Research On Design And Application Of Model Predictive Control For Automatic Train Operation System

In automatic train operation(ATO) system, ATO controller drives the train automatically instead of a driver. By controlling the moving of the train as close as possible to the target curve, the demands for time accuracy, safety, travelling comfort and energy saving can be met. Because on one hand external disturbance(slope, air resistance and line humidity) and the change of internal parameters(caused by load or temperature changes) will affect the running of the train, on the other hand there are constraints and time delay in its driving and braking system, poor parking accuracy and low travelling comfort level always happen. Model predictive control(MPC) is a kind of control algorithm optimizing online based on prediction model, rolling optimization and error correction. Compared with traditional control algorithm, it has advantages in solving time delay, constraint and parameter change problems. Under such a background, this dissertation investigates design and application of model predictive control for automatic train operation system, builds train motion model based on measured data, designs whole trajectory tracking MPC controller and illustrates its effectiveness by simulation using real station information.The main work of this dissertation can be summarized as follows:1) Building train motion model. Analyze the working princ iple of ATO system and build train motion model. Program a software to validate model and verify its accuracy with measured data.2) Design and simulation of speed tracking MPC controller. We design speed tracking MPC controller to meet the demands for time accuracy and travelling comfort. Simplify train motion model,the simplified model is made up of driving model and braking ones. Design driving and braking MPC controller respectively, propose switching strategy and show the flow chart of speed tracking MPC controller. The simulation results illustrate its effectiveness in measurable and immeasurable disturbance rejection. Analyze its characteristics comparing with existing control algorithm.3) Design and simulation of whole trajectory tracking MPC controller. We design whole trajectory tracking MPC controller to meet the demands for safety and parking accuracy. Discuss the requirements for controller in train starting, running and parking process and provide corresponding solutions. The simulation results based on real station data illustrate the effectiveness of the proposed control scheme.4) Study on the change of control performance with the change of train internal parameters. Simulate the cases when the train’s steady-state gain, time delay and time constant are different from predictive model and draw a conclusion.