【作者】 刘名；

【导师】 肖仪清；

【作者基本信息】 哈尔滨工业大学 ， 结构工程， 2008， 硕士

【摘要】 随着科学技术和社会经济的快速发展,大跨度桥梁结构在工程中应用越来越多,而这些结构主要在复杂的环境中服役,另外还有材料的老化、交变荷载的疲劳效应等因素的耦合作用,这些都会造成结构的损伤,从而导致结构抗力的衰减,加大了结构发生事故的危险。当前,结构健康检测与损伤识别已经成为土木工程学科的热点研究领域,结构损伤识别问题就是根据结构表现出的一些性态去判别结构损伤的出现、位置和程度。本文针对大跨度桥梁结构,以山东滨州黄河大桥为背景,以数值仿真和实验为手段,研究基于结构柔度和模态应变能变化的损伤识别,以及神经网络方法的损伤识别。柔度法与模态应变能法是分别基于位移模态和应变模态的损伤识别方法,本文通过数值模拟分析说明,基于位移模态的柔度法只能识别出损伤的位置,但同时又具有一定的抗燥能力,基于应变模态的模态应变能法不仅能识别出损伤的位置还能识别出损伤的程度,是一种比较优秀的损伤识别方法。本文在此基础上引入BP神经网络,以结构模态应变能变化为样本训练了一个三次神经网络,并对损伤模拟的结果进行仿真,得到较好的结果。其次,本文在对滨州黄河大桥实验室缩尺模型的实例分析基础上,验证了这三种方法的可行性。最后给出了本文的结论和本课题未来研究工作的展望。更多还原

【Abstract】 With the fast development of technology and economy, more and more large-span bridges are used in engineering. However, these structures are mainly in service in the complex environment. The coupling action of material aging and fatigue effect of alternating loads can also cause structure damage, which will result in the decrease of structure bearing capacity and enhance the possibility of structure events. Currently structural health monitoring and damage identification have become the hot research field. Structural damage identification is defined that the commencement, the location and the degree of structural damage are identified according to the behavior of structure.This paper presents the damage identification of large span structure based on structural compliance and the change of modal strain energy as well as the one of neural network method. The results are all based on modeling and test under the background of the Shandong Binzhou Yellow River highway bridge. This paper is divided into four chapters as follows:Flexibility method and the modal strain energy method is damage identification methods based on each displacement mode and the strain mode, the paper adopted by the numerical simulation analysis, flexibility method can only identify the location of the damage ,which is based on displacement mode, but at the same time it has a certain capacity to resist noise; based on the strain mode, the modal strain energy method can not only identify the location of the damage but also identify the extent of the damage, and it is a relatively good damage identification methods. In this thesis, based on the introduction of BP neural network to structural changes in the modal strain energy for the training samples three times a neural network and simulation of damage simulation results, to achieve better result. Second, this article in the Binzhou Yellow River highway Bridge in the laboratory scale model based on the analysis of the case, to verify the feasibility of the three methods.Finally, this paper presents the conclusion and the prospect of research work in the future.更多还原