基于时序模型残差分析的桥梁实时损伤预警方法

发布时间：2018-01-21 09:49

本文关键词： 桥梁损伤预警时序模型环境激励数值模拟模型试验　出处：《天津大学》2014年硕士论文　论文类型：学位论文

【摘要】：交通运输事业的发展和完善对国民经济的高速稳定增长起着重大的作用,而桥梁结构的正常运行是交通运输系统正常运转的前提。随着交通量的不断上涨,结构出现破损的现象也在增加,导致桥梁承载力下降,若不能及时识别损伤,使其累积到一定程度,一旦发生破坏,将给社会带来重大的损失。因此,对桥梁进行损伤预警研究,具有重要的理论意义和使用价值。当前国内已有的以及新建的大跨度桥梁的结构健康监测系统是通过获取结构实时监测信号分析结构系统特性的变化来发出损伤预警通知。本文提出一种基于时间序列模型分析的桥梁损伤预警方法。本文以桥梁结构实时损伤预警为主题对以下方面进行了研究:1、通过对已有的结构损伤识别方法和时间序列理论的分析和总结,提出了一种基于自回归模型和带有外部输入的自回归模型(Autoregressive models-Autoregressive models with eXogenous inputs,简称AR-ARX模型)的损伤预警方法,对响应信号进行标准化预处理,提出一种数据匹配的新方法,并对AR-ARX模型定阶方法进行了优化。通过模型残差的变化来识别结构损伤,并通过基于大量数据样本的统计评估指标进行损伤预警,对损伤预警效果进行了评估分析。2、通过对简支梁桥的数值模拟对该损伤预警方法的有效性进行了研究。对地面施加一定幅值的随机白噪声以模拟实桥的环境激励,研究了单一损伤和两处损伤情况下的损伤预警效果,并对不同噪声水平对预警效果的影响进行分析。分析结果表明该损伤预警方法可以有效识别损伤位置和损伤程度,且具有较强的抗干扰能力。3、加工制作两跨连续钢梁模型,对其进行试验研究,设置单一损伤和两处损伤。试验前先对连续梁进行了数值模拟分析。根据测试要求并结合连续梁模型确定加速度传感器布置方案,对地面进行持续随机激励模拟环境激励,采集结构在各种工况下的加速度信号,应用所提损伤预警方法进行分析处理,验证了所提方法的有效性。
[Abstract]:The development of transportation and improve the national economy plays an important role, and the normal operation of the bridge structure is the premise of the normal operation of the transportation system. With increasing traffic volume, structure damage phenomenon is also increasing, resulting in the decline of carrying capacity of bridge, if not timely identification of injury. The accumulated to a certain extent, once damaged, will bring a great loss to the society. Therefore, the damage of the bridge early warning research has important theoretical significance and practical value. The structural health monitoring system of current domestic existing and new construction of the long-span bridge is the change in the structure characteristics of the system by acquiring real-time monitoring of structure to give the signal damage warning notice. This paper proposes a bridge damage warning analysis method based on time series model. In this paper, the bridge structure real-time damage pre The police as the theme for the research on the following aspects: 1, through the analysis and summary of the structural damage identification method and time series theory, propose a method based on autoregressive model and autoregressive model with external input (Autoregressive models-Autoregressive models with eXogenous inputs, referred to as AR-ARX model) the damage alarming method, response to standard signal preprocessing, we propose a new method of data matching, and the AR-ARX model order determination method was optimized. Through the change of model residuals to identify the structural damage, and through statistical evaluation of large data samples based on damage warning, to damage alarming is evaluated by.2. Study on Numerical Simulation of the beam bridge effectiveness of the damage alarming method. Random white noise is applied to simulate the amplitude of ground bridge The study on environmental excitation, damage warning effect of a single injury and two damage cases, and the effects of different noise levels on early warning effect were analyzed. The results show that the damage alarming method can effectively identify the damage location and degree, and has strong anti-interference ability of.3, making two span continuous beam model and on its pilot study, set up a single damage and two damage. Before the tests of the continuous beam was simulated and analyzed. According to the test requirements and the continuous beam model to determine the acceleration sensor layout scheme, continuous random excitation simulation environment excitation on the ground, the acceleration signal in different conditions using the proposed acquisition structure. Damage alarming method for analysis, verify the effectiveness of the proposed method.

【学位授予单位】：天津大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U446

【引证文献】