高速动车组数据驱动无模型自适应控制方法

Data-driven Model-free Adaptive Control Method for High-speed Electric Multiple Unit

针对动车组的速度跟踪控制问题,同时考虑到现有基于模型的控制方法对系统动力学模型的依赖性,以及传统无模型自适应控制时变参数估计算法的复杂性,将改进的多输入多输出(Multiple-input multiple-output,MIMO)偏格式动态线性化无模型自适应控制(Partial form dynamic linearization-improved model-free adaptive control,PFDL-iMFAC)方法引入到动车组自动驾驶系统中.该控制方法在无模型自适应控制的基础上,考虑滑动时间窗口,增加了可调自由度和设计灵活性,并在输入准则函数中加上对能量函数的惩罚项,减少能量损耗,为动车组的跟踪精度和节能运行提供了一种优化的方法,在满足动车组速度跟踪效果好的前提下实现节能运行.最后以CRH380A动车组为对象进行仿真实验,通过与传统无模型自适应控制对比:所提出的控制算法各动力单元速度跟踪误差在±0.2 km/h以内,加速度在±0.65 m/s^(2)以内且变化平稳,比传统无模型自适应控制方法节约9.86%的能量.

For the speed tracking control problem of electric multiple unit,the dependence of the existing modelbased control methods on the system dynamic model and the complexity of the time-varying parameter estimation algorithm of the traditional model-free adaptive control are both considered.The improved multiple-input multipleoutput(MIMO)partial format dynamic linearization-improved model-free adaptive control(PFDL-iMFAC)method is introduced into the automatic train operation system.On the basis of model-free adaptive control,this control method considers the sliding time window,increases the adjustable degree of freedom and design flexibility,and adds the penalty term to the energy function in the input criterion function to reduce the energy loss.It provides a compromise method for the tracking accuracy and energy-saving operation of electric multiple unit,and realizes energy-saving operation under the premise of satisfying the good speed tracking effect of electric multiple unit.Finally,CRH380A electric multiple unit is taken as the object for simulation experiment.Compared with the traditional model-free adaptive control,the speed tracking error of each power unit in the proposed control algorithm is within±0.2 km/h,and the acceleration one is within±0.65 m/s^(2) and the change is stable,saving 9.86%of energy compared with the traditional model-free adaptive control method.