复合材料加筋板剪切后屈曲分析与优化设计

Optimization design and analysis of stiffened composite panels in post-buckling under shear

为了充分利用复合材料加筋板的后屈曲承载能力,针对复合材料加筋板的后屈曲行为开展优化设计方法的研究具有重要意义。详细探讨了筋条尺寸及密度等参数对承受面内剪切载荷作用下的复合材料双向加筋板屈曲后屈曲的影响规律。建立了复合材料加筋板考虑后屈曲响应的结构分级优化方法:在一级优化中以结构几何尺寸为设计变量,使用响应面法(RSM)拟合出结构后屈曲响应的全局近似函数,结果显示,加筋缘条的宽度及加筋的密度对屈曲承载能力有重要影响;在二级优化中采用遗传算法(GA)对复合材料铺层顺序进行优化,经过两级优化后的复合材料加筋板相比于初始设计在质量减少了3%的同时,线性屈曲位移提高了8.86倍,线性屈曲模态由局部屈曲改善为整体屈曲,同时结构的后屈曲承载能力提高了8.7%。基于解决旅行商问题(TSP)的遗传算法被调整用于固定铺层厚度的复合材料铺层顺序优化问题,经优化,结构线性屈曲特征值提高了12.76%,表明了优化方法的可行性。

In order to make full use of post-buckling load-bearing capacity of stiffened composite panel,it is meaningful to carry out research about the optimization strategy of stiffened composite panel especially in terms of its post-buckling behavior.The geometry parameter＇s influence on the buckling and post-buckling behavior of a bidirectional stiffened composite panel under in-plane shear is discussed in detail in this paper.Multi-level optimization including post-buckling capacity for composite stiffened panels under in-plane shear is presented in this article.In the 1st stage optimization,the geometry parameters are chosen as the design variables and the response surface method（RSM）is utilized to establish the global approximate function of the structure＇s post-buckling response.Results showed that the width and density of the stiffeners have important influence on the buckling bearing capacity.In the 2nd stage optimization,genetic algorithm（GA）is applied to optimize the stack sequence of the composite plies.The optimized structure shows a level of buckling displacement 8.86 times higher than that observed in the original panel and the mass decreased by 3% at the same time.The buckling mode transforms from early local buckling to global buckling,and the ultimate load capacity has also improved by 8.7% at the same time.Genetic algorithm dealing with the traveling salesman problem（TSP）is modified to solve the stack sequence optimization problem with fixed thickness.The optimized buckling eigenvalue has improved by 12.76%,which proves the validity and feasibility of the algorithm.