Gain-Scheduled Vibration Isolation Control Using Semi-Active Actuator

我研究了具有可变阻尼机构的半主动执行器的建模，并构建了新的阻尼器模型，该模型假设阻尼系数随可变阀机构线性变化，其中伺服阀可以采用节流孔开度。我假设阀的位移受一个非线性弹簧的限制。将阻尼器模型应用于汽车悬架，并介绍了涉及可变阻尼器模型的线性参数变系统。因此，我设计了增益计划控制系统，仿真结果验证了阻尼系数根据悬架状态的变化是平滑的，所设计的控制器可以完成指定的控制。此外，我已将所提出的方法应用于磁流变流体阻尼器。将磁流变阻尼器的特性识别为输入电流与阻尼系数之间的关系。我把这种关系近似为笔划速度的五分之一次方的函数。并设计了将输入电流输入到磁流变阻尼器的控制系统，使磁流变阻尼器的系数与所提出的增益调节控制器的输出相吻合。通过实验验证了该方法的有效性。实验还验证了我设计的卡尔曼滤波器可以估计冲程速度。我还研究了半主动控制系统在四自由度结构中的应用。我针对降阶模型设计了所提出的控制系统。仿真结果表明，在阻尼系数变化范围较小的情况下，该方法能有效抑制振动。在变量范围变大的情况下，通过将控制器切换为行程速度符号，发现了较高模式的激励问题。

I have examined a modelling of a semi-active actuator with a variable damper mechanism and have constructed the new damper model whose damping coefficient is assumed to be linearly changed by the variable valve mechanism, where the servo valve can adopt the orifice open rate. I assume that the valve displacement is restricted by a nonlinear spring. The damper model has been applied to a vehicle suspension and I have introduced the linear parameter varying system involving the variable damper model. Thus, I have designed the gain-scheduled control system and it has been verified from simulation results that the damping coefficient is smoothly changed according to the state of the suspension and the proposed controller can perform the specified control.Furthermore, I have applied the proposed method to a Magneto-Rheologial fluid damper. The characteristics of the MR damper were identified as the relation between input current and damping coefficient. I approximated the relation as the function of the one fifth power of stroke velocity. And I have designed the control system that put the input current to the MR damper, so that the MR damper coefficient coincides with output from the proposed gain-scheduled controller. By carrying out experiments effectiveness of the proposed method has been verified. Also it has been verified that a Kalman filter I designed can estimate the stroke velocity.I have also examined application of the semi-active control system to a four-degree-of-freedom structure. I designed the proposed control system for the reduced-order model. It has been verified from simulation results that the proposed method can suppress the vibration when the variable range of damping coefficient is relatively small. I have found out excitation problem of the higher modes by switching of the controller with sign of the stroke velocity in case that the variable range becomes large.