面向长大下坡路况的虚拟编组重载列车分层控制研究OA
Research on layered control of virtual coupling heavy-haul trains for long downhill road conditions
为解决考虑非线性阻力的重载列车在长大下坡路况下面临的群组控制问题,利用线性二次优化、扩展卡尔曼滤波和预估迭代模型预测控制设计一种分层控制策略.首先,该方法针对虚拟编组重载列车的群组控制需求,将基本阻力、附加阻力和空气制动力视为列车总扰动,对包含非线性项的总扰动的列车动力学模型建模;其次,分层控制上层采用线性二次优化模型,通过优化协调进行全局轨迹规划,计算虚拟编组重载列车在长大下坡安全约束下的最优参考轨迹;下层采用基于扩展卡尔曼滤波的预估迭代模型预测控制,利用扩展卡尔曼滤波解决测量噪声对虚拟编组控制稳定性的影响,通过预估迭代模型预测控制解决最优约束下的非线性控制,先利用预估迭代求解模型非线性项,再通过优化求解运行轨迹,实现跟踪上层参考曲线,有效控制重载列车在长大下坡路况群组间距,防止列车冲撞,提升重载铁路运力.最后,通过与自适应控制、鲁棒模型预测控制和传统模型预测控制对比,仿真结果表明,所提出的方法对长大下坡路况下群组间距实现有效控制,速度控制精度分别提高约23.9%、33.4%和45.1%,在位置控制精度上提高约26.6%、29.1%和33.6%.实验验证了分层控制策略的有效性,并证明了基于扩展卡尔曼滤波的预估迭代模型预测控制在复杂场景下的鲁棒性和准确性,为虚拟编组重载列车在长大下坡路况下的安全运行提供了有效解决方案.
To address the problem of group control of long downhill road conditions faced by the heavy-haul trains in view of the non-linearity in its frictional resistance,a condensed hierarchy of tempo-spatial transfer controls was designed using linear quadratic optimization,an extended Kalman filter,and predictive iterative model control.The first one was a solution that satisfied the group controlling needs of heavy-haul trains with virtual couples,in which basic resistance force,additional resistance force,air braking force were regarded as the total disturbance,and the dynamic train model of the total disturbance with nonlinear terms was developed.Second,the higher layer of the hierarchical control used a linear quadratic evaluation model,and a global trajectory planning through the coordination of the optimization was calculated in the optimal reference trajectory of the virtual coupling heavy-haul trains under safety constraints on the long and steep downhill slopes.Third,the predictive Nonlinear control within optimal limits was achieved with predictive iterative model,the predictive iterative method was used to calculate the nonlinear term in the model,and the running trajectory was optimized to follow the upper reference curve and maintain the distance between groups of heavy-haul trains on long and steep downhill road conditions.This allows avoiding collisions and enhancing capacity of heavy-haul railways.Lastly,a comparative study with adaptive control,robust model predictive control,and traditional model predictive control proves the effectiveness of the proposed method in controlling the group spacing down a long downhill road and improving the speed and position control accuracies by nearly 23.9%,33.4%and 45.1%and 26.6%,29.1%and 33.6%,respectively.The hierarchical control scheme performance was experimentally confirmed,and the stability and precision of the predictive control by the predictive iterative model using extended Kalman filter when applying in the complex scenes has been revealed.This is a potential solution to the safe control of the virtual coupling heavy-haul trains in long and steep downhill roads.
魏文军;王洁;刘艳浩;周仲成
兰州交通大学 自动化与电气工程学院,甘肃 兰州 730070兰州交通大学 自动化与电气工程学院,甘肃 兰州 730070兰州交通大学 自动化与电气工程学院,甘肃 兰州 730070兰州交通大学 自动化与电气工程学院,甘肃 兰州 730070
交通工程
虚拟编组重载列车预估迭代模型预测控制扩展卡尔曼滤波
virtual couplingheavy-haul trainsestimation iterationmodel predictive controlextended Kalman filter
《铁道科学与工程学报》 2026 (4)
1612-1624,13
国家自然科学基金资助项目(61863023)甘肃省自然科学基金资助项目(23JRRA868)
评论