大坝升船机地震鞭梢效应智能控制的振动台试验  

Experimental study on seismic response control of large-span building on top of ship lift towers using roof MR intelligent isolation system

作　　者：涂建维[1] 瞿伟廉[1] 陈静[2] 王雪亮[1] 

机构地区：[1]武汉理工大学道路桥梁与结构工程湖北省重点实验室,湖北武汉430070 [2]武汉理工大学自动化学院,湖北武汉430070

出　　处：《建筑结构学报》2007年第4期44-50,共7页Journal of Building Structures

基　　金：国家自然科学基金重大国际合作资助项目(50520130296)

摘　　要：大坝升船机顶部厂房与底部塔柱结构之间的侧移刚度突变造成了顶部厂房强烈的地震鞭梢效应。针对此问题,本文进行了屋盖MR智能隔震系统对大坝升船机地震鞭梢效应智能控制的振动台试验研究。屋盖MR智能隔震系统由隔震垫、磁流变(MR)阻尼器及控制系统组成,安装在厂房的柱子顶部。首先采用修正的Bingham模型得到数据来训练神经网络以反映MR阻尼器的逆向动力特性,然后与模糊算法结合通过dSPACE实时控制系统得到阻尼器的最优控制信号。试验结果表明:模糊-神经控制方法能有效控制屋盖MR智能隔震系统,使得大坝升船机顶部厂房的鞭梢效应明显减小,排架柱顶位移减小了50%,隔震层相对位移减小了57%。A vertical ship lift under earthquake excitation may suffer from a whipping effect due to the sudden change of building lateral stiffness at the top of ship lift towers. This paper presents an experimental investigation on roof magnetorheological （MR） intelligent isolation system to prevent the whipping effect on the machinery building. The roof MR intelligent isolation system which is installed at the column tip of machinery building consists of three sections： seismic isolator, MR damper and dSPACE real-time simulator. The neural network was firstly trained by data from the revised Bingham model to demonstrate the inverse dynamic characteristic of the MR damper. The optimal control signal can then be obtained from dSPACE real-time simulator system manipulated by neuro-fuzzy control technique which is combined with inverse neural network model and fuzzy algorithm. The experimental results show that the roof MR intelligent isolation system with neuro-fuzzy control technique can significantly reduce the seismic whipping effect on the building, namely 50 % reduction in displacement at the column tip and 57 % reduction in relative displacement of seismic isolator.

关 键 词：大坝升船机 模糊 神经网络 MR阻尼器 隔震 鞭梢效应 

分 类 号：TU352.1[建筑科学—结构工程]