Open AccessEfficient WENO library for OpenFOAM
Author(s) -
Jan Wilhelm Gärtner,
Andreas Kronenburg,
Tobías Martín
Publication year - 2020
Publication title -
softwarex
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.528
H-Index - 21
ISSN - 2352-7110
DOI - 10.1016/j.softx.2020.100611
Subject(s) - stencil , computer science , polygon mesh , reuse , scheme (mathematics) , reduction (mathematics) , computational science , boundary (topology) , stability (learning theory) , computational fluid dynamics , class (philosophy) , parallel computing , computer graphics (images) , mathematics , artificial intelligence , aerospace engineering , geometry , ecology , mathematical analysis , machine learning , engineering , biology
The weighted essentially non oscillating (WENO) concept is well established in research and its advantages are known, however, implementation details such as memory demand hindered the usage for general applications and general purpose libraries for many open source CFD tools do not yet exist. This paper introduces a WENO library for OpenFOAM and describes an efficient implementation. The large memory demand of the scheme is solved be reusing already calculated stencil sets and storing them in a data bank, giving a possible memory reduction of over 90%. A new class is implemented to allow the reconstruction of the stencil list on highly decomposed meshes by reconstructing a regional mesh around each processor. This avoids accuracy deficiency and stability problems at the processor boundary. Lastly, the performance of the implemented scheme is demonstrated by a standard Taylor Green Vortex test case.
Accelerating Research
Robert Robinson Avenue,
Oxford Science Park, Oxford
OX4 4GP, United Kingdom
Address
John Eccles HouseRobert Robinson Avenue,
Oxford Science Park, Oxford
OX4 4GP, United Kingdom