磁流变 ( MR)阻尼器是土木结构振动控制领域最有应用前景的半主动控制装置之一 ,但由于 MR阻尼器的高度非线性动特性 ,使得描述其逆向动特性的数学模型很难得到 ,即根据理想的阻尼力确定出 MR阻尼器所需的输入电压。然而 ,对基于 MR阻尼器的结构振动控制设计和仿真 ,这种反映阻尼力 -电压关系的逆向动特性模型是十分重要的。为此 ,本项研究针对 MR阻尼器的非线性特性 ,提出运用神经网络技术建立 MR阻尼器的神经网络模型来模拟其逆向动特性 ,神经网络模型的输出即为产生理想的阻尼力所需的输入电压。通过数值仿真结果探讨所提出的结构振动控制策略的适用性和有效性 Magneto-rheological (MR) dampers are one of the most promising semi-active control devices in the field of civil structure vibration control. However, due to the highly nonlinear dynamic characteristics of MR dampers, it is very difficult to obtain mathematical models describing their inverse dynamic characteristics , That is based on the ideal damping force to determine the required MR input voltage damper. However, for the design and simulation of MR damper-based structural vibration control, this inverse kinematic model reflecting the damping force-voltage relationship is very important. Therefore, in this study, aiming at the nonlinear characteristics of MR dampers, a neural network model of MR dampers was proposed to simulate the inverse dynamic characteristics of the MR dampers. The output of the neural network model is required to generate the ideal damping force Input voltage. The applicability and effectiveness of the proposed structural vibration control strategy are discussed by numerical simulation results