循环遗传聚类法稳定图自动分析

发布时间：2018-03-07 08:04

本文选题：自动识别　切入点：模糊聚类　出处：《浙江大学学报(工学版)》2017年03期 　论文类型：期刊论文

【摘要】：提出一种迭代的基于遗传算法和模糊聚类(IG-FC)技术的稳定图自动分析方法.该算法通过任意给定一个初始聚类数目,迭代使用遗传算法和模糊聚类技术搜索出代表真实模态的聚类中心,最终自动剔除稳定图中的虚假模态.一旦程序参数给定后,可在没有任何人为主观干扰的情况下自动识别出结构的真实模态.使用一个七自由度的弹簧质量数值模型以及加拿大某座钢筋混凝土框架结构大楼、瑞士Z24桥的加速度实测数据验证所提出算法的准确性.算例模态频率的识别结果表明:该算法可以自动且较准确地剔除稳定图上的虚假模态;可以自动判别并识别出相邻模态以及弱激励模态.
[Abstract]:An iterative automatic stability graph analysis method based on genetic algorithm (GA) and fuzzy clustering (IG-FC) technique is proposed. Genetic algorithm and fuzzy clustering technique are used to search the clustering center representing the real modal, and the false modal in the stability graph is eliminated automatically. Once the program parameters are given, The real mode of the structure can be automatically identified without any human interference. Using a seven-degree-of-freedom spring mass numerical model and a reinforced concrete frame structure building in Canada, The measured acceleration data of Z24 bridge in Switzerland verify the accuracy of the proposed algorithm. The results of modal frequency identification show that the proposed algorithm can automatically and accurately eliminate the false modes on the stability map. The adjacent modes and weak excitation modes can be identified and identified automatically.
【作者单位】：浙江大学建筑工程学院;
【基金】：国家“十二五”科技支撑计划资助项目(2012BAJ07B03)
【分类号】：TU317

【相似文献】