结构智能半主动控制系统仿真设计

通过在电液伺服材料性能实验机MTS机上进行磁流变阻尼器动态特性神经网络辨识试验，建立了MR阻尼器的神经网络逆动态特性模型。通过离线训练设计了结构的神经网络辨识器、控制器模型，并提出了MR阻尼器神经网络半主动控制方案。将所建的MR阻尼器的神经网络逆动态模型、结构的神经网络辨识器模型、以及神经网络控制器模型用于结构的神经网络半主动控制并进行仿真，结果表明所建模型及控制方案能够有效降低结构的振动响应，同时最大限度地发挥MR阻尼器的动态特性。

Intelligent Control System Simulink Design of Structure with Semi-active Technology

Based on the test of indicating the dynamical characteristics of MR damper on MTS, a neural network model of the inverse dynamical characteristics of MR damper is established. By training off-line the neural-indictor network and neural-controller network with the idealistic-test, two schemes of neural-indictor network and neural-controlerl network are designed, and the neural semi-active control policy of MR damper is proposed. Computer simulation of a numerical example shows that applying the neural semi-active control strategy with these neural network models of the inverse character of MR damper, the identification of structure, and the neural controller to the control system can reduce the response of structure effectively and take maximum advantage of the properties of MR damper.