基于小波—人工免疫算法的框架结构损伤识别研究

发布时间：2018-04-20 16:25

本文选题：损伤识别 + 小波分析　；参考：《长沙理工大学》2015年硕士论文

【摘要】：在建筑结构中框架结构是运用比较普遍的一种结构形式,因受自然或人为因素的影响,结构可能会出现损伤,对结构的安全构成威胁,当损伤积累到一定程度,甚至可能酿成重大的工程事故。故对框架结构损伤识别方法的研究意义重大。小波分析作为一种新的信号处理手段,它能在时频两域很好的表征信号的局部特征,能较好的分析信号的奇异点位置,因而小波理论已被应用到结构的损伤诊断中;人工免疫算法是模仿生物免疫系统的运行机制而建立的一种学习、优化的人工智能算法,人工免疫算法具有优越的全局搜索能力,可进行结构损伤程度的识别。本文将这两者的优点进行有机的结合,提出了小波-人工免疫算法的概念,建立了一种既能识别结构损伤位置又能确定损伤程度的小波-人工免疫算法,即通过将小波分析与人工免疫算法相结合的方法实现对框架结构的损伤位置和损伤程度的识别。以含损伤的框架结构为研究对象,采用本文提出的小波-人工免疫算法,建立框架结构有限元模型,采用Lanczos法对结构进行模态分析,得到含损伤框架结构的模态参数,运用小波分析方法对结构的模态参数进行连续小波变换得到小波系数图,通过图形中的奇异点来识别框架结构损伤的位置。然后,以基于结构的频率和模态振型建立的目标函数为抗原,以小波识别出的损伤单元的损伤程度(即问题的解)作为抗体,运用人工免疫算法进行全局寻优的迭代计算,进而实现对结构的损伤程度的识别。本文建立了一层一跨多损伤(两处损伤和三处损伤)和一层两跨多损伤(三处损伤和四处损伤)框架结构的有限元模型,设置多种工况进行数值模拟分析,运用小波分析方法,确定结构损伤的位置。在确定了结构损伤单元位置和损伤单元个数的基础上运用人工免疫算法实现对结构损伤程度的识别。本文在简单的框架结构基础上,将上述方法应用到较复杂的两层一跨多损伤(三处损伤和四处损伤)、两层两跨多损伤(两处损伤和三处损伤)的框架结构上,验证了该方法的有效性。本文提出的方法可供结构损伤诊断的工程应用参考。
[Abstract]:In building structure, frame structure is a kind of structure which is widely used. Due to the influence of natural or human factors, the structure may be damaged, which will threaten the safety of the structure, and when the damage accumulates to a certain extent, It could even lead to major engineering accidents. Therefore, it is of great significance to study the damage identification method of frame structure. Wavelet analysis, as a new signal processing method, can well characterize the local characteristics of signals in time-frequency domain and analyze the location of singularities of signals, so wavelet theory has been applied to the damage diagnosis of structures. Artificial immune algorithm (AIA) is a kind of learning and optimization artificial intelligence algorithm which imitates the operating mechanism of biological immune system. Artificial immune algorithm has superior global searching ability and can be used to identify the degree of structural damage. In this paper, the advantages of these two methods are combined organically, the concept of wavelet artificial immune algorithm is put forward, and a wavelet artificial immune algorithm is established, which can not only identify the location of structural damage but also determine the degree of damage. In other words, wavelet analysis and artificial immune algorithm are combined to identify the damage location and damage degree of frame structure. Taking the frame structure with damage as the research object, the finite element model of the frame structure is established by using the wavelet artificial immune algorithm proposed in this paper, and the modal parameters of the frame structure with damage are obtained by Lanczos method. Wavelet transform is applied to the modal parameters of the structure to obtain the wavelet coefficients. The damage location of the frame structure is identified by the singular points in the graph. Then, the target function based on the frequency and modal mode of the structure is taken as antigen, the damage degree of the damage unit identified by wavelet is taken as the antibody, and the iterative calculation of global optimization is carried out by using artificial immune algorithm. Then the damage degree of the structure can be identified. In this paper, the finite element model of one story, one span, multiple damage (two damage and three damage) and one layer two span multiple damage (three damage and four damage) is established. Determine the location of structural damage. On the basis of determining the location and the number of damage elements, the artificial immune algorithm is used to identify the damage degree of the structure. On the basis of simple frame structure, this paper applies the above method to the complicated frame structure with two layers and one span multiple damage (three damage and four damage), two story two span multiple damage (two damage and three damage). The effectiveness of the method is verified. The method presented in this paper can be used as a reference for the engineering application of structural damage diagnosis.
【学位授予单位】：长沙理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TU317

【相似文献】