高速铁路轮轨作用力的全息辨识模型  

HOLOGRAPHIC IDENTIFICATION MODEL OF WHEEL & RAIL CONTACT FORCE FOR HIHG-SPEED RAILWAY

作　　者：孙善超[1] 刘金朝[1] 王卫东[1] 刘小明[2] 魏悦广 SUN Shan-chao;LIU Jin-zhao;WANG Wei-dong;LIU Xiao-ming;WEI Yue-guang(China Academy of Railway Science,Beijing 100081,China;Institute of Mechanics,Chinese Academy of Sciences,Beijing 100190,China;College of Engineering,Peking University,Beijing 100871,China)

机构地区：[1]中国铁道科学研究院,北京100081 [2]中国科学院力学研究所,北京100190 [3]北京大学工学院,北京100871

出　　处：《工程力学》2018年第11期190-196,共7页Engineering Mechanics

基　　金：中国铁路总公司重点课题项目(2017G011-B;2017G011-E);国家自然科学基金项目(11432014;11572329)

摘　　要：结合最优控制理论、卡尔曼滤波方法以及轮轨蠕滑理论,该文建立了轮轨作用力的全息辨识模型,并将模型的反演结果与动力学解答和实验测量结果作对比,验证了模型的适用性。首先,根据车辆系统的状态空间方程,设计一个最优化加速度状态跟踪器,将轮轨作用力辨识问题转换成为最优控制策略的设计问题;然后,利用SVD(singularvaluedecomposition奇异值分解)分解技术对最优控制问题进行逆向求解,完成系统的动态载荷预测值的辨识;进一步结合卡尔曼滤波技术及加速度测试值,对辨识动载荷预测值进行正向修正,得到左、右轮轨垂向力及轮轴横向力;最后,结合轮轨蠕滑理论,将左、右轮轨横向力转换为到左、右轮轨垂向力及轮对横移量的函数,消减不适定性,从而能够分别辨识出轮轨左、右的横向力。辨识模型得到的结果与动力学解答的相关系数分别达到0.65、0.8以上,与实测载荷的相关系数分别达到0.51、0.69以上。Combined with optimal control theory,wheel rail creep theory and Kalman filtering method,a holographic model is established for the identification of wheel&rail contact forces.The identification results were compared with numerical simulation and inspection data.Firstly,the identification problem was converted into the design issues using an optimal control strategy,an optimal acceleration state tracker was developed by using the state space equation of a vehicle system.Secondly,the design issues were inversely solved using SVD technique,and the evolution history of contact forces were obtained.The predicted contact forces were positively updated using Kalman fliting method.Finally,the relation between contact forces and wheelset displacements were updated by using the wheel-rail creep theory,the ill-posedness problem was thusly solved,and independent left/right lateral forces can be obtained.The identification results were compared with the numerical simulation and the experimental finding,and it showed that the correlation coefficient of identification lateral/vertical forces and simulation forces were 0.69/0.8,and that the correlation coefficient of identification lateral/vertical forces and verification forces were 0.51/0.69.

关 键 词：最优控制 SVD分解 卡尔曼滤波 轮轨蠕滑理论 全息轮轨力辨识模型 

分 类 号：U211.5[交通运输工程—道路与铁道工程]