遗传算法用于结构物内部缺陷识别的逆分析  

Inverse Analysis Estimating Inner Defects of Structure Using Genetic Algorithm

作　　者：邵金林[1,2] 沈成武[1,2] 唐小兵[1,2] 沈延春[1,2] 

机构地区：[1]武汉交通科技大学土木建筑工程系 [2]武汉职工医学院医疗系

出　　处：《武汉交通科技大学学报》1997年第3期275-280,共6页Journal of Wuhan University of Technology(Transportation Science & Engineering)

摘　　要：尝试用遗传算法与有限元结合，对悬臂梁的内部空洞进行识别，得到满意的逆分析结果．由于缺陷识别问题可以转化为有约束的目标函数极小化问题，表明遗传算法可用于优化计算．由于遗传因子用二进制形式表示，以及染色体的遗传和自然淘汰等特点，使其对离散型参数的多峰优化问题尤为有效．Crossover,mutation and selection of genetic algorithm are indroduced in detail.Inner voids of a cantilever beam are estimated using combination of genetic algorithm and finite element method.This estimation can be transtated into global optimal calculation with a multi peak value and separated parameters.It is difficult to obtain the global optimum using traditional optimal algorithm because of the discontinuity of the estimate space in this problem.By contrast,the genetic algorithm based on the principle of genetics and natural selection is efficient.Satisfied results for the estimation are obtained.

关 键 词：遗传算法 逆分析 缺陷 识别 悬臂梁 有限元 

分 类 号：TB11[理学—数学]