Intelligent Vibration Control System with Actuators using MR Fluid

本课题研究的是一种基于磁流变液的智能振动控制系统。磁流变流体是一类可以通过外加磁场强度控制其流变特性的功能流体。通过将磁流变液应用于传统的阻尼器，可以实现半主动阻尼器，其阻尼特性可以由外部施加信号控制。磁流变液一般由悬浮在矿物油或硅油等液体中的铁磁性、微米大小的颗粒组成，颗粒在没有磁场的情况下，通过强表面活性剂均匀地分散在液体中。在磁场作用下，颗粒聚集形成链状团簇结构。当对MR流体施加一定的剪切变形时，团簇可能会摩擦或刷过相反的表面，从而产生摩擦力，表现出流体的宾厄姆塑性行为。摩擦力在很大程度上取决于磁通密度的大小，因此建立闭环磁路是非常重要的。在本项目中，为了预测半主动执行器的磁流变效应，首先估算了磁流变液的等效磁导率。用元胞自动机模拟磁场作用下的星团形成，这是一种离散建模技术。考虑到模拟颗粒结构，采用有限元程序进行了磁性分析，并估算了等效磁导率。在此基础上，构造了可变阻尼器，并将其设置在结构模型上。采用由神经网络组成的智能控制系统对磁流变阻尼器进行控制。实验表明，将磁流变液与神经网络控制系统相结合，实现了智能振动控制系统。

This project is on an intelligent vibration control system with actuators using MR fluid. MR fluid is known as a class of functional fluids whose rheological properties can be controlled by the applied magnetic field strength. By applying MR fluid to some conventional damper, semi-active dampers may be realized, and their damping properties can be controlled by an external applied signal. MR fluid is generally composed of ferromagnetic, micron-sized particles suspended in liquid such as mineral oil or silicone oil, and the particles are dispersed uniformly in the liquid by strong surface active agent without magnetic field. Under magnetic field, the particles aggregate and form chain-like cluster structures. When some shear deformation is applied to MR fluid, the clusters may scrub or brush the opposite surface and friction force may be produced, which show Bingham plastic behavior of the fluid. The friction force strongly depends on the strength of magnetic flux density, and it is very important to make closed loop magnetic circuit.In the present project, the equivalent magnetic permiability of MR fluid was first estimated in order to predict MR effect of the semi-active actuators .The cluster formation under a magnetic field is simulated by Cellular Automata, which is known as a discrete modeling technique. Considerting the simulated particle structures, magnetic analysis was carried out by FEM code and equivalent permiability was estimated. Based on the results obtained above, a variable damper was constructed and set on a structural model. The MR damper was controlled by an intelligent control system composed of neural network. It was shown experimentally that an intelligent vibration control system was realized by the conbination of MR fluid and neural network control system.