基于模型修正与时序分析的结构损伤识别方法研究

发布时间：2018-06-22 21:43

本文选题：结构健康监测 + 结构损伤识别　；参考：《兰州理工大学》2013年博士论文

【摘要】：随着世界经济与科学技术的快速发展,现代结构设计不断呈现出大型化、复杂化、多样化的趋势,而这些结构设计使用寿命较长、影响力较大,一旦失事,将会造成严重的生命财产损失。因此,为了保障结构的安全性、完整性与耐久性,在许多新建的大型结构和基础设施上增设了长期结构健康监测系统,对结构状态进行实时监测,为实现结构状态评估提供依据。结构损伤识别(Structural damage identification)作为结构健康监测技术的核心,对掌握结构工作状态以及评估结构安全性具有重要的意义。尽管在过去的20年内结构损伤识别得到了广泛的研究,但离实际工程应用还有一定的距离,还需要进行深入研究。基于此,本文通过模型修正与时序分析方法对结构损伤识别进行系统研究,主要取得以下几个方面的研究成果： (1)详细介绍了新型智能优化算法——和声搜索算法(Harmony Search Algorithm,缩写为HSA)的基本原理以及现有比较常见的改进方法。针对改进的HSA在搜索范围较大时,收敛速度慢且容易陷入局部最优的问题,提出了一种快速和声搜索算法(Fast Harmony Search Algorithm,缩写为FHSA),该算法既保证了最优解的集中性(intensification),也增加了寻求最优解时的多样性(diversification),最后给出了FHSA的适用性建议； (2)HSA作为一种新近提出的启发式智能优化算法,在许多领域得到应用,但在结构健康监测领域则鲜有见到。本文将HSA引入到结构健康监测领域,并成功应用于结构有限元模型修正。在分析中,将结构有限元模型修正问题转换为非线性优化问题,以模态参数的残差作为目标函数,分别对模态参数为确定性量与不确定性量两种情况进行分析。但在分析中发现,FHSA无法给出选择随机数产生的分布准则,因此对FHSA进行了改进,提出了一种改进的快速和声搜索算法(Improved Fast Harmony Search Algorithm,缩写为IFHSA).通过数值算例验证了HSA在有限元模型修正中的可行性,并将其应用于广州塔benchmark中的缩减模型修正。这不仅拓宽了HSA的应用领域,也为有限元模型修正提供了另一种思路； (3)将HSA成功应用于结构损伤识别。当采用模型修正方法对结构损伤进行识别时,可以将结构损伤识别转换为优化问题。在分析中,将频率误差、模态置信度以及残余力向量(residual force vector)以不同权重值组合从而建立目标函数,利用IFHSA逐步实现对结构参数模型的修正,进而对结构损伤进行识别。通过算例对该方法在无噪声干扰与有噪声干扰情况下的识别结果进行了分析。分析结果表明,该方法可以用于结构损伤识别,并且具有一定的抗噪性； (4)时间序列模型作为信息的凝聚器,可用于建立反映结构特性的参数模型。通过对时间序列模型进行理论分析,提出了一种基于时间序列预测模型的结构损伤识别方法。在分析中,首先对结构响应数据进行预处理,建立结构在完好工况下参考自回归(Auto-regression,缩写为AR)预测模型,利用参考AR预测模型计算待识别工况的残差,通过定义待识别工况与参考AR预测模型的残差的方差之比作为损伤指标,对结构损伤进行识别。通过算例分析表明,该损伤指标不仅可以判断结构是否发生损伤,而且还可以识别结构损伤位置； (5)基于传递函数可以反映结构输入输出相互关系(结构特性)的思想,提出了一种基于ARX (Auto Regressive model with eXogenous input,缩写为ARX)模型建立伪传递函数(Pseudo-Transfer Function,缩写为PTF)的损伤定位方法,并结合自然激励技术对其进行改进,使其可用于环境激励下结构的损伤识别。通过数值模拟与6层剪切型框架模型试验,验证了该方法在损伤识别中的有效性。最后将其应用于广州塔缩减模型的损伤识别,并对该方法在高耸结构损伤识别中应用进行了研究。
[Abstract]:With the rapid development of the world economy and science and technology, modern structural design has been showing a trend of large, complicated and diversified, and these structures have long life and great influence. If they fail, they will cause serious loss of life and property. Therefore, in order to ensure the safety, integrity and durability of the structure, there are many A long-term structural health monitoring system is added to the new large structure and infrastructure to monitor the state of the structure in real time. The structure damage identification (Structural damage identification), as the core of the structural health monitoring technology, is the core of the structural health monitoring and the evaluation of structural security. It is of great significance. Although structural damage identification has been widely studied in the past 20 years, it has a certain distance from the practical engineering application and needs further research. Based on this, this paper systematically studies the structural damage identification by model modification and time series analysis method, and mainly studies the following aspects. Achievements:
(1) the basic principle of the new intelligent optimization algorithm (Harmony Search Algorithm, abbreviated HSA) and the common improvement methods are introduced in detail. In this paper, a fast harmonic search algorithm (Fast) is proposed to solve the problem that the improved HSA has a slow convergence speed and easy to fall into the best local problem when the search range is large. Harmony Search Algorithm, abbreviated as FHSA), the algorithm not only guarantees the centrality of the optimal solution (intensification), but also increases the diversity of the optimal solution (diversification). Finally, the applicability of the FHSA is given.
(2) HSA, as a newly proposed heuristic intelligent optimization algorithm, has been applied in many fields, but rarely seen in the field of structural health monitoring. In this paper, HSA is introduced into the field of structural health monitoring and is successfully applied to the modification of structural finite element model. In the analysis, the problem of structural finite element model correction is converted to nonlinear optimization. The model parameter is used as the objective function, and the modal parameters are analyzed in two cases, which are deterministic and uncertain. However, in the analysis, it is found that FHSA can not give the distribution criterion of selecting random numbers. Therefore, FHSA is improved and an improved fast harmonic search algorithm (Improved Fast) is proposed. Harmony Search Algorithm, abbreviated as IFHSA). Through numerical examples, the feasibility of HSA in the finite element model correction is verified and applied to the revision of the reduced model in the Canton Tower benchmark. This not only widens the application field of the HSA, but also provides another way of thinking for the modification of the finite element model.
(3) HSA is successfully applied to structural damage identification. When the structural damage is identified by the model correction method, structural damage identification can be converted into an optimization problem. In the analysis, the frequency error, the modal confidence and the residual force vector (residual force vector) are combined with different weights to establish the target function, and the IFHSA is used in the analysis. The structural damage is identified by the gradual realization of the structural parameter model, and the results of the method are analyzed in the case of no noise interference and noise interference. The analysis results show that the method can be used for structural damage identification and has certain noise resistance.
(4) as a coagulator of information, the time series model can be used to establish the parameter model that reflects the structural characteristics. By theoretical analysis of the time series model, a structural damage identification method based on time series prediction model is proposed. In the analysis, the structural response data is pretreated first, and the structure is built in perfect condition. The prediction model of Auto-regression (AR) is used to calculate the residual error of the working condition by the reference AR prediction model, and the structural damage is identified by defining the variance ratio of the variance of the residual difference between the identified and the reference AR prediction models. Whether the structure is damaged or not, it can also identify the location of structural damage.
(5) a method of damage location based on ARX (Auto Regressive model with eXogenous input and abbreviated ARX) model based on the transfer function which can reflect the relationship between the input and output of the structure is proposed, which is modified with the natural excitation technology and modified with the natural excitation technology. It can be used for damage identification of structures under environmental excitation. Through numerical simulation and 6 layer shear frame model test, the effectiveness of the method in damage identification is verified. Finally, the method is applied to the damage identification of the Canton Tower reduction model, and the application of this method to the identification of the damage of the towering structure is studied.
【学位授予单位】：兰州理工大学
【学位级别】：博士
【学位授予年份】：2013
【分类号】：TU317

【相似文献】