电控机械式变速箱换档过程迭代学习控制  

Iterative Learning Control for Gear Shifting Process in Electrical Mechanical Transmission

作　　者：何凯 林成涛[1] 李亮[1,2] 王翔宇 HE Kai;LIN Chengtao;LI Liang;WANG Xiangyu(State Key Laboratory of Automotive Safety and Energy,Tsinghua University,Beijing 100084;Collaborative Innovation Center of Electric Vehicles,Beijing 100084)

机构地区：[1]清华大学汽车安全与节能国家重点实验室,北京100084 [2]电动车协同创新中心,北京100084

出　　处：《机械工程学报》2019年第4期84-90,共7页Journal of Mechanical Engineering

基　　金：国家重点研发计划(2017YFB0103502);国家自然科学基金(51675293)资助项目

摘　　要：电控机械式变速箱(Electrical mechanical transmission, EMT)的换档过程是一个复杂的非线性多刚体动力学过程。由于系统建模困难,不确定性大,采用传统的控制算法往往无法得到满意的控制效果,从而极大地限制了换档过程控制方法的发展。针对这些问题,分析了换档过程的特性,将整个换档过程视为重复性过程,提出采用迭代学习控制的方法来优化换档过程。为此,采用未知输入观测器对换档阻力进行估计;然后设计线性状态反馈控制器得到初始控制输入;进而引入迭代学习控制方法来补偿控制误差,并且考虑了迭代学习过程中期望轨迹的设置,以得到良好的换档控制效果;最后,通过台架试验验证了算法的可行性和有效性。Gear shifting process in electrical mechanical transmission(EMT) is a complicated nonlinear multi-body dynamic process. Because of the difficulty in system modeling and huge uncertainty of this process,it is difficult to obtain favorable control performance by using traditional control methods. Hence,these problems largely hinder the development of corresponding control methods for this process. In face of these problems,the paper analyzes the characteristics of the entire process and consider that the process is a repetitive execution,then presents an approach of adopting iterative learning control(ILC) method to optimize gear shifting process. To realize this purpose,unknown input observer is used to estimate the gear shifting load force,then design a linear state feedback controller to generate initial control inputs. After that,ILC is adopted to compensate the control error and considering the construction of the desired trajectory. Finally,experimental results have shown that the proposed scheme is feasible and efficient.

关 键 词：迭代学习控制 换档过程 电控机械式变速箱 未知输入观测器 

分 类 号：TG156[金属学及工艺—热处理]