基于变步长极值搜索的列车黏着控制方法研究  

Study on Adhesion Control Method of Train Based on Variable Step-size Extremum Seeking Algorithm

作　　者：柳海科 袁忠于 LIU Haike;YUAN Zhongyu(The Mechanical Engineering College of Lanzhou Jiaotong University,Lanzhou 730070 Gansu,China)

机构地区：[1]兰州交通大学机电工程学院,兰州730070

出　　处：《铁道机车车辆》2021年第6期45-52,共8页Railway Locomotive & Car

摘　　要：为了有效地解决由于外界因素导致的轮轨黏着牵引力利用率不高的问题,提出用全维状态观测器来估测黏着系数,同时采用带遗忘因子递推最小二乘算法(FFRLS)以获得黏滑特性曲线的斜率值,进而引入变步长梯度下降法搜寻最优蠕滑速度以调节转矩补偿值,可实现在恶劣轮轨路况下高速列车黏着力仍可高效利用的目的。文中将在Matlab/Simulink平台对此研究策略的可行度和有效性进行验证。In order to effectively solve the problem that wheel-rail traction can not be fully utilized due to external factors,a full-dimen⁃sional state observer is proposed to calculate the adhesion coefficient,and using the forgetting factor least squares algorithm(FFRLS)to obtain the slope of the adhesion-slip characteristic curve.Then using variable step-size gradient descent method to seek the optimal slip speed to adjust the torque compensation value,so that the adhesion of high-speed trains can still be used efficiently under terrible wheel and rail conditions.In this paper,the Matlab/Simulink simulation is used to verify the feasibility and effectiveness of the proposed method.

关 键 词：黏着控制 变步长梯度下降法 带遗忘因子最小二乘法 黏着力 黏滑曲线斜率 

分 类 号：U260.115[机械工程—车辆工程]