复杂追踪识别结构参数及损伤

发布时间：2018-08-14 15:17

【摘要】：近年来国家对基础设施的投资规模不断扩大,各种大型复杂结构不断涌现。它们在服役期间,由于外界和自身因素的影响,不可避免地发生损伤。为了及早发现结构损伤,降低对人们生命及财产安全的威胁,一个有效的手段就是在大型结构上安装健康监测系统。作为结构健康监测系统的核心内容,结构参数识别和损伤识别一直是国内外学者的研究焦点,致力于寻找一种大型复杂结构在线监测的损伤诊断方法。本文将复杂追踪算法应用于土木工程结构模态参数和损伤识别中,为结构识别提供新思路,丰富和完善结构健康监测理论。本文主要研究内容及成果如下:(1)阐述了本课题的研究背景及意义,对当前结构模态参数识别和损伤识别的研究方法及存在的问题进行了系统论述,并简单介绍了盲源分离的发展及在结构识别中的应用。(2)简单介绍了复杂追踪的相关数学知识、基本原理、预处理方法,并对目标函数的建立以及梯度优化算法进行了论述。原梯度算法迭代过程使用固定步长和非线性函数,使得算法的自适应性不足。为此,根据峭度及最优步长对原梯度算法的迭代过程进行了改进,形成了适用性更广、分离效果更好的自适应复杂追踪算法。(3)发展出复杂追踪结合希尔伯特变换识别结构模态参数的新方法。对多自由度结构系统动力响应与复杂追踪模型之间的关系进行了探讨,通过复杂追踪算法分离出结构的各阶模态响应及振型,再利用希尔伯特变换解析出结构的模态频率及阻尼比。采用不同的模型验证了该方法准确识别结构模态参数的可靠性。(4)发展出复杂追踪结合集合经验模态分解识别结构损伤的新方法。结构损伤会引起响应信号的突变,通过集合经验模态分解提取包含损伤分量的高频本征模函数,再利用复杂追踪算法,从噪声信号中分离出该损伤分量。根据信号突变点的位置识别损伤发生的时间,再根据混合矩阵识别损伤发生的位置。通过不同的模型,验证了该方法可准确有效地识别结构损伤发生的时刻与位置。
[Abstract]:In recent years, the scale of national investment in infrastructure has been expanding, and various large and complex structures are emerging. During their service, injuries inevitably occur due to the influence of external and self-factors. In order to detect structural damage as early as possible and reduce the threat to people's life and property, an effective method is to install health monitoring system on large structures. As the core of structural health monitoring system, structural parameter identification and damage identification have been the research focus of scholars at home and abroad, dedicated to find a large and complex structure online monitoring damage diagnosis method. In this paper, the complex tracking algorithm is applied to the modal parameters and damage identification of civil engineering structures, which provides a new idea for structural identification and enriches and perfects the theory of structural health monitoring. The main contents and achievements of this paper are as follows: (1) the research background and significance of this subject are expounded, and the research methods and existing problems of structural modal parameter identification and damage identification are systematically discussed. The development of blind source separation and its application in structure recognition are briefly introduced. (2) the mathematical knowledge, basic principle, preprocessing method of complex tracing are briefly introduced, and the establishment of objective function and gradient optimization algorithm are discussed. The iterative process of the original gradient algorithm uses fixed step size and nonlinear function, which makes the algorithm self-adaptive. Therefore, the iterative process of the original gradient algorithm is improved according to the kurtosis and the optimal step size, resulting in a wider applicability. An adaptive complex tracking algorithm with better separation effect is proposed. (3) A new method of identifying structural modal parameters by complex tracing combined with Hilbert transform is developed. The relationship between dynamic response and complex tracking model of multi-degree-of-freedom structural system is discussed. The modal responses and modes of structures are separated by complex tracking algorithm. Then the modal frequency and damping ratio of the structure are analyzed by Hilbert transform. Different models are used to verify the reliability of the method. (4) A new method is developed to identify structural damage with complex tracing and set empirical mode decomposition. Structural damage can cause sudden change of response signal. The high frequency eigenmode function containing damage component is extracted by means of set empirical mode decomposition, and the damage component is separated from noise signal by complex tracking algorithm. The time of damage is identified according to the location of the signal mutation point, and then the location of the damage is identified according to the mixed matrix. Through different models, it is verified that the method can accurately and effectively identify the time and position of structural damage.
【学位授予单位】：苏州科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TU317

【参考文献】