Computational Technology Resources - CCC

Ch. Song¹, J.Q. Zhang², A.S. Kumar¹ and Y.F. Zhan¹

¹Centre for Infrastructure Engineering and Safety, University of New South Wales, Sydney, Australia
²Key Laboratory of Urban Security and Disaster Engineering of Ministry of Education, Beijing University of Technology, Beijing, China

doi:10.4203/ccc.6.14.1

Ch. Song, J.Q. Zhang, A.S. Kumar, Y.F. Zhan, "Towards Fully-automated High-performance Scaled Boundary Finite Element Analysis", in P. Ivanyi, J. Kruis, B.H.V. Topping, (Editors), "Proceedings of the Seventeenth International Conference on Civil, Structural and Environmental Engineering Computing", Civil-Comp Press, Edinburgh, UK, Online volume: CCC 6, Paper 14.1, 2023, doi:10.4203/ccc.6.14.1

Keywords: scaled boundary finite element method, mesh generation, octree mesh, parallel computing, statics, dynamics.

Abstract

This paper presents the development of the scaled boundary finite element method to benefit from modern technologies for geometrical modelling and high-performance computing. The scaled boundary finite element method allows the use of arbitrarily shaped star-convex polyhedral elements. The greater flexibility in spatial discretization than standard finite elements facilitates automatic mesh generation. A simple and efficient octree algorithm is developed to mesh geometric models given in common formats such as conventional CAD, STL, digital images, and point clouds. By identifying suitable transformations of the octree cells, a mesh can be deconstructed into a limited number of unique cell patterns. A pattern-by-pattern method for computing matrix-vector products in explicit dynamics and iterative solvers is developed. The operations grouping elements of the same pattern reduce the memory requirement and improve the parallel computation efficiency. Numerical examples of large-scale problems with complex geometries are presented. A significant speedup is observed for these examples with up to 1 billion degrees of freedom and running on up to 16,384 computing cores.

download the full-text of this paper (PDF, 19039 Kb)

go to the previous paper
go to the next paper
return to the table of contents
return to the volume description

	Computational & Technology Resources an online resource for computational, engineering & technology publications
	not logged in - login
Front Page Browse CCC CCP CSETS CTR IJRT Other Authors Search Purchase Guide FAQ Contact us	Civil-Comp Conferences ISSN 2753-3239 CCC: 6 PROCEEDINGS OF THE SEVENTEENTH INTERNATIONAL CONFERENCE ON CIVIL, STRUCTURAL AND ENVIRONMENTAL ENGINEERING COMPUTING Edited by: P. Ivanyi, J. Kruis and B.H.V. Topping Paper 14.1 Towards Fully-automated High-performance Scaled Boundary Finite Element Analysis Ch. Song¹, J.Q. Zhang², A.S. Kumar¹ and Y.F. Zhan¹ ¹Centre for Infrastructure Engineering and Safety, University of New South Wales, Sydney, Australia ²Key Laboratory of Urban Security and Disaster Engineering of Ministry of Education, Beijing University of Technology, Beijing, China doi:10.4203/ccc.6.14.1 Full Bibliographic Reference for this paper Ch. Song, J.Q. Zhang, A.S. Kumar, Y.F. Zhan, "Towards Fully-automated High-performance Scaled Boundary Finite Element Analysis", in P. Ivanyi, J. Kruis, B.H.V. Topping, (Editors), "Proceedings of the Seventeenth International Conference on Civil, Structural and Environmental Engineering Computing", Civil-Comp Press, Edinburgh, UK, Online volume: CCC 6, Paper 14.1, 2023, doi:10.4203/ccc.6.14.1 Keywords: scaled boundary finite element method, mesh generation, octree mesh, parallel computing, statics, dynamics. Abstract This paper presents the development of the scaled boundary finite element method to benefit from modern technologies for geometrical modelling and high-performance computing. The scaled boundary finite element method allows the use of arbitrarily shaped star-convex polyhedral elements. The greater flexibility in spatial discretization than standard finite elements facilitates automatic mesh generation. A simple and efficient octree algorithm is developed to mesh geometric models given in common formats such as conventional CAD, STL, digital images, and point clouds. By identifying suitable transformations of the octree cells, a mesh can be deconstructed into a limited number of unique cell patterns. A pattern-by-pattern method for computing matrix-vector products in explicit dynamics and iterative solvers is developed. The operations grouping elements of the same pattern reduce the memory requirement and improve the parallel computation efficiency. Numerical examples of large-scale problems with complex geometries are presented. A significant speedup is observed for these examples with up to 1 billion degrees of freedom and running on up to 16,384 computing cores. download the full-text of this paper (PDF, 19039 Kb) go to the previous paper go to the next paper return to the table of contents return to the volume description
Back to top	©Civil-Comp Limited 2023 - terms & conditions