Parameter Identification And Its Application In Urban Train Control Model

With the expansion of city scale,traffic congestion is becoming more and more serious in big cities.Urban rail transit has a high efficiency in the utilization of land,which has the characteristics of large volume,fast speed,safe and comfortable,and little impact on the environment.The development of urban rail transit is an important way to solve the problem of urban traffic congestion.Currently,urban rail transit is still in a state of rapid development in China.The parameter identification of train model is necessary for engineering application and theoretical research,which can meet the urban rail transit operation in different weather conditions and diff’erent environment requirements and improve performance of train controller.According to the summary of the train model.traction and braking characteristics,the basic resistance model of train,train traction control model and train brake control model are established.The traction control model of high speed stage and low speed stage is proposed through the analysis of the running speed curve of the urban rail train.The train traction process is divided into the stage of traction establishment,traction process and traction removal.The braking process is divided into the stage of braking establishment,the braking process and the electro-pneumatic brake.Using the error class model to describe the train control process.For the parameter identification of train basic resistance,an improved genetic algorithm is proposed based on the double convergence criterion in this paper.By setting the criterion of judgment for accuracy convergence,the improved genetic algorithm has accuracy and f-ast speed.By comparing the basic resistance parameters of sunny days and showers,the basic resistance parameters of showers changed significantly.For the parameter identification of train control model,the double-closed loop control frame with the motor controlled is proposed by analyze the ATO structure.The train control model is divided into train control equation error model,train control output error model and train control Box-Jenkins model.Recursive least squares class algorithm and stochastic gradient algorithm are adopted to identify parameters.The auxiliary model identification theory was used to solve the problem of unobserved variables of the parameter identification.The multi-innovation identification theory was used to solve the problem of slow convergence of the stochastic gradient algorithm.The online self--tune controller based on white and colored noise train control error class model parameter identification was applied to the train braking stage.The stopping accuracy of the train has great uncertainty by analysis parking accuracy of t Beijing yizhuang line.The online self-tune controller based on white and color noise train control error model parameter identification guarantee the stopping accuracy within 1cm and avoid the influence of the change of environment and train parameters on the parking precision.