支持系统级仿真与优化的流场模型降阶技术  

Fluid field model order reduction method for system-level simulation and optimization

作　　者：郝文涛[1] 田凌[1] 童秉枢[1] 

机构地区：[1]清华大学精密仪器与机械学系,北京100084

出　　处：《清华大学学报（自然科学版）》2007年第11期1965-1968,共4页Journal of Tsinghua University(Science and Technology)

基　　金：国家"八六三"高技术项目(2006AA04Z112)

摘　　要：为降低仿真对计算机软硬件系统的要求并提高计算速度,设计了基于降阶模型的系统级仿真与优化流程,并将本征正交分解(POD)方法应用到流固耦合仿真的流场降阶建模过程中,提高了仿真数据的重用度。在理论研究的基础上,对一个在空气中振动的微梁进行分析,并构建其降阶模型。结果表明:基于POD的降阶建模方法在保证计算精度的前提下,能有效减少流场自由度,加快流固耦合仿真的速度。该方法可用于加速产品系统级仿真和优化过程。A system-level simulation and optimization process was designed based on reduced-order models to reduce simulation costs and speed up computations. The proper orthogonal decomposition （POD） method was used to reduce the order of the fluid model in the fluid-structure interaction simulation, which reused much of the simulation data. A micro beam vibrating in air was simulated with a reduced-order model. The results show that the POD-based method effectiveiy reduces t,be number of degrees of freedom of the f;uid model, with the precision of the reduced-order model being still comparable with that of the full models. Therefore, the method can be used to speed up system-level simulations to optimize designs.

关 键 词：系统级仿真 流固耦合 降阶模型 本征正交分解 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]