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

Title: Extending Theory for Domain Decomposition to Irregular Subdomains.

Abstract

Abstract not provided.

Authors:
; ;
Publication Date:
Research Org.:
Sandia National Lab. (SNL-NM), Albuquerque, NM (United States)
Sponsoring Org.:
USDOE National Nuclear Security Administration (NNSA)
OSTI Identifier:
1137304
Report Number(s):
SAND2007-1737C
523731
DOE Contract Number:
AC04-94AL85000
Resource Type:
Conference
Resource Relation:
Conference: Proposed for presentation at the 17th International Conference on Domain Decomposition Methods held July 3-7, 2006 in St. Wolfgang/Strobl, Austria.
Country of Publication:
United States
Language:
English

Citation Formats

Dohrmann, Clark R., Axel Klawonn, and Olof Widlund. Extending Theory for Domain Decomposition to Irregular Subdomains.. United States: N. p., 2007. Web.
Dohrmann, Clark R., Axel Klawonn, & Olof Widlund. Extending Theory for Domain Decomposition to Irregular Subdomains.. United States.
Dohrmann, Clark R., Axel Klawonn, and Olof Widlund. Thu . "Extending Theory for Domain Decomposition to Irregular Subdomains.". United States. doi:. https://www.osti.gov/servlets/purl/1137304.
@article{osti_1137304,
title = {Extending Theory for Domain Decomposition to Irregular Subdomains.},
author = {Dohrmann, Clark R. and Axel Klawonn and Olof Widlund},
abstractNote = {Abstract not provided.},
doi = {},
journal = {},
number = ,
volume = ,
place = {United States},
year = {Thu Mar 01 00:00:00 EST 2007},
month = {Thu Mar 01 00:00:00 EST 2007}
}

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:
  • A method for a fast solution of the Poisson equation on irregular regions is proposed. The method, which may be called the Boundary Integral-based Domain Decomposition (BIDD) method, seems to be particularly efficient for parallel processing of large scale boundary value problems. For instance, in a hierachical multiprocessor system like the Cedar architecture, each cluster and each processor of a cluster can solve an independent subproblem resulting from the BIDD approach. 13 refs.
  • Abstract not provided.
  • A large class of radiative transfer and particle transport problems contain highly diffusive regions. It is possible to reduce computational costs by solving a diffusion problem in diffusive subdomains instead of the transport equation. This enables one to decrease the dimensionality of the transport problem. In this paper we present a methodology for decomposition of a spatial domain of a transport problem into transport and diffusion subregions. We develop methods for solving one-group problems in 1D slab geometry. To identify and locate diffusive regions, we develop metrics for measuring transport effects that are based on the quasidiffusion (Eddington) factor. Wemore » present the results of test problems that demonstrate the accuracy of the proposed methodology.« less
  • For accurate structural interpretation of dipmeter data, it is important to project well bores onto sections correctly. Current methods include the DVA and tangent plots of Bengtson's statistical curvature analysis techniques (SCAT). However, if the well bore crosses domain boundaries, i.e., surfaces across which the projection type (cylindrical or conical) or structural axis changes, all current methods fail: the plots show composite patterns that are difficult to interpret, if not misleading. Thus, methods for detecting domain boundaries, so that data from different domains can be analyzed separately, become important. The authors have developed and extensively tested a rapid, graphical methodmore » based on a series of AD plots - plots of apparent dip in arbitrary directions vs. measured depth. (SCAT's T and L plots thus become AD plots in the T and L directions.) In addition, the authors have developed an algorithm for calculating the most likely L direction of a given data set. They apply this calculation to equally spaced intervals and show the results on a separate plot, thereby providing interpreters with another aid for detecting domain boundaries. Another important requirement is to determine how azimuth-frequency distribution (AFD) varies with depth. The traditional method - displaying rose diagrams at equally-spaced intervals - has two drawbacks: first, interpreters must study this display carefully to understand how AFD varies with depth; second, rose diagrams themselves are often misleading. In contrast, only a quick glance at a stack of rectangular historgrams is necessary for determining how AFD varies with depth; moreover, historgrams always depict AFD faithfully.« less
  • The emerging discipline of domain analysis, modeling, and engineering, or D-AME, has received most attention from the field of systematic software reuse, where the term {open_quotes}domain{close_quotes} usually denotes a well-scoped area of functionality within a set or class of software systems. A central challenge in D-AME research has been in defining processes and representations sufficiently general to apply in the diverse organizational and technical environments in which D-AME can make useful contribution. The systematic reuse community has established ambitious goals for what a D-AME process should address, such as the ability to support design for reuse for all products andmore » processes of the software life cycle, and applicability beyond software domains: e.g., to domains such as business processes, product variability models, or more generally, domains of shared knowledge about particular technical areas of expertise. In practice, though, the search for generalized domain analysis processes and methods has been fraught. with difficulties. Obstacles include: adoption of a too-narrow conception of the nature of {open_quotes}domains{close_quotes}; tight coupling of D-AME process and methods with software engineering representations; and a consequent lack of understanding of the unique aspects of D-AME as a qualitative process. This paper discusses the goals for the extensibility of D-AME, the primary barriers to achieving these goals, and specific features of the Organization Domain Modeling (ODM) methodology that address these issues. ODM is structured as a core life cycle process model which is broadly applicable to diverse domains and organizational contexts. The core process is augmented by a set of supporting methods which facilitate tailorability, for example, by encapsulating commitments to specific software design representations and processes.« less