基于磁流变悬置的整车建模与振动控制的研究

摘要

With the results of dynamic performance test of magneto-rheological ( MR) mount as data sam-ples, the forward and inverse models for MR mount are established by adopting identification technique with general-ized regression neural network ( GRNN) , and the identification models are used for mount control. Then a 10 DOF vehicle dynamic model with MR mount is set up, a fuzzy controller is designed with the signals of engine rotation speed and accelerations at mounting points as input, and a simulation on the semi-active control of MR mount is conducted. The results show that the fuzzy controller designed has better results of broadband vibration isolation, and the vibration acceleration peak at mounting points reduces significantly, verifying the correctness and effective-ness of GRNN model and fuzzy controller algorithm.%以磁流变悬置动态性能试验结果为数据样本，利用广义回归神经网络( GRNN)辨识方法建立了磁流变悬置的正、逆模型，并将辨识模型用于悬置系统的控制；建立了基于磁流变悬置的整车10自由度动力学模型，以发动机转速和悬置点处的加速度信号为输入，设计了模糊控制器，对磁流变悬置进行半主动控制的仿真。结果表明：该模糊控制器具有较好的宽频隔振效果，悬置点处的振动加速度峰值明显减小，验证了GRNN模型和模糊控制器算法的正确性和有效性。