Adaptive Speed Tracking Control For High-speed EMU

In recent years, with the rapid development of Chinese high-speed railway, its mileage will reach nearly 20000 kilometers and it has 2200 operating quantity of Electric Multiple Units(EMU) at the end of 2015. At present, Chinese high-speed railway engineering construction level is in the forefront of the world, however, the research on the dynamic model and the optimization control of the high-speed train running process is insufficient relatively, which limits the sustainable development of China’s high-speed railway. The dynamic modeling and optimal control of high-speed train running process, to achieve the high precision tracking for a given operating speed, are the core technologies of EMU in the future high speed. Thus, its fundamental research is of great significance.This thesis establishes the EMU elemental point, distributed model and multi-intelligent body model, according to the nonlinearity, parameter uncertainties in the dynamics of high-speed EMU operation process and the coupling relationship between each two vehicles. Based on these models, aiming to realize the high accuracy tracking, the adaptive control and the improved adaptive multivariable generalized predictive control are respectively studied. Finally, a distributed predictive control method of EMU based on multi-agent is designed for the structure of multi power unit. The main contents and contributions are as follows:Firstly, aiming at the model parameters time-varying of high-speed EMU elemental model due to drag coefficient uncertainties, the adaptive generalized predictive controller is designed. For the time-varying model parameters of high-speed train, based on the dynamics mechanism model of high-speed EMU, according to the actual operation process data and traction/braking characteristic curve of EMU, the corresponding process model of operation is established by using recursive least squares method to identify the model parameters. For the given speed tracking control, the traction/braking force of EMU is real-time generated by using the rolling optimization and feedback correction control strategy.Secondly, since the operation dynamic performance of EMU is affected by the non modeling dynamic from nonlinear air resistance in high-speed EMU running process, the system structure is described by a linear adaptive model and a high order nonlinear term, that is the system structure of non-modeled dynamic integration. A nonlinear adaptive controller with high order nonlinear compensator is designed to realize tracking control for a given speed by using the Adaptive Networkbased Fuzzy Inference System(ANFIS) and the one-to-one mapping to estimate the high order nonlinear term, and the stability and convergence of the proposed adaptive control algorithm are proved.Again, different from the adaptive generalized predictive control method in the control process only modifying the model parameters and not modifying the weighted coefficient, which leads to the problem of the train control effect is poor in the starting, braking, or operating conditions of the conversion stage, this thesis develops an improved adaptive multivariable generalized predictive control method. According to the actual power allocation structure, the operation process of EMU is described as a multi input and multi output system, and the distributed model is established. The improved adaptive multivariable generalized predictive control method is adopted to track the given speed, which is not only updates the model parameters but also adjusts the controller parameters according to the new model parameters, and can improve the control performance of the EMU.Next, according to the dynamic structure characteristics of high-speed EMU, each power unit is described as an agent, based on the operating data and traction/ braking characteristic curve of each power unit, the multi-agent model of EMU is established by using the sub space identification method. The control objective is synchronous tracking a given speed of each agent. The distribution of traction/ braking force of each power unit is optimized by using the distributed coordination control algorithm with the optimization of the neighborhood, which meets the practical requirements of high-speed EMU. Compared with the centralized control of the multi variable, the distributed control of the small size of each agent can solve the problem of large scale online control in the running process of EMU, and reduce the size and complexity of the problem and improve the control performance.Finally, the simulation and verification platform for high-speed EMU running process are analyzed, and the structure and function of the simulation platform is designed. The main functions of the system are manual and automatic driving, reading and managing line data, operating environment, and traction calculation by the control objectives, EMU train data management, the results output and other auxiliary functions. By using the type of CRH380A EMU data and the actual line data of Beijing-Shanghai high speed railway from Xuzhou East to Tengzhou East, the effectiveness of the aforementioned modeling and control methods for EMU is verified.