Adaptive P Control Based On Neural Network And Its Application In Train Traction Control

Due to the impact of road conditions and external environment,the train wheelsets are still faced with unbalanced such as slippage or jolt in the process of traction or braking.In this dissertation,the dynamic model for the train wheels is improved,and a controller for the speed tracking is designed based on the neural network adaptive algorithm.In addition,the synchronous controller based on the cross control method is designed to coordinate the control between the wheelsets.The research mainly includes the following parts:Firstly,the mechanics mathematical model of the train wheelset is established and improved by analyzing the force condition.Emphasis is placed on the analysis of influence by external disturbances that may be encountered during the train operation,the efficiency factor of equipment,and the effects of differences among multiple objects.The adjacent cross-coupling control method is used to solve the problem of inconsistent control between multiple wheelsets.The tracking controller and synchronous coordination controller are designed.Secondly,a control strategy is proposed based on the neural network adaptive algorithm.The number of neurons and the selection of activation function are further analyzed in the neural network structure.The generalized neuron self-increasing algorithm and the new network structure of multi-layered grouping are adopted.In the network structure,each group can select different activation function and adjust the number of neurons according to the error value.This strategy enhances the network structure and improves the learning ability of the network with adjustable number of neurons.Then,the speed tracking controller and synchronous coordination controller are designed respectively,based on the neural network adaptive algorithm.The speed tracking controller is used to control the speed of the train driving wheelset and the synchronous coordination controller is used to coordinate the multiple wheelsets.The advantages of the proposed strategy are shown by comparing with the existing algorithms.Finally,the feasibility of the algorithm and the effectiveness of the controller are verified via simulations according to the data of the target running lines.The results show that the designed controllers can meet the goal of train speed control and reduce the speed difference among each wheelset with improved performance.It helps reduce the physical difference caused by the wheel rail or other equipment and ensure the smooth running of the train.