中山大学学报(自然科学版) ›› 2020, Vol. 59 ›› Issue (4): 79-88.doi: 10.13471/j.cnki.acta.snus.2019.06.20.2019B062

• • 上一篇    下一篇

基于改进稳态遗传算法的桥梁有限元模型修正 

秦世强, 张亚州, 康俊涛   

  1. 武汉理工大学土木工程与建筑学院,湖北 武汉 430070

  • 收稿日期:2019-06-20 出版日期:2020-07-25 发布日期:2020-07-25
  • 作者简介:秦世强(1987年生),男;研究方向:桥梁健康监测与模型修正;E-mail:shiqiangqin@whut.edu.cn

Bridge finite element model updating based on improved steady state genetic algorithm 

QIN Shiqiang, ZHANG Yazhou, KAGN Juntao   

  1. Swarm of Civil Engineering and ArchitectureWuhan University of TechnologyWuhan 430070China

  • Received:2019-06-20 Online:2020-07-25 Published:2020-07-25

摘要: 由于测试误差和结构参数的不确定性,有限元模型修正的局部最优解和全局最优解均有可能是真实解。为了同时获取模型修正的局部最优解和全局最优解,文章提出一种改进的稳态遗传算法(ISSGA)。该算法通过一种双角度算子来判定目标函数的可行解,并通过定义可行解的伴侣解不断优化解的位置,实现目标函数局部最优和全局最优解的寻找。通过两个测试函数和一座混凝土箱梁桥模型修正案例,验证了ISSGA算法的精度、稳定性和计算效率,并明确了算法中各个参数的取值依据。结果表明:ISSGA可同时获得目标函数的局部和全局最优解;双角度算子可有效避免局部最优解的遗漏;ISSGA算法为获得模型修正合理解提供了可能。

关键词: 桥梁工程, 模型修正, 稳态遗传算法, 优化问题, 多解问题

Abstract: Due to the test errors and the uncertainties of structural parameters, the best local and global  solutions are all possible to be the true solution of finite element model updating. To simultaneously obtain the local and global best solutions in model updating, an improved steady state genetic algorithm (ISSGA) is proposed. The proposed ISSGA utilizes a double angle operator to select alternative solutions of the objective function. The solutions are iteratively optimized by defining the solutions pairs, and the global and local best solutions are eventually obtained. Two benchmark functions and model updating of a concrete box girder bridge are employed to validate ISSGA regarding the accuracy, stability and calculation efficiency. In addition, the principle to determine the algorithm parameters is clarified. The results show that the ISSGA can obtain the global and the best local of the objective function at same time. The double angle operator can avoid the omission of the best local. ISSGA provides a possibility to obtain the solution of model updating wi the best justification.

Key words: bridge engineering, model updating, steady state genetic algorithm, optimization, multiple solutions problem

中图分类号: