东篱科研大数据发现系统（DRDS）

位置：成果数据库 > 期刊 > 期刊详情页

大型桥梁动力检测测点优化的改进遗传算法及其应用

期刊名称：西安建筑科技大学学报(自然科学版). 2006, 38(5): 625-628(EI)
时间：0
分类：TU317[建筑科学—结构工程]
作者机构：[1]哈尔滨工业大学土木工程学院,黑龙江哈尔滨150090
相关基金：国家自然科学基金资助项目（50278029）
相关项目：重大工程结构健康监测及其集成系统

关键词：大型桥梁, 动力检测, 测点优化, 遗传算法, 参数识别, long-span bridges , dynamic test , optimal sensor placement , genetic algorithm , parametric identification

中文摘要：

回顾了基于动力检测的传感器优化布置准则和方法，提出了一种应用改进遗传算法，服务于大型桥梁动力检测的测点优化方法，并将该方法具体应用到了哈尔滨四方台大桥的动力检测中。该算法改进了约束条件，对于传统遗传算法在大型结构应用时收敛慢且易陷入局部最优的缺陷进行了自适应和全面交叉改进，这种改进大大加快了收敛速度，并确保该算法能搜索到最优值．把经典的优化准则——有效独立准则，模态置信准则，模态应变能准则等以适应度的形式嵌入改进遗传算法中，得出各自的优化布置．通过对哈尔滨四方台大桥模型的仿真分析，证明改进的遗传算法在搜索能力、计算效率、可靠性等相对于传统遗传算法有较大的改善，搜索能力明显优于经典的序列法．在此基础上选取三种典型方法应用于哈尔滨四方台大桥的检测中，用实际采样得到的响应数据进行模态参数辨识，得出了该结构的振型，通过实际应用证明了上述方法的可行性。

英文摘要：

The research efforts to find best possible sensor locations in structural health monitoring field have received considerable attention recently for those attached to the results of modal parameter identification tightly. Based on the former methods and criteria of optimal sensor placement for modal test, a new method of optimum sensor placement using Improved Genetic Algorithm （IGA） for dynamic test is proposed and applied in detection of Harbin Sifangtai Bridge. The algorithm converges swiftly and has a strong ability to find the best value by improving the constraint condition, thoroughly crossing the auto mutation ratio. According to Effective Independence criterion （El）, Modal Assurance Criterion （MAC） and Modal Kinetic Energy method （MKE）, three placement designs are produced. The computational simulation of Harbin Sifangtai Bridge model using all the methods above has been implemented. The simulation results show that accuracy, reliability and efficiency of IGA are much better than traditional Genetic Algorithm and the searching ability of IGA is stronger than classical sequential sensor placement algorithm. Among those methods, three placement designs are adopted in detection of Harbin Sifangtai Bridge. Through the output signal, mode shapes of the bridge are identified by combined method of the NEXT and ERA to evaluate validity of the methods above.

同期刊论文项目