帝国主义竞争算法的改进及其在结构识别中的应用
发布时间：2018-09-04 14:21
【摘要】：在役工程结构在恶劣自然环境、超负荷运营以及材料疲劳老化等多重因素影响下存在严重的安全隐患。传统的无损检测和可靠性评估方法需要知道结构缺陷的大致位置,且无法探测到结构内部缺陷,已经不能满足实用要求。为了弥补以上缺陷,近些年来该领域的学者们开展了诸多结构健康监测系统的研究。结构识别(结构参数识别和损伤识别)作为结构健康监测的核心内容,其识别方法繁多,但至今为止仍没有一个适用于实际工程的有效方法。本文对一种群体智能算法——帝国主义竞争算法(Imperialist Competitive Algorithm,ICA)进行了研究,针对算法的局限性进行改进,提出了寻优能力更强的改进算法,并将其引入到结构模态参数识别和损伤识别中。主要内容和研究及成果如下:(1)系统介绍了结构模态参数识别和损伤识别的研究现状、主要方法以及应用背景,由此提出了将ICA应用于结构识别的研究思路和主要研究内容;(2)详细介绍了有关于IC A的背景、基本原理以及计算流程等,同时简要阐述了ICA的主要应用领域;(3)针对ICA实际运用过程中会出现早熟收敛和陷入局部最优的缺陷,引入PSO算法中全局最优思想改进同化方程,并利用小波变异替代殖民地革命过程中的随机变异,发展出基于全局最优思想的改进帝国主义竞争算法——GBICA。通过标准测试函数测试结果可知,GBIC A相较于IC A在寻优精度、寻优能力和跳出局部最优值三方面能力上均有较大提升;(4)鉴于已知激励下的结构模态参数识别在实际应用中很难实现,本文研究环境激励下基于智能优化算法的模态参数识别,将参数识别问题转化为最优化问题,一次性识别出结构模态参数。通过数值模拟和实例分析可得:一方面,GBIC A相较于IC A在模态参数识别精度上有明显提高。另一方面,在不同噪声环境下,改进后的GBICA识别结果更加稳定,抗噪性能优于ICA;(5)由广义柔度灵敏度方法识别损伤的不足之处出发,利用广义柔度矩阵差构建目标函数,结合优化算法,将损伤识别问题转化为最优化问题。通过数值模型的损伤识别分析可得:一方面,无论是单处损伤还是多处损伤,ICA和GBICA两种方法均能快速、准确识别出结构的损伤位置和损伤程度。另一方面,相较于ICA,GBICA在损伤程度识别精度上明显提高,且在高噪声环境下,改进之后的GBICA稳定性明显优于ICA,说明GBICA具有更强的鲁棒性。
[Abstract]:Under the influence of many factors, such as harsh natural environment, overload operation and fatigue aging of materials, there are serious hidden dangers to the safety of in-service engineering structures. Traditional methods of nondestructive testing and reliability evaluation need to know the approximate location of structural defects and can not detect the internal defects of the structure, which can no longer meet the practical requirements. In recent years, scholars in this field have carried out a lot of research on structural health monitoring system. Structural identification (structural parameter identification and damage identification) as the core content of structural health monitoring, there are many identification methods, but up to now, there is still no effective method for practical engineering. In this paper, a colony intelligence algorithm-imperialist competition algorithm (Imperialist Competitive Algorithm,ICA) is studied. Aiming at the limitation of the algorithm, an improved algorithm with better searching ability is proposed. It is introduced into structural modal parameter identification and damage identification. The main contents and results are as follows: (1) the research status, main methods and application background of structural modal parameter identification and damage identification are introduced systematically. This paper puts forward the research idea and main research content of applying ICA to structure recognition. (2) the background, basic principle and calculation flow of IC A are introduced in detail. At the same time, the main application fields of ICA are briefly described. (3) aiming at the defects of premature convergence and falling into local optimum in the practical application of ICA, an improved assimilation equation is improved by introducing the global optimal idea in the PSO algorithm. By replacing the random variation in the colonial revolution with wavelet mutation, an improved imperialist competition algorithm based on the global optimal idea, GBICA, is developed. Through the test results of standard test function, we can see that the precision of IC A is better than that of IC A. (4) in view of the fact that structural modal parameter identification under known excitation is difficult to be realized in practical application, this paper studies modal parameter identification based on intelligent optimization algorithm under environment excitation. The parameter identification problem is transformed into an optimization problem, and the structural modal parameters are identified at one time. Through numerical simulation and example analysis, it can be concluded that, on the one hand, compared with IC A, the accuracy of modal parameter identification of GBICA is obviously improved. On the other hand, under different noise conditions, the improved GBICA recognition results are more stable, and the anti-noise performance is better than that of ICA; (5). Based on the generalized flexibility sensitivity method, the objective function is constructed by using the generalized flexibility matrix difference. Combined with the optimization algorithm, the damage identification problem is transformed into an optimization problem. Through the damage identification analysis of the numerical model, it can be concluded that, on the one hand, both single and multiple damage can be quickly identified by ICA and GBICA, and the damage location and damage degree can be accurately identified. On the other hand, compared with ICA,GBICA, the accuracy of damage recognition is obviously improved, and in high noise environment, the improved GBICA stability is obviously better than that of ICA, which shows that GBICA has stronger robustness.
【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
[Abstract]:Under the influence of many factors, such as harsh natural environment, overload operation and fatigue aging of materials, there are serious hidden dangers to the safety of in-service engineering structures. Traditional methods of nondestructive testing and reliability evaluation need to know the approximate location of structural defects and can not detect the internal defects of the structure, which can no longer meet the practical requirements. In recent years, scholars in this field have carried out a lot of research on structural health monitoring system. Structural identification (structural parameter identification and damage identification) as the core content of structural health monitoring, there are many identification methods, but up to now, there is still no effective method for practical engineering. In this paper, a colony intelligence algorithm-imperialist competition algorithm (Imperialist Competitive Algorithm,ICA) is studied. Aiming at the limitation of the algorithm, an improved algorithm with better searching ability is proposed. It is introduced into structural modal parameter identification and damage identification. The main contents and results are as follows: (1) the research status, main methods and application background of structural modal parameter identification and damage identification are introduced systematically. This paper puts forward the research idea and main research content of applying ICA to structure recognition. (2) the background, basic principle and calculation flow of IC A are introduced in detail. At the same time, the main application fields of ICA are briefly described. (3) aiming at the defects of premature convergence and falling into local optimum in the practical application of ICA, an improved assimilation equation is improved by introducing the global optimal idea in the PSO algorithm. By replacing the random variation in the colonial revolution with wavelet mutation, an improved imperialist competition algorithm based on the global optimal idea, GBICA, is developed. Through the test results of standard test function, we can see that the precision of IC A is better than that of IC A. (4) in view of the fact that structural modal parameter identification under known excitation is difficult to be realized in practical application, this paper studies modal parameter identification based on intelligent optimization algorithm under environment excitation. The parameter identification problem is transformed into an optimization problem, and the structural modal parameters are identified at one time. Through numerical simulation and example analysis, it can be concluded that, on the one hand, compared with IC A, the accuracy of modal parameter identification of GBICA is obviously improved. On the other hand, under different noise conditions, the improved GBICA recognition results are more stable, and the anti-noise performance is better than that of ICA; (5). Based on the generalized flexibility sensitivity method, the objective function is constructed by using the generalized flexibility matrix difference. Combined with the optimization algorithm, the damage identification problem is transformed into an optimization problem. Through the damage identification analysis of the numerical model, it can be concluded that, on the one hand, both single and multiple damage can be quickly identified by ICA and GBICA, and the damage location and damage degree can be accurately identified. On the other hand, compared with ICA,GBICA, the accuracy of damage recognition is obviously improved, and in high noise environment, the improved GBICA stability is obviously better than that of ICA, which shows that GBICA has stronger robustness.
【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】
相关期刊论文 前10条
1 常军;巩文龙;;量子粒子群结合小波变换识别结构模态参数[J];振动与冲击;2014年23期
2 常军;刘大山;;基于量子粒子群算法的结构模态参数识别[J];振动与冲击;2014年14期
3 张毅刚;刘才玮;吴金志;彭天明;;适用空间网格结构模态识别的改进功率谱峰值法[J];振动与冲击;2013年09期
4 常军;;在役桥梁结构损伤位置识别的综合指标方法研究[J];振动与冲击;2011年10期
5 狄生奎;张爱丽;汲生伟;;基于柔度矩阵的梁式结构损伤识别[J];兰州理工大学学报;2011年04期
6 朱宏平;余t，
本文编号：2222411
本文链接：https://www.wllwen.com/jianzhugongchenglunwen/2222411.html
