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

Title: Toward Reliable Validation of HPC Network Simulation Models

Authors:
; ; ; ; ; ; ; ; ;
Publication Date:
Research Org.:
Lawrence Livermore National Lab. (LLNL), Livermore, CA (United States)
Sponsoring Org.:
USDOE
OSTI Identifier:
1413174
Report Number(s):
LLNL-CONF-733848
DOE Contract Number:
AC52-07NA27344
Resource Type:
Conference
Resource Relation:
Conference: Presented at: Winter Simulation Conference, Las Vegas, NV, United States, Dec 03 - Dec 06, 2017
Country of Publication:
United States
Language:
English
Subject:
97 MATHEMATICS, COMPUTING, AND INFORMATION SCIENCE

Citation Formats

Mubarak, M, Jain, N, Domke, J, Wolfe, N, Ross, C, Li, K, Bhatele, A, Carothers, C D, Ma, K L, and Ross, R B. Toward Reliable Validation of HPC Network Simulation Models. United States: N. p., 2017. Web.
Mubarak, M, Jain, N, Domke, J, Wolfe, N, Ross, C, Li, K, Bhatele, A, Carothers, C D, Ma, K L, & Ross, R B. Toward Reliable Validation of HPC Network Simulation Models. United States.
Mubarak, M, Jain, N, Domke, J, Wolfe, N, Ross, C, Li, K, Bhatele, A, Carothers, C D, Ma, K L, and Ross, R B. Mon . "Toward Reliable Validation of HPC Network Simulation Models". United States. doi:. https://www.osti.gov/servlets/purl/1413174.
@article{osti_1413174,
title = {Toward Reliable Validation of HPC Network Simulation Models},
author = {Mubarak, M and Jain, N and Domke, J and Wolfe, N and Ross, C and Li, K and Bhatele, A and Carothers, C D and Ma, K L and Ross, R B},
abstractNote = {},
doi = {},
journal = {},
number = ,
volume = ,
place = {United States},
year = {Mon Jun 26 00:00:00 EDT 2017},
month = {Mon Jun 26 00:00:00 EDT 2017}
}

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:
  • The safety assessment of a proposed repository for radioactive waste involves predictions of the long-term behavior of the repository system using quantitative modeling of natural and engineered subsystems. Considering that nuclear waste disposal programs gradually leave a generic stage and enter a stage of application, it is clear that the need for validated models will increase. There seems to be general consensus that future performance assessment model systems should be more firmly based on experimental and natural evidence than current model systems. In October 1987, the international INTRAVAL project was started in Stockholm as an international effort toward validation ofmore » geosphere models. The project has been initiated by the Swedish Nuclear Power Inspectorate (SKI) and was prepared by an ad hoc group with representatives from eight national and international organizations. The purpose of the project is to increase the understanding of how various processes and geohydrological structures of importance for the transport of radionuclides from a repository to the biosphere can be properly described by mathematical models and how models developed for this purpose can adequately simulate the nuclide transport during short as well as very long time periods.« less
  • Under the auspices of Pacific Northwest National Laboratory’s Signature Discovery Initiative (SDI), the research team developed a series of Bayesian Network models to assess multi-source signatures of nuclear programs. A Bayesian network is a mathematical model that can be used to marshal evidence to assess competing hypotheses. The purpose of the models was to allow non-expert analysts to benefit from the use of expert-informed mathematical models to assess nuclear programs, because such assessments require significant technical expertise ranging from the nuclear fuel cycle, construction and engineering, imagery analysis, and so forth. One such model developed under this research was aimedmore » at assessing the consistency of open-source information about a nuclear facility with the facility’s declared use. The model incorporates factors such as location, security and safety features among others identified by subject matter experts as crucial to their assessments. The model includes key features, observables and their relationships. The model also provides documentation, which serves as training materials for the non-experts.« less
  • An exact treatment of the positron and electron in a two-chain, Path Integral Monte Carlo (PIMC) simulation is used to calculate both self-annihilation and pickoff rates at finite temperature. It has already been demonstrated that this technique can reproduce and extend results of simple theories of positrons and positronium (Ps) in spherical voids. Here, we include the effect of the linear dielectric response of a homogeneous material on the annihilation rate of positrons and Ps. In addition, we find lifetimes and structural information for Ps in cylindrical channels, both with and without adsorbed fluid atoms.
  • This paper describes new research being performed to improve understanding of seismic waves generated by underground nuclear explosions (UNE) by using full waveform simulation, high-performance computing and three-dimensional (3D) earth models. The goal of this effort is to develop an end-to-end modeling capability to cover the range of wave propagation required for nuclear explosion monitoring (NEM) from the buried nuclear device to the seismic sensor. The goal of this work is to improve understanding of the physical basis and prediction capabilities of seismic observables for NEM including source and path-propagation effects. We are pursuing research along three main thrusts. Firstly,more » we are modeling the non-linear hydrodynamic response of geologic materials to underground explosions in order to better understand how source emplacement conditions impact the seismic waves that emerge from the source region and are ultimately observed hundreds or thousands of kilometers away. Empirical evidence shows that the amplitudes and frequency content of seismic waves at all distances are strongly impacted by the physical properties of the source region (e.g. density, strength, porosity). To model the near-source shock-wave motions of an UNE, we use GEODYN, an Eulerian Godunov (finite volume) code incorporating thermodynamically consistent non-linear constitutive relations, including cavity formation, yielding, porous compaction, tensile failure, bulking and damage. In order to propagate motions to seismic distances we are developing a one-way coupling method to pass motions to WPP (a Cartesian anelastic finite difference code). Preliminary investigations of UNE's in canonical materials (granite, tuff and alluvium) confirm that emplacement conditions have a strong effect on seismic amplitudes and the generation of shear waves. Specifically, we find that motions from an explosion in high-strength, low-porosity granite have high compressional wave amplitudes and weak shear waves, while an explosion in low strength, high-porosity alluvium results in much weaker compressional waves and low-frequency compressional and shear waves of nearly equal amplitude. Further work will attempt to model available near-field seismic data from explosions conducted at NTS, where we have accurate characterization of the sub-surface from the wealth of geological and geophysical data from the former nuclear test program. Secondly, we are modeling seismic wave propagation with free-surface topography in WPP. We have model the October 9, 2006 and May 25, 2009 North Korean nuclear tests to investigate the impact of rugged topography on seismic waves. Preliminary results indicate that the topographic relief causes complexity in the direct P-waves that leads to azimuthally dependent behavior and the topographic gradient to the northeast, east and southeast of the presumed test locations generate stronger shear-waves, although each test gives a different pattern. Thirdly, we are modeling intermediate period motions (10-50 seconds) from earthquakes and explosions at regional distances. For these simulations we run SPECFEM3D{_}GLOBE (a spherical geometry spectral element code). We modeled broadband waveforms from well-characterized and well-observed events in the Middle East and central Asia, as well as the North Korean nuclear tests. For the recent North Korean test we found that the one-dimensional iasp91 model predicts the observed waveforms quite well in the band 20-50 seconds, while waveform fits for available 3D earth models are generally poor, with some exceptions. Interestingly 3D models can predict energy on the transverse component for an isotropic source presumably due to surface wave mode conversion and/or multipathing.« less