A Study Of The Heavy-haul Trains Locomotive Brake Control Strategy Based On Fuzzy Model Predictive Control

To meet the ever-increasing demand for the railway transportation, the operation of heavy-haul trains is an effective way to improve transport capacity. Train Braking is the chief problem to be solved for heavy-haul transport. Due to the longer train formation and increased weight of the train, the original brake control methods are difficult to achieve rapid, accurate and synchronized brake. Consequently the longitudinal force of the train is increased and it will easily lead to major accidents such as coupler broken and derailment. In this thesis, the control strategies of the heavy-haul trains locomotive brake system are studied thoroughly from three aspects: the model predictive control Fuzzy modeling, the rolling optimization and the feedback correction. The main research works are as follows:The working principle of the new heavy-haul trains locomotive braking system is analyzed. For the main controlled object of the system Brake-cylinder, which is time-variation, delay and nonlinear, by adopting the modeling method based on T-S fuzzy model, an optimization method based on Co-evolutionary Genetic Algorithm is proposed to acquire a set of optimization model parameters. Then, the T-S fuzzy model based on the Fuzzy C-Means clustering algorithm for the heavy-haul trains locomotive brake control system is built to improve the accuracy of model identification as well as ensure the rapidness of identification. Based on it, by simulating and comparing several typical control algorithms for the heavy-haul train locomotive brake control system, a predictive control algorithm based on T-S fuzzy model is introduced.In order to reduce the influence of the model parameter, the noise coupling and the random interference on the stability of the heavy-haul trains locomotive brake control system, a predictive control rolling optimization method with optimizing strategy is proposed. The optimization method based on fuzzy genetic is used to obtain the global optimal solution, which is used as a controller output. An optimize group selection strategy based on the recursive least squares algorithm is applied to improve the rapidness and stability of the regulation in the predictive control algorithm. A feedback correction method with compensation is proposed. By using kernel principal component and adjacent support vector machines based soft measurement technology, the dead time of high-speed electrovalve is accessed, combining with pressure feedback caused by air pressure thermal effect, the compensation amount for feedback correction is calculated using compensation formula with weighted factors, which is to reduce the serious nonlinearity and delay problems on locomotive brake control system brought by dead time of the high-speed electrovalve and the transduction process of the compressed gases in brake cylinders.According to Lyapunov stability criteria, a T-S fuzzy model-based predictive controller is designed and its stability is proved. Aiming at the -features of the heavy-haul train locomotive brake control system, such as delay in brake cylinder pressure control process, time-variation of parts parameters and non-linear process of gas exchange with uncertainty interference, the locomotive brake control system was simulated and analyzed to validate the effectiveness of the proposed predictive control strategy.For the issue of multiple brake sources synchronized brake in haul train, by analyzing the impact on synchronized brake by two key factors, which are the steady-state error produced by precision control in the process of brake control and the press margin caused by the difference of control features in dynamic adjusting course, the T-S fuzzy model predictive control strategy is developed as a solution for the heavily-haul train locomotive brake control. Moreover, the train Longitudinal Dynamics Model is adopted to analyze the force on the coupler during the brake course, and the simulation results are compared to prove the effectiveness of the proposed predictive control strategy for synchronized brake.The heavy-haul train locomotive brake control system adopts PC104 framework based high-performance embedded hardware system, works on the QNX real-time multi-tasking operating system platform and uses ISaGRAF software development tools based on client/server model to achieve the T-S fuzzy model based predictive control strategy. With embedded software programming skills, the computing time is reduced, thus improved the real-time feature of the control system. By accomplishing laboratory double-heading debugging, the control system has been applied to the heavy-haul train on Shenhua railway for freightage. The practical application shows that the strategy can achieve synchronized brake rapidly and accurately while completely meet the technical indicators of the heavy-haul train locomotive brake control system.