由Biot模型描述的多孔材料的Optimization
Optimization of the porous material described by the Biot model
摘要:
本文研究了粘性流体饱和多孔局部周期性材料微结构的形状优化问题。在宏观层面上,多孔材料是由有效介质系数定义的Biot模型,包括排水骨架弹性,Biot应力耦合,Biot压缩系数,和达西流动模型的水力渗透率。通过均匀化,这些系数的计算使用的代表性单元格的特征响应组成的弹性固体骨架和粘性孔隙流体。为了达到最优化的目的,推导了问题在连续水平上的灵敏度分析。我们提供的目标函数的Biot模型系数相对于由B样条框,其中嵌入整个代表性的细胞描述的底层孔隙形状构成的灵敏度。我们认为材料设计问题的框架中,一个单一的代表性的细胞的布局进行了优化。然后,我们提出了一个顺序线性化方法的两个尺度的问题,其中局部微结构优化相对于宏观设计标准。数值实验报告,其中包括刚度最大化的限制,允许足够的渗透性,反之亦然。讨论了设计各向异性、样条盒参数化等问题。为了避免网格重构,采用了基于内射约束的几何正则化技术。
英文摘要:
The paper is devoted to the shape optimization of microstructures generating porous locally periodic materials saturated by viscous fluids. At the macroscopic level, the porous material is described by the Biot model defined in terms of the effective medium coefficients, involving the drained skeleton elasticity, the Biot stress coupling, the Biot compressibility coefficients, and by the hydraulic permeability of the Darcy flow model. By virtue of the homogenization, these coefficients are computed using characteristic responses of the representative unit cell consisting of an elastic solid skeleton and a viscous pore fluid. For the purpose of optimization, the sensitivity analysis on the continuous level of the problem is derived. We provide sensitivities of objective functions constituted by the Biot model coefficients with respect to the underlying pore shape described by a B-spline box which embeds the whole representative cell. We consider material design problems in the framework of which the layout of a single representative cell is optimized. Then we propose a sequential linearization approach to the two-scale problem in which local microstructures are optimized with respect to macroscopic design criteria. Numerical experiments are reported which include stiffness maximization with constraints allowing for a sufficient permeability, and vice versa. Issues of the design anisotropy, the spline box parametrization are discussed. In order to avoid remeshing a geometric regularization technique based on injectivity constraints is applied.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。