Research On Direct Torque Intelligent Control Of Train Traction Transmission System

With the rapid development of economy,people need more trains with the hign quality and performance for transportation and travel.The train control system has always been drawn the attention of experts.The optimization of train performance can not only promote the green environmental protection,but also to enhance the comprehensive national strength and capacity,which is one of the key developmental disciplines in china.It has been an important research field for the high quality and stable operation of traction transmission system with trains.Firstly,this paper analyzes the general composition of the train traction drive system and the working principle of each components aim for optimizing the core part which called the three-phase asynchronous motor.From the mathematical model of asynchronous motor,we study the core idea and the basis of direct torque control(DTC)in-depth.It was based on get the deviation of stator flux and electromagnetic torque and follow the flux linkage area to choose voltage vector selection for achieving reasonable control of motor.Although the direct torque control way is used widely in the field of AC speed regulation field,it also has some defects such as unstable switching frequency,large ripple of torque,abnormal fluctuations under the condition of low speed and so on.Therefore,we consider the combination of intelligent control strategy to optimize it,so that it can not only maintain the good performance of direct torque control,but also overcome its own shortcomings.Then,in order to sovle these problems caused by the mutation of motor speed or load,or fluctuation of torque,we design a fuzzy adaptive PI regulator combined with fuzzy control which owns good robustness.We try to use the fuzzy PI controller instead of the traditional speed control ways.This intelligent way can utilize the proportion of integral coefficient to feedback dynamic data online at any work time,so we can get an improved stable system.Furthermore,because neural network have the characteristics of parallel computing and strong fault-tolerant ability,which can deal with the nonlinear relationship accurately and quickly,we introduce this intelligent algorithm model and build a BP neural network for the optimization of the voltage vector controller.However,by controlling the switch state selection table to control the voltage space vector,and then control the motor running state of this complex process,although the use of fuzzy control can be better chosen the switching table,but it is heavily dependent on artificially set fuzzy quantity of subjectivity,and no self learning adaptation function.If fuzzy rules set excessively accurate that would affect the response performance of the system.In this regard,the neural network is introduced with the ability of parallel processing and self training and learning ability of the neural model,for the demand of handling this nonlinear relationship.Finally,the traditional direct torque control system model and the system which is optimized by intelligent strategy are both constructed on MATLAB.By comparing the results of the two simulation models,it is simply known that the simulation results verified that these control schemes are more feasible and stable and achieved the anticipative effects.