改进分层遗传算法在斜拉桥主梁损伤识别中的应用

发布时间：2018-03-18 01:35

本文选题：斜拉桥　切入点：损伤识别　出处：《土木建筑与环境工程》2014年06期 　论文类型：期刊论文

【摘要】：标准遗传算法在解决像斜拉桥这类复杂结构的损伤识别问题时会出现提前收敛,即所谓"早熟"的现象。为了避免此现象的发生,提高损伤识别的效率与精度,提出一种基于改进分层遗传算法的斜拉桥主梁损伤识别方法。采用索力变化作为优化目标函数,将3种具有不同遗传算子的标准遗传算法与变量微调和灾变策略相结合,形成了一种具有灾变特性的分层遗传算法,以实验室独塔斜拉桥模型作为研究对象进行了数值仿真,结果表明:改进的分层遗传算法成功的避免了标准遗传算法"早熟"现象的发生,能快速有效的完成斜拉桥主梁各种损伤的识别;同时对此方法进行抗噪性分析发现,该方法具有良好的抗噪能力。
[Abstract]:In order to avoid this phenomenon and improve the efficiency and accuracy of damage identification, the standard genetic algorithm can solve the damage identification problem of complex structures such as cable-stayed bridges, which is called "premature convergence". A damage identification method for cable-stayed bridge girder based on improved hierarchical genetic algorithm is proposed. Using the variation of cable force as the optimization objective function, three standard genetic algorithms with different genetic operators are combined with variable fine-tuning and catastrophe strategy. A hierarchical genetic algorithm with catastrophic characteristics is developed, and the model of single-tower cable-stayed bridge in laboratory is used as a numerical simulation object. The results show that the improved hierarchical genetic algorithm can successfully avoid the occurrence of "premature" phenomenon of standard genetic algorithm, and can quickly and effectively identify all kinds of damage of main girder of cable-stayed bridge. This method has good anti-noise ability.
【作者单位】：石家庄铁道大学工程力学系;
【基金】：国家自然科学基金(50778116) 河北省自然科学基金(E2012210061) 河北省教育厅重点项目(ZH2012068)
【分类号】：U448.27

【参考文献】