Computational Technology Resources - CCP

Keywords: domain decomposition, scalable algorithms, parallel algorithms, contact problems.

Summary

Since 1992, the research at the Department of Applied Mathematics of the VŠB-Technical University of Ostrava has been focused to development of scalable algorithms for contact problems and contact shape optimization problems. These algorithms are based on FETI domain decomposition methods which are well known for their parallel and numerical scalability. Our algorithms were originally implemented using a C++ library called OOSol (Object Oriented SOLvers) [1]. This library benefits from modularity and extensibility, nevertheless the slow development of the code caused by the lack of availability of advanced auxiliary algorithms necessary for debugging of complex algorithms, was its biggest disadvantage. Therefore we started to implement simultaneously all these algorithms into a new library that is developed in The Mathworks Matlab environment [2] which is equipped with many of these helpful functions. We call this library MatSol (MATlab SOLvers) [3]. Several years ago the Mathworks company introduced the Matlab Distributed Computing Engine which allowed Matlab functions to run on parallel computers. Hence, the MatSol has the full functionality of the OOSol library including the parallel algorithms and recently represents our primary testing and developing library. At the present time the library already contains many tools for the solution of realistic engineering problems such as problems in structural mechanics, electromagnetics, acoustics etc.

In our paper, we present the functionality of the MatSol library to the solution of realistic contact problems with millions of degrees of freedom showing parallel and numerical scalability of the implemented total FETI method [4,5]. The interface between MatSol and ANSYS or COMSOL [6,7] is presented as well. This feature allows to simply plug the MatSol library into the commercial finite element packages. Some comparisons of the commercial solver ANSYS and MatSol is shown. The efficiency of the solving algorithms will be presented on such complex problems as composite modelling or contact shape optimization problems.

References

1: OOSol library, http://www.am.vsb.cz/oosol
2: MATLAB - The Language of Technical Computing, http://www.mathworks.com
3: MatSol library, http://www.am.vsb.cz/matsol
4: Z. Dostál, "Optimal Quadratic Programming Algorithms, with Applications to Variational Inequalities", 1st edition, Springer US, New York, 2009.
5: Z. Dostál, D. Horák, R. Kucera, "Total FETI - an easier implementable variant of the FETI method for numerical solution of elliptic PDE", Commun. Numer. Methods Eng., 22, 1155-1162, 2006. doi:10.1002/cnm.881
6: ANSYS, http://www.ansys.com
7: COMSOL Muliphysics, http://www.comsol.com

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £140 +P&P)

	Computational & Technology Resources an online resource for computational, engineering & technology publications
	not logged in - login
Front Page Browse CCP CSETS CTR IJRT Other Authors Search Purchase Guide FAQ Contact us	Civil-Comp Proceedings ISSN 1759-3433 CCP: 91 PROCEEDINGS OF THE TWELFTH INTERNATIONAL CONFERENCE ON CIVIL, STRUCTURAL AND ENVIRONMENTAL ENGINEERING COMPUTING Edited by: B.H.V. Topping, L.F. Costa Neves and R.C. Barros Paper 116 MatSol: Parallel Algorithms based on Domain Decomposition Methods V. Vondrák, T. Kozubek, A. Markopoulos, T. Brzobohatý and R. Kucera Department of Applied Mathematics, VŠB-Technical University of Ostrava, Czech Republic doi:10.4203/ccp.91.116 purchase the full-text of this paper Full Bibliographic Reference for this paper , "MatSol: Parallel Algorithms based on Domain Decomposition Methods", in B.H.V. Topping, L.F. Costa Neves, R.C. Barros, (Editors), "Proceedings of the Twelfth International Conference on Civil, Structural and Environmental Engineering Computing", Civil-Comp Press, Stirlingshire, UK, Paper 116, 2009. doi:10.4203/ccp.91.116 Keywords: domain decomposition, scalable algorithms, parallel algorithms, contact problems. Summary Since 1992, the research at the Department of Applied Mathematics of the VŠB-Technical University of Ostrava has been focused to development of scalable algorithms for contact problems and contact shape optimization problems. These algorithms are based on FETI domain decomposition methods which are well known for their parallel and numerical scalability. Our algorithms were originally implemented using a C++ library called OOSol (Object Oriented SOLvers) [1]. This library benefits from modularity and extensibility, nevertheless the slow development of the code caused by the lack of availability of advanced auxiliary algorithms necessary for debugging of complex algorithms, was its biggest disadvantage. Therefore we started to implement simultaneously all these algorithms into a new library that is developed in The Mathworks Matlab environment [2] which is equipped with many of these helpful functions. We call this library MatSol (MATlab SOLvers) [3]. Several years ago the Mathworks company introduced the Matlab Distributed Computing Engine which allowed Matlab functions to run on parallel computers. Hence, the MatSol has the full functionality of the OOSol library including the parallel algorithms and recently represents our primary testing and developing library. At the present time the library already contains many tools for the solution of realistic engineering problems such as problems in structural mechanics, electromagnetics, acoustics etc. In our paper, we present the functionality of the MatSol library to the solution of realistic contact problems with millions of degrees of freedom showing parallel and numerical scalability of the implemented total FETI method [4,5]. The interface between MatSol and ANSYS or COMSOL [6,7] is presented as well. This feature allows to simply plug the MatSol library into the commercial finite element packages. Some comparisons of the commercial solver ANSYS and MatSol is shown. The efficiency of the solving algorithms will be presented on such complex problems as composite modelling or contact shape optimization problems. References 1 OOSol library, http://www.am.vsb.cz/oosol 2 MATLAB - The Language of Technical Computing, http://www.mathworks.com 3 MatSol library, http://www.am.vsb.cz/matsol 4 Z. Dostál, "Optimal Quadratic Programming Algorithms, with Applications to Variational Inequalities", 1st edition, Springer US, New York, 2009. 5 Z. Dostál, D. Horák, R. Kucera, "Total FETI - an easier implementable variant of the FETI method for numerical solution of elliptic PDE", Commun. Numer. Methods Eng., 22, 1155-1162, 2006. doi:10.1002/cnm.881 6 ANSYS, http://www.ansys.com 7 COMSOL Muliphysics, http://www.comsol.com purchase the full-text of this paper (price £20) go to the previous paper go to the next paper return to the table of contents return to the book description purchase this book (price £140 +P&P)
Back to top	©Civil-Comp Limited 2023 - terms & conditions