基于遗传算法的混凝土三维非稳态温度场反分析  

Inverse analysis of three dimensional unsteady temperature field with genetic algorithms

作　　者：张宇鑫 宋玉普[2] 王登刚[3] 

机构地区：[1]上海师范大学建筑工程学院,上海201418 [2]大连理工大学海岸与近海工程国家重点实验室,辽宁大连116024 [3]同济大学建筑工程学院,上海200092

出　　处：《计算力学学报》2004年第3期338-342,367,共6页Chinese Journal of Computational Mechanics

摘　　要：大体积混凝土结构施工期间的合理温度控制问题非常重要,而精确进行温度计算所需的一些材料参数往往不易直接测得,需要根据一些易测得的量进行反求。本文基于三维瞬态温度场有限元求解理论与反问题理论,建立了混凝土三维瞬态温度场反问题求解数值模型。运用遗传算法寻求非线性反演问题全局最优解,只需要若干点温度实测值便可实现混凝土多个热学参数如绝热温升、导温、导热系数及热交换系数等的同时反演,算例对本文反演方法的反演精度及数值稳定性给出了满意的证明。Suitable temperature control is important for the construction of mass concrete structures. Some material parameters for temperature computation cannot be acquired by direct measurement. According to some values easily to acquire, the inverse computation is used. By the finite element method, an inverse problem solution model for three-dimensional transient temperature field is established. By using the genetic algorithms to get the optimum solution of nonlinear inverse problem, the multi-thermal parameters, such as adiabatic temperature, temperature conduction, heat conduction and heat exchanging parameters can be obtained simultaneously with only temperature values of a few test points. Numerical examples show the high accuracy and numerical stability of the method.

关 键 词：温度 反问题 热学参数 有限元/遗传算法 

分 类 号：O39[理学—工程力学] TV315[理学—力学]