【作者】 张强；

【导师】 何玉敖；

【作者基本信息】 天津大学 ， 结构工程， 2008， 博士

【摘要】 土木工程结构有钢筋混凝土结构、钢结构等多种形式,它们均有强烈的非线性特征。经典的结构抗震设计均采用线弹性模型,只能计算多遇地震即所谓小震阶段所受力。而大震阶段验算结构应为非线性结构。结构控制是结构抗震中有发展前途的一种方法。目前对线性结构的控制理论研究已较为成熟。但对非线性结构的控制研究较少。本文主要针对非线性结构的结构控制方法作了研究,主要内容包括:1、首先研究对非线性结构直接进行非线性控制。将非线性控制(基于滞回非线性动态分离的位移加速度反馈结构控制方法)方法引入到非线性结构振动控制中,对其在非线性结构振动控制中的应用作了初步的研究,并建立了非线性结构振动控制系统的非线性控制模型。仿真结果表明将非线性控制方法应用于非线性振动控制是可行的,达到同样控制效果所需控制力比一般线性最优控制方法都要小,尤其在强震下更为显著。2、采用逆系统方法,将非线性结构精确反馈线性化,得到伪线性结构,再采用传统的线性控制理论进行控制。研究结果表明,该方法效果比一般的线性理论控制方法要好,扩大了线性理论适用范围。3、采用神经网络逆系统方法,结合神经网络及逆系统方法的优点,对不易建模的强非线性结构进行控制。研究表明,该方法结合了逆系统方法和神经网络方法的优点,可用于复杂非线性结构的抗震控制。建模时可将原结构化为线性和非线性模块的叠加,以减少建模难度。4、将预测控制与神经网络相结合,利用神经网络强大的非线性建模能力建立原结构模型和逆系统模型。采用基于神经网络的预测控制方法,对强非线性结构进行非线性预测控制。对线性控制和采用神经网络预测的主动控制方法的效果作了比较研究。可以看出,进行神经网络预测控制后,位移得到有效控制,控制效果显著。更多还原

【Abstract】 There are many kinds of civil engineering structures such as reinforced concrete structures and steel structures and so on, and they all have strong nonlinear characteristics. In classics seismic design of structures, with the linearity and resilience presume, we can only analyze frequent earthquakes. But in major earthquakes phase, the calculating model of structures ought to be nonlinear. Structure control is one kind of promising means to structure seismic design. At the moment, control theory for the linear structure has already been developed very maturely. Yet it is still a difficult problem for seismic control of nonlinear structure. The research aimed mainly at nonlinear structure control means .The main research work consists of four aspects:1. Directly nonlinear control is applied to nonlinear structures. With introduction of nonlinear vibration control method into nonlinear structures, it is got as the preliminary research result to the application in the nonlinear structure vibration, and it has been established for nonlinear model to nonlinear structure vibration control system. The simulation result indicates that applying the nonlinear control method to the nonlinear vibration control is feasible, with smaller control force than general linear optimum control method to achieve the similar control effect, especially more remarkable under strong earthquakes.2. Through accurately feedback linearization, and with inverse system approach ,it can change from nonlinear structures to pseudo linear system to which traditional linear control theory could be applied. Research results show the means is superior to linear theory control method and the application field is extended.3. With neural network inverse system method which combining the neural network approach with inverse system approach, it is controlled for strong nonlinear building vibration which is difficult to set up model. From this research, it makes known that the approach have both merits of inverse system approach and that of neural network approach, and it is usable for seismic control of complex nonlinear structures. Original structures may be transformed into superimpose of linear module and nonlinear module in order to decrease difficulty in building model.4. With predictive control approach based on neural network, nonlinear prediction control was implemented for strong nonlinear structure vibration. The comparison research has been made between the effect of linear control method and that of active predict control method by using neural network. It can be seen through the neural network predict control that structure displacement is under effective control, which shows that the effect of this kind of control method is remarkable.更多还原