动车组列车制动系统 Hammerstein 模型的广义预测控制研究

摘要

Considering the importance of the EMU braking controller in ATO ,this paper proposed the General-ized Predictive Control (GPC ) over the EM U braking system based on its Hammerstein model .Firstly ,the Hammerstein model was seen as a series system that included a static subsystem and a dynamic subsystem .The dynamic subsystem was described as the CARIMA model .The static subsystem was described as a static func-tion .The Mind Evolutionary Algorithm (MEA) amended the errors of the model of the dynamic subsystem caused by time delay and external disturbance .Based on these thoughts the non-line generalized predictive con-troller of the dynamic subsystem was designed and the intermediate variable ,namely ,the input of the dynamic subsystem ,was obtained .Secondly ,a corroding to the characteristics of the EMU breaking system ,the interme-diate variable was constrained in order to correspond to the actual control value .Then ,the inverse algorithm of the static function was executed and the input of the EM U breaking system was achieved .Finally ,simulation re-sults of EM U CRH2 prove that the M EA amending method is effective and the GPC is superior to PID .%鉴于动车组列车制动控制在运行和ATO中的重要性，以 Hammerstein模型为基础，设计动车组列车制动系统的广义预测控制GPC器。把Hammerstein模型看作静态子系统和动态子系统的串联，动态子系统辨识为CARIM A模型，用思维进化算法M EA辨识由动态子系统纯延时环节和外界干扰造成的模型误差，设计基于M EA误差修正的GPC器，得出中间量。根据动车组列车制动特性对中间量进行约束化处理，使处理后的中间量和制动级位实现一一对应关系。对描述静态子系统的静态函数求逆，得到制动系统的制动级位。以C R H2型动车组为仿真对象，比较PID和GPC的控制效果，证明M EA修正误差的有效性，验证GPC器控制动车组列车制动系统的优越性。