construction and easing the traffic burden.However,the mechanism affecting the running speed of high-speed train is complex.It is very difficult to establish high-precision train dynamics model and design simple and reliable speed tracking controller.First of all,the aerodynamic resistance of the train in high-speed operation can not be ignored,and the aerodynamic resistance is proportional to the square of the speed,which becomes a difficult nonlinear term in the dynamic model of high-speed train.Secondly,when the components in the train system are aging or damaged,the characteristics of the high-speed train system itself will change,which will inevitably result in the change of system structure parameters.Thirdly,in the actual high-speed train operation,with the change of working conditions and environment,there are uncertain factors such as external interference and limited input and output.If these uncertain factors are not considered,there will be some unmodeled errors.Therefore,in view of the complex and variable characteristics of high-speed train dynamics system,this paper proposes the method of characteristic modeling to analyze the high-speed train dynamics,and establishes the equivalent low-order slow time-varying characteristic model,and then designs the all-coefficient adaptive controller based on the builtcharacteristic model and tracking control requirements to realize the speed tracking adaptive control.Finally,on this basis,the influence of system delay is further considered,and the error prediction model is analyzed and established.Combined with the error prediction,the all-coefficient adaptive control considering the delay is realized.The research content of this paper mainly consists of the following aspects:Firstly,the conventional mathematical model of the longitudinal dynamics of the train is analyzed based on the physical mechanism of the high-speed train,and the applicability of characteristic modeling is analyzed based on this dynamics basis.Using the measured data of high-speed train operation and the least squares identification method with forgetting factor to identify slow time-varying characteristic parameters,then the equivalent low-order slow time-varying characteristic model is obtained.The characteristic model avoids the actual physical mechanism of the train,and the modeling is based on the object's dynamic characteristics and environmental characteristics,which greatly reduces the complexity of the model,thereby facilitating the design of simple and reliable low-order controllers.Secondly,according to the traction / braking characteristics of the train,the ideal running speed curve of the train is drawn.Based on the built train dynamic characteristic model and all-coefficient adaptive control theory,the speed tracking control analysis and combination research are carried out for different controllers,and the tracking control performance isevaluated by the mean square error.Simulation results show that the combination of nonlinear golden section controller,maintenance / tracking controller and logic integral controller has the best tracking performance.Finally,considering the delay of the system,the error prediction model is designed according to all coefficient adaptive control theory,and the prediction performance of BPNN,SVM and LSSVM are compared.A sliding window is established according to the timeliness and correlation of samples to filter and weight training samples,thereby further improving the prediction accuracy of the prediction model.Combined with the prediction model,the all coefficient adaptive tracking control considering the system delay is realized.