升船结构地震鞭梢效应的MR智能隔震控制
|
摘要
提出了用屋盖磁流变(MR)智能隔震系统实现升船结构地震鞭梢效应的神经网络预测半主动控制方法.首先,根据反向传播(BP)误差逆传播网络较强的泛化能力,运用该网络来预测未知地震波,在此基础上,根据地震波的预测值来确定"开关-耗能"半主动控制策略;然后,在该控制策略和预测神经网络的帮助下完成神经网络控制器的训练过程;最后,运用该控制器对结构实施智能控制.应用本控制方法对三峡升船结构屋盖MR智能隔震系统进行了仿真,计算结果表明:层间位移和柱底弯矩均有明显的减小,能有效抑制升船结构顶部厂房地震鞭梢效应,是一种简单并易于实现的智能控制装置.
The method of neural network predicts semi-active control was proposed to apply to the roof MR intelligent isolation system to prevent the whipping effect.First,because of the extension of BP error inverse transfer network,it was used to predict unknown earthquake wave.Based on the predicted result,"on-off" semi-active control strategy was determined.Second,with the aid of semi-active control strategy and BP error inverse transfer network,neural network controller was successfully trained.Finally,the trained neural network controller was applied to the roof MR intelligent isolation system of the ship lift structure.The imitation result indicate: the system can prevent the seismic whipping effect of the top machinery building of ship lift structure, displacement between layers and the moment at the bottom of the pillar are obviously restrained.So,it is an easy and practical intelligent control device.
The method of neural network predicts semi-active control was proposed to apply to the roof MR intelligent isolation system to prevent the whipping effect.First,because of the extension of BP error inverse transfer network,it was used to predict unknown earthquake wave.Based on the predicted result,"on-off" semi-active control strategy was determined.Second,with the aid of semi-active control strategy and BP error inverse transfer network,neural network controller was successfully trained.Finally,the trained neural network controller was applied to the roof MR intelligent isolation system of the ship lift structure.The imitation result indicate: the system can prevent the seismic whipping effect of the top machinery building of ship lift structure, displacement between layers and the moment at the bottom of the pillar are obviously restrained.So,it is an easy and practical intelligent control device.
引文
[1]吴杰芳,陈敏中.三峡工程垂直升船机建筑结构整体抗震计算[J].长江科学院学报,1997,14(3):43-46.
[2]吕明云.大坝升船机地震鞭梢效应的MR智能半主动控制研究[D].武汉:武汉理工大学土木工程与建筑学院,2002.
[3]袁曾任.人工神经元网络及其应用[M].北京:清华大学出版社,1992.
[4]焦李成.神经网络系统理论[M].西安:西安电子科技大学出版社,1990.
[5]Ghaboussi J,Joghataie A.Active control of struc-tures using neural networks[J].J Engrg Mech Div,ASCE,1995,121(4):555-567.
[6]瞿伟廉,王军武,徐幼麟,等.MR智能阻尼器耦联的带裙房高层建筑结构地震反应的半主动控制[J].地震工程与工程振动,2000,20(4):87-95.
[7]Nikzad K,Ghaboussi J,Paul S.A study of actuatordynamics and delay compensation using neuro-con-trollers[J].J Engrg Mech Div,ASCE,1996,122:966-975.
[2]吕明云.大坝升船机地震鞭梢效应的MR智能半主动控制研究[D].武汉:武汉理工大学土木工程与建筑学院,2002.
[3]袁曾任.人工神经元网络及其应用[M].北京:清华大学出版社,1992.
[4]焦李成.神经网络系统理论[M].西安:西安电子科技大学出版社,1990.
[5]Ghaboussi J,Joghataie A.Active control of struc-tures using neural networks[J].J Engrg Mech Div,ASCE,1995,121(4):555-567.
[6]瞿伟廉,王军武,徐幼麟,等.MR智能阻尼器耦联的带裙房高层建筑结构地震反应的半主动控制[J].地震工程与工程振动,2000,20(4):87-95.
[7]Nikzad K,Ghaboussi J,Paul S.A study of actuatordynamics and delay compensation using neuro-con-trollers[J].J Engrg Mech Div,ASCE,1996,122:966-975.
版权所有：© 2023 中国地质图书馆 中国地质调查局地学文献中心