【作者】 卢平；

【导师】 许斌；

【作者基本信息】 湖南大学 ， 结构工程， 2008， 硕士

【摘要】 大量工业与民用建筑和社会基础设施在长期的服役过程中受到使用荷载及各种自然和人为因素的作用,不断出现损伤累积和功能退化,极端情况下甚至引发突发性灾难。对结构健康状况和损伤进行有效的评估已经成为国际土木工程界面临的一项紧迫课题,其中结构参数和损伤识别是其关键和核心问题之一。人工神经网络由于具有能以任意精度逼近任何线性和非线性函数关系的能力而在土木工程相关领域中得到了广泛的应用,包括结构参数识别和损伤评估。但以往基于神经网络和结构响应时间序列的结构损伤识别方法往往只能给出定性的结论而不能得出定量的结论,本文提出了一种基于神经网络和结构加速度响应时程的结构参数和损伤识别方法,并通过模型结构实验得到验证。另一方面,由于远程结构健康监测具有自动、实时、在线的特点,是对实际工程结构进行监测的重要课题。本文对湖南大学土木工程学院研究开发的网络结构实验室(Networked Structural Laboratory, NetSLab)的通讯平台进行升级,实现了结构动力响应测量时间序列远程传输,并与本文提出的结构参数和损伤识别方法相结合,为实现远程在线监测提供了一个新工具。本文首先回顾了传统的结构损伤检测和基于振动的损伤识别方法,并对人工神经网络在土木工程中的应用进行了总结。其次,提出了一种直接运用结构动力响应测量、基于神经网络的时域结构参数直接识别方法,并采用振动台实验的实测加速度响应对此方法进行验证。基于结构运动方程的离散解,对参数识别方法的理论基础以及两个神经网络的构建依据进行了阐明。为了对目标结构进行参数识别,首先假定一个参考结构并构建一个神经网络来描述参考结构的加速度响应时间序列之间的映射关系,即建立参考结构的非参数模型。然后,定义加速度响应预测差值均方根向量作为评价指标用于参数识别,并构建参数识别用神经网络来描述评价指标与结构参数之间的关系。编制了相关程序,运用一个框架结构模型振动台实验的加速度响应实测时间序列对模型的结构参数进行了识别。同时对此框架模型的损伤状态下的结构参数进行了识别。结果表明该方法具有较高的识别精度,识别结果可靠。最后,介绍了网络结构实验室的通讯平台,并进行了相关程序编制工作。对一个框架结构模型振动台实验的加速度响应时程,采用该通讯平台进行了远程数据传输试验。结果表明,该通讯平台能可靠的运行,具有实际应用的前景。更多还原

【Abstract】 Many civil infrastructures are now deteriorating due to aging, misuse, lacking proper maintenance, and, in some cases, overstressing as a result of increasing load demands and changing environments. Failure of these infrastructures often leads to a high social consequence. It is therefore critical to evaluate their current reliability, performance, and condition for the prevention of potential catastrophic events. Structural identification and damage detection have become an increasingly important research topic for health monitoring, performance assessment and safety evaluation of engineering structures.On one hand, because of the ability to approximate arbitrary continuous functions, neural networks have drawn considerable attention in civil engineering for identification in a non-parametric manner. However, because of the nonparametric characteristics, most of the proposed methods for structural health monitoring and damage detection with neural networks can just be used to give a qualitative indication or information that damage might be present in the structure, no quantitative identification can be determined. On the other hand, remote structural health monitoring system can provide abundant information for structural identification and damage detection due to its self-monitoring, long-term and on-line characteristics, and it has been applied in some important infrastructures. A networked structural laboratory (NetSLab) platform originally for networked remote collaborative test has been developed at Hunan University recently. The NetSLab platform provides a potential way for remote data transformation. In this study, modifications have been made on the developed NetSLab platform in order to realize structural dynamic response time series measurement transfer.Firstly, a comprehensive review on the traditional damage detection techniques, vibration-based global identification methodologies, and the application of artificial neural network in civil engineering are made. Secondly, a novel two neural networks based structural parameters identification methodology with the direct use of structural acceleration time series is proposed and validated by a shaking table test of a model structure. The rationality of the methodology is explained and the theory basis for the construction of the two neural networks is described according to the discrete time solution of the state space equation. An evaluation index called the root mean square of the acceleration prediction difference vector (RMSAPDV) is defined and employed to identify structural parameters. Based on the trained acceleration-based neural network modeling for a reference structure, and the parameter evaluation neural network that describes the relation between structural parameters and the components of the corresponding RMSAPDVs, the structural parameters of the model frame structure with known mass distribution are identified by the direct use of acceleration measurements. Results show that structural parameters can be identified with acceptable accuracy. The performance of the proposed methodology for a damaged model structure is also studied. Finally, after the updating of the NetSLab, a networked structural parameters identification system based on the novel identification methodology and the updated network platform is developed and test on remote time series file transfer is carried out. Results show that the updated network platform has the potential of becoming a practical tool for remote health monitoring of civil infrastructures.更多还原