A UNIFIED MULTILEVEL FRAMEWORK OF UPSCALING AND DOMAIN DECOMPOSITION

被引:0
|
作者
Sandvin, Andreas [1 ]
Nordbotten, Jan M. [1 ]
Aavatsmark, Ivar [1 ]
机构
[1] CIPR, N-5007 Bergen, Norway
关键词
Multiscale; preconditioner; porous media; FINITE-VOLUME METHOD;
D O I
暂无
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
We consider multiscale preconditioners for a class of mass-conservative domain-decomposition (MCDD) methods. For the application of reservoir simulation, we need to solve large linear systems, arising from finite-volume discretisations of elliptic PDEs with highly variable coefficients. We introduce an algebraic framework, based on probing, for constructing mass-conservative operators on a multiple of coarse scales. These operators may further be applied as coarse spaces for additive Schwarz preconditioners. By applying different local approximations to the Schur complement system based on a careful choice of probing vectors, we show how the MCDD preconditioners can be both efficient preconditioners for iterative methods or accurate upscaling techniques for the heterogeneous elliptic problem. Our results show that the probing technique yield better approximation properties compared with the reduced boundary condition commonly applied with multiscale methods.
引用
收藏
页码:1060 / 1067
页数:8
相关论文
共 50 条
  • [31] A Unified Framework for Distance-Aware Domain Adaptation
    Wang, Fei
    Ding, Youdong
    Liang, Huan
    Gao, Yuzhen
    Che, Wenqi
    2020 25TH INTERNATIONAL CONFERENCE ON PATTERN RECOGNITION (ICPR), 2021, : 1796 - 1803
  • [32] A unified framework for visual domain adaptation with covariance matching
    Karn, Ravi Ranjan Prasad
    Sanodiya, Rakesh Kumar
    Bajpai, Priyam
    KNOWLEDGE-BASED SYSTEMS, 2023, 278
  • [33] Unified Optimal Transport Framework for Universal Domain Adaptation
    Chang, Wanxing
    Shi, Ye
    Tuan, Hoang Duong
    Wang, Jingya
    ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 35 (NEURIPS 2022), 2022,
  • [34] A unified framework for cross-domain sentiment classification
    Wu, Qiong
    Liu, Yue
    Shen, Huawei
    Zhang, Jin
    Xu, Hongbo
    Cheng, Xueqi
    Jisuanji Yanjiu yu Fazhan/Computer Research and Development, 2013, 50 (08): : 1683 - 1689
  • [35] A Domain Decomposition Multilevel Preconditioner for Interpolation with Radial Basis Functions
    Haase, Gundolf
    Martin, Dirk
    Schiffmann, Patrick
    Offner, Guenter
    LARGE-SCALE SCIENTIFIC COMPUTING, LSSC 2017, 2018, 10665 : 499 - 506
  • [36] BEYOND AUTOMATED MULTILEVEL SUBSTRUCTURING: DOMAIN DECOMPOSITION WITH RATIONAL FILTERING
    Kalantzis, Vassilis
    Xi, Yuanzhe
    Saad, Yousef
    SIAM JOURNAL ON SCIENTIFIC COMPUTING, 2018, 40 (04): : C477 - C502
  • [37] A Multilevel Domain Decomposition Approach for Studying Coupled Flow Applications
    Aulisa, E.
    Cervone, A.
    Manservisi, S.
    Seshaiyer, P.
    COMMUNICATIONS IN COMPUTATIONAL PHYSICS, 2009, 6 (02) : 319 - 341
  • [38] A Multilevel Domain Decomposition Solver Suited to Nonsmooth Mechanical Problems
    Iceta, Damien
    Alart, Pierre
    Dureisseix, David
    DOMAIN DECOMPOSITION METHODS IN SCIENCE AND ENGINEERING XVIII, 2009, 70 : 113 - 120
  • [39] A DOMAIN DECOMPOSITION PRECONDITIONER BASED ON A CHANGE TO A MULTILEVEL NODAL BASIS
    TONG, CH
    CHAN, TF
    KUO, CCJ
    SIAM JOURNAL ON SCIENTIFIC AND STATISTICAL COMPUTING, 1991, 12 (06): : 1486 - 1495
  • [40] A unified approach to Boolean function decomposition for multilevel FPGA-based design
    Boole, E
    Boule, K
    Chapenko, V
    PROGRAMMABLE DEVICES AND SYSTEMS 2001, 2002, : 215 - 219