考虑环境因素影响的结构健康监测损伤识别

发布时间：2018-03-31 01:29

本文选题：结构健康监测　切入点：损伤识别　出处：《大连理工大学》2013年硕士论文

【摘要】：土木工程结构安全关系到人民生命和财产安全,为早期发现结构损伤或退化,最大化结构使用周期并减小寿命周期成本,确保结构的安全性和可靠性,在结构服役期间进行在线结构健康监测是必要的,而结构损伤识别算法是结构健康监测的核心所在。近三十年来,大量的损伤识别算法不断涌现,但由于实际工程的复杂性,使得其在工程应用中受到极大的限制。近年来,考虑环境因素和运行状态的识别算法,成为结构健康监测领域的研究热点。将环境因素和运行条件考虑到损伤识别方法中,使得损伤识别算法向工程应用领域迈进了一大步,本文针对结构损伤识别问题做了以下几方面工作： (1)阐述了结构健康监测技术在土木工程领域发展的重要地位,综述了结构健康监测领域中常见的损伤识别方法,详细分析了基于振动的损伤识别方法的优缺点。针对各种环境因素和结构运行状态,总结了其对结构参数不利的影响,进而对近年来形成的两种主要考虑环境因素和运行状态的损伤识别方法做一综述。 (2)应用单元模态应变能法于输电铁塔损伤识别中,单元模态应变能利用结构振型计算,而结构振型是受环境因素和运行状态影响相对较小的一种结构参数。详细地推导了单元模态应变能法,并在输电塔截断模型的有限元模型和试验模型中对该方法进行了测试。 (3)将盲源分离算法首次应用于考虑环境因素的损伤识别算法中,详细地介绍了盲源分离模型和典型算法(AMUSE法和FastICA法)。应用盲源分离方法对识别地结构的频率进行信号源分离,认为改变结构频率有环境源和损伤源,通过分离,可以去除环境源对频率信号的改变,进而能够突出结构损伤对频率的改变,使得结构损伤识别效率大大提高,从而减少误判和漏判。 (4)提出了一种基于协整的考虑环境和运行变量的损伤识别方法。将结构邻近位移序列应用于协整中,提出协整对法去定位损伤,产生的残差序列用于损伤预警,当该残差超过了控制线,表示结构可能有损伤发生,通过各协整对余量跳跃的幅值和方向可以判断损伤发生的位置,模拟中该方法在环境因素影响下取得了良好的识别结果。当考虑非线性环境因素时,残差表现出非稳态特性,可进一步通过自相关过滤算法加以改善。
[Abstract]:The safety of civil engineering structures is related to the safety of people's lives and property. In order to detect the damage or degradation of structures early, maximize the life cycle of structures and reduce the cost of life cycle, the safety and reliability of structures can be ensured. It is necessary to carry out on-line structural health monitoring during the period of structural service, and structural damage identification algorithm is the core of structural health monitoring. In the past 30 years, a large number of damage identification algorithms have been emerging, but due to the complexity of practical engineering, many damage detection algorithms have been developed. In recent years, the recognition algorithm considering environmental factors and running state has become a research hotspot in the field of structural health monitoring. Environmental factors and operation conditions are taken into account in damage identification methods. The damage detection algorithm is a big step forward to the engineering application field. In this paper, the following work has been done on the structural damage identification problem:. This paper expounds the important position of structural health monitoring technology in the field of civil engineering, and summarizes the common damage identification methods in the field of structural health monitoring. The advantages and disadvantages of damage identification method based on vibration are analyzed in detail. Then two damage identification methods which mainly consider environmental factors and operating state are reviewed. (2) the element modal strain energy method is applied to the damage identification of transmission tower, and the element modal strain energy is used to calculate the mode of the structure. The modal shape of the structure is a kind of structural parameter which is less affected by the environmental factors and the running state. The element mode strain energy method is derived in detail and tested in the finite element model and the experimental model of the truncation model of transmission tower. (3) Blind source separation algorithm is first applied to damage identification algorithm considering environmental factors. The blind source separation model and typical algorithms, such as AMUSE method and FastICA method, are introduced in detail. The blind source separation method is used to separate the frequency of the identified ground structure. It is considered that changing the frequency of the structure has environmental source and damage source. By separating, the change of frequency signal caused by environmental source can be removed, and then the change of frequency of structural damage can be highlighted, and the efficiency of structural damage identification can be greatly improved. So as to reduce miscarriage of justice and omission. In this paper, a method of damage identification based on cointegration considering environment and running variables is proposed. The adjacent displacement sequence of structure is applied to cointegration, and the method of co-integration pair is proposed to locate the damage. The residual sequence is used for early warning of damage. When the residual error exceeds the control line, it indicates that damage may occur in the structure, and the location of the damage can be determined by the amplitude and direction of each cointegration pair margin jump. Under the influence of environmental factors, the method achieves good recognition results. When nonlinear environmental factors are considered, the residuals show unsteady-state characteristics, which can be further improved by autocorrelation filtering algorithm.
【学位授予单位】：大连理工大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TU317

【参考文献】