skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: A parallel, volume-tracking algorithm for unstructured meshes

Abstract

Many diverse areas of industry benefit from the use of volume of fluid methods to predict the movement of materials. Casting is a common method of part fabrication. The accurate prediction of the casting process is pivotal to industry. Mold design and casting is currently considered an art by industry. It typically involves many trial mold designs, and the rejection of defective parts is costly. Failure of cast parts, because residual stresses reduce the part`s strength, can be catastrophic. Cast parts should have precise geometric details that reduce or eliminate the need for machining after casting. Volume of fluid codes will help designers predict how the molten metal fills a mold and where ay trapped voids remain. Prediction of defects due to thermal contraction or expansion will eliminate defective, trial mold designs and speed the parts to market with fewer rejections. Increasing the predictability and therefore the accuracy of the casting process will reduce the art that is involved in mold design and parts casting. Here, recent enhancements to multidimensional volume-tracking algorithms are presented. Illustrations in two dimensions are given. The improvements include new, local algorithms for interface normal constructions and a new full remapping algorithm for time integration. Thesemore » methods are used on structured and unstructured grids.« less

Authors:
; ;  [1];  [2]
  1. Los Alamos National Lab., NM (United States)
  2. Cambridge Power Computing Associates, MA (United States)
Publication Date:
Research Org.:
Los Alamos National Lab., NM (United States)
Sponsoring Org.:
USDOE, Washington, DC (United States)
OSTI Identifier:
383638
Report Number(s):
LA-UR-96-2420; CONF-9605211-2
ON: DE96014239; TRN: AHC29621%%84
DOE Contract Number:
W-7405-ENG-36
Resource Type:
Conference
Resource Relation:
Conference: Parallel CFD `96, Capri (Italy), 20-23 May 1996; Other Information: PBD: [1996]
Country of Publication:
United States
Language:
English
Subject:
99 MATHEMATICS, COMPUTERS, INFORMATION SCIENCE, MANAGEMENT, LAW, MISCELLANEOUS; 36 MATERIALS SCIENCE; 32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATION; METALS; CASTING; FLUID MECHANICS; ALGORITHMS; FOUNDRIES; COMPUTER-AIDED MANUFACTURING; PARALLEL PROCESSING; MESH GENERATION; CASTING MOLDS; FORTRAN; ITERATIVE METHODS

Citation Formats

Mosso, S.J., Swartz, B.K., Kothe, D.B., and Ferrell, R.C. A parallel, volume-tracking algorithm for unstructured meshes. United States: N. p., 1996. Web.
Mosso, S.J., Swartz, B.K., Kothe, D.B., & Ferrell, R.C. A parallel, volume-tracking algorithm for unstructured meshes. United States.
Mosso, S.J., Swartz, B.K., Kothe, D.B., and Ferrell, R.C. Tue . "A parallel, volume-tracking algorithm for unstructured meshes". United States. doi:. https://www.osti.gov/servlets/purl/383638.
@article{osti_383638,
title = {A parallel, volume-tracking algorithm for unstructured meshes},
author = {Mosso, S.J. and Swartz, B.K. and Kothe, D.B. and Ferrell, R.C.},
abstractNote = {Many diverse areas of industry benefit from the use of volume of fluid methods to predict the movement of materials. Casting is a common method of part fabrication. The accurate prediction of the casting process is pivotal to industry. Mold design and casting is currently considered an art by industry. It typically involves many trial mold designs, and the rejection of defective parts is costly. Failure of cast parts, because residual stresses reduce the part`s strength, can be catastrophic. Cast parts should have precise geometric details that reduce or eliminate the need for machining after casting. Volume of fluid codes will help designers predict how the molten metal fills a mold and where ay trapped voids remain. Prediction of defects due to thermal contraction or expansion will eliminate defective, trial mold designs and speed the parts to market with fewer rejections. Increasing the predictability and therefore the accuracy of the casting process will reduce the art that is involved in mold design and parts casting. Here, recent enhancements to multidimensional volume-tracking algorithms are presented. Illustrations in two dimensions are given. The improvements include new, local algorithms for interface normal constructions and a new full remapping algorithm for time integration. These methods are used on structured and unstructured grids.},
doi = {},
journal = {},
number = ,
volume = ,
place = {United States},
year = {Tue Oct 01 00:00:00 EDT 1996},
month = {Tue Oct 01 00:00:00 EDT 1996}
}

Conference:
Other availability
Please see Document Availability for additional information on obtaining the full-text document. Library patrons may search WorldCat to identify libraries that hold this conference proceeding.

Save / Share:
  • We develop a new algorithm for performing parallel S{sub n} sweeps on unstructured meshes. The algorithm uses a low-complexity list ordering heuristic to determine a sweep ordering on any partitioned mesh. For typical problems and with ''normal'' mesh partitionings we have observed nearly linear speedups on up to 126 processors. This is an important and desirable result, since although analyses of structured meshes indicate that parallel sweeps will not scale with normal partitioning approaches, we do not observe any severe asymptotic degradation in the parallel efficiency with modest ({le}100) levels of parallelism. This work is a fundamental step in themore » development of parallel S{sub n} methods.« less
  • We develop a new algorithm for performing parallel S{sub n} sweeps on unstructured meshes. The algorithm uses a low-complexity list ordering heuristic to determine a sweep ordering on any partitioned mesh. For typical problems and with ''normal'' mesh partitionings we have observed nearly linear speedups on up to 126 processors. This is an important and desirable result, since although analyses of structured meshes indicate that parallel sweeps will not scale with normal partitioning approaches, we do not observe any severe asymptotic degradation in the parallel efficiency with modest ({le} 100) levels of parallelism. This work is a fundamental step inmore » the development of parallel S{sub n} methods.« less
  • We discuss selected aspects of a new parallel three-dimensional (3-D) computational tool for the unstructured mesh simulation of Los Alamos National Laboratory (LANL) casting processes. This tool, known as {bold Telluride}, draws upon on robust, high resolution finite volume solutions of metal alloy mass, momentum, and enthalpy conservation equations to model the filling, cooling, and solidification of LANL castings. We briefly describe the current {bold Telluride} physical models and solution methods, then detail our parallelization strategy as implemented with Fortran 90 (F90). This strategy has yielded straightforward and efficient parallelization on distributed and shared memory architectures, aided in large partmore » by new parallel libraries {bold JTpack9O} for Krylov-subspace iterative solution methods and {bold PGSLib} for efficient gather/scatter operations. We illustrate our methodology and current capabilities with source code examples and parallel efficiency results for a LANL casting simulation.« less
  • The Los Alamos National Laboratory (LANL) is currently developing a new casting simulation tool (known as Telluride) that employs robust, high-resolution finite volume algorithms for incompressible fluid flow, volume tracking of interfaces, and solidification physics on three-dimensional (3-D) unstructured meshes. Their finite volume algorithms are based on colocated cell-centered schemes that are formally second order in time and space. The flow algorithm is a 3-D extension of recent work on projection method solutions of the Navier-Stokes (NS) equations. Their volume tracking algorithm can accurately track topologically complex interfaces by approximating the interface geometry as piecewise planar. Coupled to their fluidmore » flow algorithm is a comprehensive binary alloy solidification model that incorporates macroscopic descriptions of heat transfer, solute redistribution, and melt convection as well as a microscopic description of segregation. The finite volume algorithms, which are efficient, parallel, and robust, can yield high-fidelity solutions on a variety of meshes, ranging from those that are structured orthogonal to fully unstructured (finite element). The authors discuss key computer science issues that have enabled them to efficiently parallelize their unstructured mesh algorithms on both distributed and shared memory computing platforms. These include their functionally object-oriented use of Fortran 90 and new parallel libraries for gather/scatter functions (PGSLib) and solutions of linear systems of equations (JTpack90). Examples of their current capabilities are illustrated with simulations of mold filling and solidification of complex 3-D components currently being poured in LANL foundries.« less
  • A new algorithm for performing parallel S{sub n} sweeps on unstructured meshes is developed. The algorithm uses a low-complexity list ordering heuristic to determine a sweep ordering on any partitioned mesh. For typical problems and with 'normal' mesh partitionings, nearly linear speedups on up to 126 processors are observed. This is an important and desirable result, since although analyses of structured meshes indicate that parallel sweeps will not scale with normal partitioning approaches, no severe asymptotic degradation in the parallel efficiency is observed with modest ({<=}100) levels of parallelism. This result is a fundamental step in the development of efficientmore » parallel S{sub n} methods.« less