Comfort Analysis Of Train Based On Mixed Control Algorithms

Comfort was a persistent topic in train operation with the increase of train speed. Aiming at controlling the change of acceleration, comfort of train was improved through adjusting the speed of the train by Automatic Train Operation systems in many studies.Based on the train-track system model, the advantages and disadvantages of T-S fuzzy control, neural networks and genetic algorithm were investigated in this paper. Main studies were concentrated on controlling active suspension of model in order to decrease amplitude of train vibration. Therefore, mixed control algorithm was presented and the results were showed to be effective through the proposed control algorithm.Firstly, the train-track system model and control algorithm were introduced. Meanwhile, the status of the train vibration model and the control algorithms were also analyzed.Secondly, the train vibration model was given, and then focused on introducing its semi-structured model.Active suspension was regarded as the control object through analyzing the model.Control algorithms and optimized method were adopted on the basis of comparing T-S fuzzy controller with conventional Mamdani fuzzy controller.Thirdly, control strategy based on the selected model was established. Recent years the control scheme of single algorithm was usually adopted. In order to have better effects, parameters of T-S fuzzy controller were selected according to PID method in this paper, and active suspension became the object of T-S fuzzy controller. The superiority was verified through simulation.Fourthly, since the number of rules was zero at the beginning, the if-part of T-S fuzzy controller was optimized by neural network in order to generate fuzzy rules automatically according to kinds of uncertain factors. The experimental result demonstrated that amplitude of vibration was better.Finally, the then-part of T-S fuzzy controller was optimized by genetic algorithm and the better optimized result was gained by comparing it with the ordinary T-S fuzzy controller. Two algorithms were then applied on T-S fuzzy control system simultaneously to improve the comfort performance, thus the mixed control algorithm was proved to be efficient.