Research On Speed Tracking Predictive Control Of High Speed Trains

China has the largest high-speed railway network in the world,with nearly 30,000 kilometers of operation mileage far exceeding other countries in the world.The continuous construction and vigorous development of high-speed railway comprehensively improve the quality and efficiency of Chinese railway network travel,meanwhile promote the coordinated and balanced development of Chinese regional economy.Chinese self-developed high-speed railway operation control system has achieved a leading position in the world,which is constantly innovating and improving.As one of its key technologies,Automatic Train Operation(ATO)has been successfully applied in urban rail transit,and high-speed train ATO technology is also in full swing.An important responsibility of ATO technology is controlling the safe and efficient operation of train.Therefore,how to ensure the smooth and safe operation of high-speed trains is a research hotspot of ATO speed tracking control in recent years.Therefore,this thesis closely follows the development pace of ATO technology,and researches an intelligent speed tracking control algorithm with high tracking accuracy and strong robustness.Due to the complicated operation conditions of high-speed trains,the operation process has obvious nonlinearity,time-varying and external interference.So the model of high-speed train running process is established by using neural network and train speed tracking control is realized by using generalized predictive control in the thesis.The specific research contents are as follows:On the basis of studying the modeling methods of high-speed trains by domestic and foreign scholars,this thesis chooses the data-driven method to establish the high-speed train operation process model.According to the actual train input and output data,the radial basis function neural network model is trained offline by supervised and unsupervised learning method so that the non-linear model of train operation process can be established.At the same time,the established neural network model is adjusted online to ensure an effective model while the train is running.The design of the speed tracking control algorithm adopts a generalized predictive control algorithm with strong robustness and self-adaptive ability.Firstly,the generalized predictive control algorithm is analyzed in detail,then the trained radial basis neural network model is instantaneously linearized and used as the predictive model of generalized predictive control.Finally,owing to the complexity of the calculation process of the generalized predictive control,the generalized predictive control algorithm is improved to simplify the calculation and improve the control performance of the control algorithm.Considering the speed tracking accuracy,comfort and parking accuracy of high-speed trains,the multi-objective function with control input constraints is constructed.The controlalgorithm is designed that based on multi-objective function of generalized predictive speed tracking.By solving the quadratic optimization problem of the multi-objective function with constraints,which realize high-speed train speed distance tracking control to ensure the accuracy of train speed position tracking and passenger comfort.Finally,the CRH380 AL high-speed train is used as the simulation object,and the designed control algorithm is numerically simulated and verified.The results show that the designed control algorithm can track the given speed distance curve accurately and has faster response speed.At the same time,the passenger experience has also been greatly improved.It has certain reference value for the design of ATO speed tracking control algorithm.