Contract%20Year%201%20Review%20Computational%20Structural%20Mechanics%20(CSM)

1
Contract Year 1 ReviewComputational Structural
Mechanics (CSM)

• J. Tinsley Oden
• TICAM, The University of Texas at Austin
• 14 May 2002

2
Strategic Goals for CSM

• Top priorities
• Provide the DoD HPC user community with the
  latest tools, technical support, and productivity
  enhancements for CSM computer simulations
• Provide the means for technology transfer and
  transition support, training content, user
  community development, and on-site support for
  CSM
• Increase reliability and efficiency of CSM
  simulations by incorporating improved
  technologies into legacy codes
• Extract useful information from those simulations

3
Strategic Goals for CSM

• Provide courses, workshops, conferences and
  seminars on modern CSM methodology
• Collaborate with MSIs on technical work and
  courses
• Technical projects on
• New high-fidelity algorithms
• Development of technologies for level-2 code
  coupling for multi-physics applications
• Development of formalism and algorithms to
  enhance verification and validation of computer
  simulations
• Implementation of damage and failure models for
  multi-componentand non-isotropic materials
• Development of enhanced algorithms and codes for
  Parallel Adaptive mesh Refinements and Error
  Control
• Improvements in contact-impact algorithms for
  highly dynamicenvironments
• Improved parallel partitioning and mesh
  generation strategies for CSM applications
• Exploration of developing technologies for
  Advanced CSM, including computational frameworks
  and integrated scalable technologies
• ( - indicates project funded during FY02)

4
CSM FY02 Project Tasks

• CSM 001 Integrated Scalable Technology and
  Unified Framework for CSM and Contact-Impact
  Applications (Not funded due to budget cuts)
• CSM 007 Code Coupling Technologies for
  Multi-Physics Applications
• CSM 009 High-Fidelity Algorithms for Coupled and
  Multi-Physics Applications

5
CSM Core Support

• FAPOC roles and responsibilities
• Lead a team of experts within CSM to support
  technology transfer/transition
• Coordinate with other FAs to achieve cross-FA
  integration
• Coordinate overall program to insure full
  participation of MSIs
• FAPOC core support activities
• Sharouz Aliabadi, CAU - TICAM visit, lecture,
  discussion of joint projects
• Graham Carey - visit to ERDC, technical
  discussions on code coupling
• Ron Hinrichsen, ASC on-site - TICAM visit,
  lecture, project discussions
• Rick Weed, ERDC on-site - TICAM visit, technical
  discussions
• HPC 2002 - Littlefield, Carey, Boisseau organized
  sessions
• David Littlefield - Contact with Ray Bell and
  Greg Bessette, Sandia, concerning ZAPOTEC
  development
• Graham Carey visit to MSU and Bharat Soni,
  organized joint CSM/CFD workshop
• J. Tinsley Oden - Joint CSM-CFD Workshop,
  scheduled for July 2002

6
CSM Core Support

• On-site activities ERDC (Richard Weed)
• Implementation of latest versions of LLNL
  structural dynamics codes at ERDC MSRC to support
  Pentagon retrofit program
• Implementation of CTH with TICAM enhancements at
  ERDC MSRC
• Implementation of ZAPOTEC code on ERDC SGI and
  Compaq systems
• Developed initial package of code modifications
  to support coupled ocean simulation code PIPS
  (cross-CTA support)
• Support TICAMs effort to define PET projects,
  brief local users on TICAM PET, provide input for
  program reviews
• On-site activities ASC (Ron Hinrichsen)
• Hydrodynamic Ram (HRAM) Modeling
• HRAM Simulator Design And Modeling
• MANPADS Penetration Modeling
• Flutter Modeling of Damaged Wings
• Adhesively Bonded/Co-cured Composite Damage
  Modeling

7
Technical (Metrics)

• User codes impacted by PET
• Codes directly impacted CTH and EPIC
• Codes indirectly impacted (i.e. methods developed
  can be applied to these codes) PARADYN, PRONTO,
  ALEGRA, ALE3D
• Other codes indirectly impacted (broadened
  capabilities due to CFD/CSM and CSM/Thermo
  coupling) CALORE, ZAPOTEC
• Other codes used and evaluated
• Pre- and post-processing codes CUBIT (licensed
  and in use on-site), MESQUITE (will be licensed
  when released)
• Dynamic load balancing codes CHACO and ZOLTAN
  (both licensed from SNL, used in code coupling
  work)

8
Technical (Metrics)

• User Contacts and Assistance
• CSM User Advisory Board
• Photios Papados (ARL)
• James Baylot (ERDC)
• Alexandra Landsberg (NSWC)
• Anthony Palazotto (AFIT)
• Stephen Schraml (ARL)
• Mike Giltrud (Govt. CTA lead, DTRA)
• Others
• Joey Gill (NAVAIR)
• Steve Akers (ERDC)
• Robert Hall (ERDC)
• Monty Moshier (AFRL)

9
Technical (Metrics)

• Tools introduced to community
• Unidirectional coupling of CTH and DYNA3D
  installed at ERDC MSRC and in use
• Portions of adaptive mesh refinement work have
  been introduced into production version of AMRCTH
• Development of other tools is in progress, not
  yet introduced

10
Technical (Metrics)

• Activities and Presentations
• Prudhomme, S., J.T. Oden, A. Romkes, and F.
  Pascal, Searching for a computable error
  estimator for Discontinuous Galerkin
  approximations of elliptic problems, S'eminaire
  "Groupe de travail num'erique", given at the
  Laboratoire Analyse Num'erique et EDP,
  Universit'e Paris-Sud, Orsay, France, October 10,
  2001.
• PET Kick Off Oden and Carey met with DoD Users
  and administrators and presented strategic plan,
  The University of Texas at Austin, October 25-26,
  2001.
• Carey, G. F., Seminar, ERDC, Vicksburg, MS,
  Projection Strategies and Interface Treatments
  for Finite Element Modeling , January 23, 2002.
• Carey, G. F., Keynote Lecture, Algorithm
  Adaptivity Workshop (organized by Bharat Soni)
  Engineering Research Center (ERC), Mississippi
  State University, Jan 24-25, 3002.

11
Technical (Metrics)

• Activities and Presentations
• Oden, J. Tinsley, A Theory for Estimation of
  Modeling and Approximation Error in Computational
  Science and Engineering A Formalism for
  Verification and Validation of Computer
  Simulations, 2002 Fowler Distinguished Lecture,
  Texas AM University, College Station, Texas,
  February 19-20, 2002.
• Feb 28-Mar 4, S. Aliabadi at TICAM Gave seminar
  on "Large-Scale High Performance Computing
  Techniques for Simulation of Hydrodynamics
  Problems Involving Floating Objects"
• Carey and number of his students and post docs
  attended Finite Element Rodeo at Texas AM, March
  1-2, Robert McLay gave a talk.
• Carey attended Finite Element Circus, College
  Park, MD, March 8-9.
• Steve Bova from Sandia visited UT and gave a
  presentation on code coupling for CALORE via
  SIERRA, March 11-12.
• Carey to attend DoD Simulation Workshop in
  Marina del Rey, California, April 8-10.

12
Technical (Metrics)

• Activities and Presentations
• David Littlefield and Saeed Iqbal attended
  HPC02, April 14-18, 2002, and present papers.
  Saeed Iqbal presenting Load Balancing with
  Variable Number of Processors on Commodity
  Clusters (S. Iqbal, G. F. Carey, M.A. Padron, J.
  P. Suarez and A. Plaza). David Littlefield
  organized a session (co-ordinated with G. F.
  Carey) Jay Boisseau organized a session
  (coordinated with G. F. Carey) GFC is HPC
  National Committee member.
• Carey scheduling a visit with Photios Papados at
  ARL
• Carey scheduling visits to Sandia and LANL on
  meshing, software, and code coupling
• June 10-14, DoD Users Group Meeting TICAM,
  Robert McLay presenting Physics Coupling Issues
  and a New Treatment Using Dial-an-Operator (R.
  McLay, B. Barth, G. Carey)
• J. Tinsley Oden, Plenary lecture, Adaptive
  modeling in computational solid mechanics, II
  Congreso Internacional de Métodos Numéricos en
  Ingenieria y Ciencias Aplicadas Guanajuato,
  Mexico, January 17-19, 2002

13
Technical (Metrics)

• Technical reports and articles (published,
  submitted or pending since last report)
• Suarez, J. P., G. F. Carey and A. Plaza,
  Graph-based data structures for skeleton-based
  refinement algorithms, CNME 17(12) 903-910,
  2001.
• Iqbal, S. and G. F. Carey, Neural nets for Mesh
  Assessment, TICAM Report 02-02, The University
  of Texas at Austin, January 2002.
• Suarez, J. P., A. Plaza and G. F. Carey, The
  Propagation Problem in Longest-Edge Refinement,
  in preparation, March 2002.
• Romkes, A., J.T. Oden, and S. Prudhomme, A
  priori and a posteriori error analysis of
  stabilized Discontinuous Galerkin Methods, TICAM
  Report, in preparation.
• Prudhomme, S., J.T. Oden, A. Romkes, and F.
  Pascal, A Priori and A Posteriori Error
  Estimation for Discontinuous Galerkin Finite
  Element approximations, Proceedings of the Sixth
  U. S. National Congress on Computational
  Mechanics, Dearborn, MI, August 1-4, 2001.

14
Technical (Metrics)

• Technical reports and articles (published,
  submitted or pending since last report)
• Littlefield, D. L., A brief description of new
  algorithms incorporated into CTH A model for
  rigid obstacles and an interface for coupling
  with structural codes, TICAM Report No. 01-36,
  November 2001.
• Littlefield, D. L., Coupled Lagrangian/Eulerian
  simulation of blast/structure interaction with
  large-scale parallel computing, Proceedings of
  the High Performance Computing Symposium, April
  14 18, San Diego, CA (2002).
• Iqbal, S., Carey, G., Padron, M. A., Suarez, J.
  P. and Plaza, A., Load balancing with variable
  number of processors on commodity clusters,
  Proceedings of the High Performance Computing
  Symposium, April 14 18, San Diego, CA (2002).

15
CSM 007 Code Coupling Technologies for
Multiphysics Applications

• Objective
• Develop a suite of algorithms and codes that
  will permit code coupling in a parallel computing
  environment for CSM simulation tools

• Funding from proposal as approved 144,500
• PI Graham F. Carey

• Milestones (from year 1) development of
  Lagrangian/Eulerian Interface, development of
  projection algorithms, numerical tests
• Deliverables (from Year 1) Technical Reports and
  Papers
• Progress Indicator yellow (but remarkable
  progress has been made considering funding for
  this project began in March)
• Projection algorithms are being developed
• ALE interface has been developed, numerical
  tests indicate the validity of approach but
  further refinements are needed

16
CSM 009 High-Fidelity Algorithms for Coupled
and Multi-Physics Applications

• Objective
• Carry out new work on finite element
  methodology. Of particular interest are the
  development, analysis, testing and implementation
  of new Discontinuous Galerkin techniques.

• Funding from proposal as approved 130,000
• PI J. Tinsley Oden

• Milestones (from Year 1) a priori and a
  posteriori error estimation, development of
  methodology and testing software, numerical
  experiments
• Deliverables (from Year 1) Technical reports and
  papers
• Progress Indicator yellow (but remarkable
  progress has been made considering funding for
  this project began in March)
• A priori error estimates developed, numerical
  tests have been performed
• A comprehensive report describing the methods and
  numerical tests has been prepared

17
Core Financial Summary
18
Financial Summary CSM 007 Code Coupling
Technologies for Multiphysics Applications
19
Financial Summary CSM 009 High-Fidelity
Algorithms for Coupled and Multi-Physics
Applications
20
Staffing

• Team Members
• J Tinsley Oden - professor, director of TICAM
• Graham F. Carey professor, director of CFD Lab
• David L. Littlefield research scientist
• Serge Prudhomme research associate
• Robert T. McLay research associate
• Albert Romkes graduate research assistant
• Saeed Iqbal graduate research assistant
• Lorraine Sanchez executive assistant
• Richard Weed ERDC on-site
• Ron Hinrichsen ASC on-site
• Michael Giltrud CTA, DTRA
• Number of staff departed 0
• However funds have been inadequate to maintain
  original team
• Number of positions filled 0
• Recruitment effort unsuccessful for ARL on-site
  lead

21
Summary

• Funding level a problem
• We are currently funded at a level 30-40 less
  than PET 1, and it will be difficult to continue
  to maintain an effective program without
  increases
• Multi-year planning a problem
• Student and post-doc windows a problem need at
  least 2-year base funding
• Late funding makes team retention difficult
• Project continuity interrupted
• Multi-year projects originally started during
  FY02 not continued in FY03