基于非支配排序遗传算法的塔机有限元模型修正

摘要 为了建立一个能准确反映结构实际状态的有限元模型，提出了一种基于非支配排序遗传算法Ⅱ(NSGA-Ⅱ)的有限元模型修正方法.首先建立初始有限元模型，基于二次响应面法，得到有效的响应面替代模型，然后采用NSGA-Ⅱ对该模型进行修正，最终建立了满足工程精度要求的可靠的有限元模型.给出了某型塔机有限元模型修正的工程算例，将修正后的计算结果与实测数据相比较，说明了基于NSGA-Ⅱ多目标优化算法对于有限元模型修正具有理想的效果，修正后的有限元模型能准确反映结构力学特性.
Abstract：A finite element model updating method based on the non-dominated sorting genetic algorithm Ⅱ (NSGA-Ⅱ) was proposed， which ensures that the established finite element model can accurately reflect the actual state of the structure. Firstly， the finite element model was established， and the effective response surface substitution model was obtained based on the quadratic polynomial response surface method. Then the response surface model was updated with NSGA-Ⅱ. Finally a reliable finite element model was obtained which can satisfy the requirements of engineering precision. An engineering example of a tower crane’s finite element model updating was provided. In accordance with the measured data， the results indicated that the multi-objective optimization algorithm based on NSGA-Ⅱ has an ideal effect for the finite element model updating， and the updated finite element model can accurately reflect the mechanical properties of the structure.