扭杆式履带车辆机电悬架的非线性随机最优振动控制1)
NONLINEAR STOCHASTIC OPTIMAL CONTROL OF ELECTROMECHANICAL SUSPENSION OF TORSIONAL TRACKED VEHICLE1)
-
摘要: 履带车辆在崎岖的路面上行驶时,其车体在随机路面作用下产生剧烈振动,严重影响作战人员的工作环境与行进间的射击精度,因此需要对车体的振动加以控制。本文研究了考虑电机执行器有界时扭杆式履带车辆机电悬架非线性随机振动的最优控制。首先根据扭杆式悬架空间结构和空气弹簧力学模型,建立了机电悬架的非线性随机动力学模型。随后,基于随机动态规划原理建立机电悬架系统动态规划方程,考虑到电机执行器的执行能力有限,求解该方程确定了随机最优有界控制策略。最后,通过对比受控与未受控系统的响应统计量,对提出控制策略的控制效果进行评估,并讨论了控制回路时滞对控制效果的影响规律。结果表明提出的控制策略可以有效降低履带车辆机电悬架系统的非线性随机振动,并且控制策略对系统参数变化具有较好的鲁棒性,为保证主动控制力的有效性,时滞时间应控制在0.15 s以内。Abstract: When a tracked vehicle runs on a rough road surface, the vehicle body usually undergoes strong vibration, which significantly affects the working environment of the combat personnel and the shooting accuracy during the motion. Therefore, the vibration control of the vehicle body is much required. In this paper, a nonlinear stochastic optimal vibration control strategy for the electromechanical suspension of a torsional tracked vehicle with consideration of the actuator saturation is proposed. Firstly, the nonlinear stochastic dynamic model of the suspension system is built based on the space structure of the torsional bar and the nonlinear model of the air spring. Then, based on the stochastic dynamical programming principle, the dynamical programming equation of the electromechanical suspension system is established. The optimal bounded control strategy is formulated by solving the dynamical programming equation with consideration of the saturation of the actuator. Finally, the effectiveness of the proposed control strategy is evaluated by comparing the statistics of the responses of optimally controlled and uncontrolled systems. The influence of the time-delay on the control effectiveness is also discussed. Numerical results show that the proposed control strategy can effectively mitigate the nonlinear stochastic vibration of the electromechanical suspension, and the control strategy is found to be quite robust against the variations of the system parameters. To ensure the effectiveness of the control force, the time delay should be controlled within 0.15 second.