流固耦合问题的并行数值模拟算法与应用

项目名称： 流固耦合问题的并行数值模拟算法与应用

项目编号： No.91430215

项目类型： 重大研究计划

立项/批准年度： 2015

项目学科： 数理科学和化学

项目作者： 张林波

作者单位： 中国科学院数学与系统科学研究院

项目金额： 350万元

中文摘要： 本项目研究流固耦合问题的并行数值模拟算法及在血流动力学、人工心脏及水轮发电机中的应用，主要研究内容包括：（1）研究流固耦合ALE方法的稳定有限元离散格式和离散耦合非线性系统的高效可扩展求解算法，特别是非线性迭代中产生的线性方程组的高效预条件子；（2）研究流固耦合的全欧拉方法，设计基于XFEM的有限元离散方法和相应离散系统的高效可扩展求解方法；（3）发展非结构网格的流固耦合并行软件支撑框架，基于框架研制流固耦合并行有限元计算程序；（4）运用所发展的并行计算程序，开展以血流动力学、人工心脏及水轮机等为背景的流固耦合大规模并行数值模拟实验及应用。本项目的目标是发展流固耦合的先进计算方法和可扩展计算程序，为相关应用领域提供数值模拟工具，力求取得具有国际先进水平和高显示度的典型流固耦合问题的数值模拟结果。

中文关键词： 流固耦合；数值模拟；有限元；并行计算；血流动力学及人工心脏

英文摘要： In this project, we focus on the study of efficient algorithms and the development of parallel programs for numerical simulations of fluid-structure interactions (FSI), with emphasis on applications to hemodynamics, artificial heart，and hydro turbine. The main research subjects include: (1) Design stable FEM discretizations of FSI based on the Arbitrary Eulerian-Lagrangian (ALE) approach, develop efficient algorithms for solving the discrete coupled nonlinear systems of equations, and construct efficient preconditioners for solving the linear systems arising from the nonlinear iterations; (2) Study XFEM discretizations of FSI using the full Eulerian approach and algorithms for solving the discrete nonlinear systems of equations; (3) Develop a software framework, based on which we design parallel programs for numerical simulations of FSI on unstructured grids; (4) Carry out large-scale parallel numerical FSI simulations for several important applications. The objective of this project is three-fold: to develop advanced algorithms and scalable parallel programs for numerical simulations of FSI, to provide numerical tools for related application fields, and to perform state-of-the-art numerical simulations of typical FSI problems.

英文关键词： fluid-structure interaction;numerical simulation;finite element method;parallel computing;hemodynamics and artificial heart