主动悬架用直线电机模型预测推力控制

针对主动悬架用直线电机高精度与高效率的控制需求，充分研究直线电机以及主动悬架动力学特性，建立直线电机驱动模型与主动悬架二自由度参考模型。为了改善传统直线电机直接推力控制的动态性能，提出一种改进的模型预测推力控制方法。该方法将逆变器产生的7个非零电压矢量作为备选矢量，并融合预测模型来计算出控制周期内的电机运行状态参数，基于成本函数最小值原理挑选出最优电压矢量，并将其作用于逆变器产生驱动电机所需的电压。为了解决数字控制系统的固有问题，提出延时补偿技术，保证对电机能够进行实时控制；对于逆变器开关频率不固定而引起的开关损耗等问题，通过在成本函数中加入开关频率项，在选择最优电压矢量的同时还降低了逆变器的整流频率与开关损耗；另外为了提高电机推力的效率以及减小推力与磁链波动，提出最大推力电流比与占空比优化控制技术，提高直线电机的动态控制性能。基于MATLAB/Simulink与dSPACE联合仿真，并搭建直线电机与主动悬架硬件测试平台，对所提出的控制方法进行验证，同时对主动悬架系统的动力学性能也进行了仿真与试验测试。试验结果表明：相较于传统的直接推力控制，所提出的控制策略使电机能够拥有更快的稳态速度、更小的电磁力与磁链波动以及更低的开关频率；各工况下，轮胎动载荷试验与仿真结果均方根值的相对误差分别为12.3%、4.47%、6.3%；悬架动行程试验与仿真结果均方根值的相对误差分别为10.3%、8.86%、10.6%；车身加速度试验与仿真结果均方根值的相对误差分别为6.23%、9.12%、7.2%。由计算结果可知，各评价指标的相对误差均在13%以内，验证了仿真结果的正确性，证明了模型预测推力控制对于提升电机与悬架动力学性能的有效性，能够实现对车辆悬架系统全局工况性能最优，协调控制悬架系统的动力学性能。

Improved Model Predictive Thrust-force Control of Linear Motors for Active Suspensions

To meet the high-precision and high-efficiency control requirements of the linear motor used in an active suspension, the dynamic characteristics of the motor and the suspension were fully studied, and a driving model of the motor and a 2-DOF reference model of the suspension were established. To improve the dynamic performance of traditional linear-motor direct thrust control, an improved model predictive thrust-force control method was proposed. In this method, seven non-zero voltage vectors generated by the inverter were selected as candidate vectors, and the prediction model was fused to calculate the motor running-state parameters in the control cycle. Based on the principle of the minimum cost function, the optimal voltage vector was selected and applied to the inverter to generate the required voltage for driving the motor. To address the inherent problems of a digital control system, delay compensation technology was proposed to ensure real-time control of the motor. For the switching loss caused by the unstable switching frequency of the inverter, the switching frequency and switching loss of the inverter were reduced in the selection of the optimal voltage vector by adding the switching frequency term into the cost function. In addition, to improve the efficiency of motor thrust force and reduce the fluctuation of thrust and flux, maximum thrust current ratio and duty ratio optimization control were proposed to improve the dynamic control performance of the linear motor. Based on MATLAB/Simulink and dSPACE, a hardware test platform for the linear motor and active suspension was constructed. The control method was verified, and the dynamic performance of the active suspension system was simulated and tested. The relative errors of the root mean square of the tire dynamic load test and simulation results are 12.3%, 4.47%, and 6.3%, respectively; those of the suspension dynamic stroke test and simulation results are 10.3%, 8.86%, and 10.6%, respectively, and those from the acceleration test and the simulation results are 6.23%, 9.12%, and 7.2%, respectively. According to the calculation results, the relative error of each evaluation index is less than 13%. This verifies the correctness of the simulation results and demonstrates the effectiveness of model predictive thrust control in improving the dynamic performance of the motor and suspension, achieving globally optimal performance of the vehicle suspension system and coordinating its dynamic performance.