基于降阶模型的斜拉索振动的半主动神经网络控制

Semi-active neuro-control of cable vibration based on reduced-order model

根据模态降阶理论,获得了斜拉索-阻尼器系统的降阶模型,有效地缩减了系统的自由度.根据ER/MR阻尼器特性和主动控制中LQG控制理论,建立了面向速度剪切的半主动LQG控制方法,并获得了很好的控制效果.本文设计了神经网络的观测器,使用的传感器数目大大减少,根据智能控制理论,设计了神经网络控制器,并提出采用该神经网络作斜拉索半主动控制的控制算法.振动仿真的结果表明,经过离线训练后的神经网络观测器和控制器,有效地抑制了斜拉索的振动.

This paper designs a semi-active neuro-control of cable vibration based on a reduced-order model to suppress the cable vibration, using the smart elctro/magneto rheological (ER/MR) dampers. The smart semi-active ER/MR damper has great advantages in vibration control in civil engineering, and has attracted much attention from researchers. The paper proposes a neuro-estimator and neuro-controller for the semi-active vibration control of stay cables. The effectiveness of the proposed neural network control strategies is numerically verified by applying to a 12 m long scale-model cable prototype connected with an MR damper near the lower anchorage. The generalization of the proposed two neural networks is verified, which shows the adaptive capability of the configured neuro-controllers under the dynamic excitation distinct from training. The analysis results show that the proposed control strategies can effectively implement semi-active vibration control of stay cables in the use of MR dampers.