为消除磁流变阻尼器逆向模型输出阻尼力与控制电流和活塞相对位移之间存在的非线性滞回特性以提高泛化程度和控制实时性,利用反向传播(BP)神经网络建立磁流变阻尼器逆向模型,采用遗传算法对神经网络的结构、阈值和权值进行了优化,对预测电流进行了误差分析.将该优化的磁流变阻尼器神经网络逆向模型应用到汽车半主动悬架控制系统中进行仿真,结果表明:与期望电流值相比,优化后 BP 神经网络逆向模型的控制电流预测误差较优化前减小;与优化前相比,采用优化后的逆向模型的悬架的垂向加速度、俯仰角加速度和侧倾角加速度均减小,控制实时性增强.
Eliminating the non-linear hysteresis characteristic among the magnetorheological(MR) damper force,the relative displacement and the control current leads to improve the degree of generalization and control of real-time.The genetic algorithm is used to optimize BP neural network reverse model of MR damper,including structure,weight and threshold.The error of the predicted current is analyzed.The MR damper reverse model is applied to semi-active vehicle suspension control system.The results show:The prediction error of the control current of the BP neural network is lower than before the optimization.Vertical acceleration,pitch acceleration and roll acceleration of the suspension are reduced,compared with before the optimization.It enhances real-time control of the system.