ORNL/JICS/UTK CFD Workshop
CFD Workshop Home > Agenda > ORNL/JICS/UTK CFD Workshop

First ORNL/JICS/UTK Joint CFD (Computational Fluid Dynamics) Workshop
December 15th, 2006
JICS Auditorium (in bldg. 5100 facing the visitor center), ORNL

| CFD Workshop Home | Agenda | Abstracts |


Abstracts


AMROC - A Cartesian SAMR Framework for Large-scale Simulation of Hyperbolic Flow Problems
Ralf Deiterding, Computational Mathematics Group, Oak Ridge National Laboratory

Finite-volume-based shock capturing schemes are the most appropriate numerical methods for computational fluid dynamics of compressible high speed flows. Very high order discretizations are nowadays available, but achieving the theoretical possible order of convergence can be very cumbersome on unstructured meshes. On the other hand, structured (Cartesian) methods are only applicable to simple geometries and cannot directly be employed for many problems arising in science and engineering.

The talk presents an approach to utilizing Cartesian schemes for real-world configurations that combines the ghost-fluid idea with block-structured adaptive mesh refinement (SAMR). A scalar level set function storing the distance information to the boundary surface is used to consider arbitrary geometries on the Cartesian mesh without ambiguities. Minor approximation inaccuracies near the embedded boundary are alleviated by increasing the resolution non-uniformly. Various examples will be given which demonstrate that the approach can be efficiently used for serious computational investigations. Among the simulations described are shock-induced mixing, fully resolved detonation waves with detailed chemistry in realistic experimental devices, and three-dimensional fluid-structure interaction simulations. The fluid-structure interaction cases involve large deformation, fracture and fragmentation and confirm the applicability to problems with major topology evolutions. The object-oriented software design and distributed memory performance will be sketched briefly.

[1] R. Deiterding. AMROC - Blockstructured Adaptive Mesh Refinement in Object-oriented C++. http://amroc.sourceforge.net

[2] R. Deiterding, F. Cirak, S. P. Mauch, D. J. Hill, C. Pantano, J. C. Cummings, and D. I. Meiron. Virtual Test Facility: A virtual shock physics facility for simulating the dynamic response of materials.
http://www.cacr.caltech.edu/asc

Back to Agenda


Multiphysics/Multiscale Modeling of Heterogeneous Chemically Reacting Flows
Sreekanth Pannala, Computer Science and Mathematics Division, ORNL

Gas-solid chemically reacting flows are omnipresent in many multiphase flow reactors in various industries like Chemical, Fossil and Nuclear. The challenging aspect of modeling these reacting flows are the wide range of both temporal and spatial scales encountered in these systems. This talk would give an overview of the Computational Fluid Dynamics (CFD) aspects of a gas-solid contact reactor and the steps to couple the continuum scale simulations to the other simulations at both smaller and larger scales.

Back to Agenda


A Rational LES CFD Theory for PICMSS Implementation
AJ Baker, University of Tennessee

A newly established and verified, mathematically robust rational LES CFD theory is developed for fluid-thermal systems analysis. Using a high-order Pade' approximation of the Gaussian filter, coupled with formal Fourier transformations, the theory establishes an analytical solution for three of the four LES theory-generated Reynolds stress tensors, also a Reynolds heat flux vector for non-isothermal flows. The resultant coupled system of non-linear, initial-value, harmonic elliptic boundary value PDEs explicitly contains the filter scale, as well as reducing the need for modeling to only the subgrid scale (SGS) dissipation tensor (heat flux vector). The rational LES theory clearly defines acceptable scale ordering for such models, confirming that legacy SGS models are typically of inappropriate order in filter dimension. Referencing a recent PhD dissertation validation exercise, the end goal is implementation into the JICS scalable parallel simulation platform (PICMSS) to support a new level of simulation fidelity for pertinent fluid-thermal systems on Leadership Class supercomputers.

Back to Agenda


Two-phase Simulations in Support of the SNS Liquid Mercury Target Development
Ashraf Ibrahim, SNS

Investigations at the Oak Ridge National Laboratory (ORNL) Spallation Neutron Source (SNS) facility are proceeding in the area of two-phase flow due to the anticipated performance benefits of strategically introducing gas into the liquid mercury target. As part of an effort to validate computational fluid dynamics (CFD) models, capability to predict free-surface gas-liquid flows, simulations and experiments of gas injection in stagnant water have been carried out. The Volume of Fluid (VOF) model, implemented in the commercially available CFD code CFX-10 from ANSYS, was used to simulate the unsteady two-phase flow of gas injection in a stagnant liquid. Flow visualization data were obtained with a high-speed camera for the comparison of predicted and measured bubble sizes and shapes at various stages of the bubble growth, detachment, and gravitational rise. The VOF model is validated by comparing detailed bubble sizes and shapes at various stages of the bubble growth and detachment, with the experimental measurements at different gas flow rates. The acoustic sound signals emitted at the time of bubble detachment and resonating frequency of the bubble were recorded with a microphone and then subsequently correlated to the visual images for a specific event. As the gas flow rate increases, the irregularity, amplitude and frequency fluctuations increase. The acoustic signature aspect is particularly interesting since it has applicability to the injection of gas in liquid mercury, which is opaque

Back to Agenda


Computational Fluid Dynamics Applications in Microscale
Cheng-Xian Lin, The University of Tennessee

With the rapid advancement in computer science and infrastructure, computational fluid dynamics (CFD) as a tool for research and design has penetrated almost every industrial sector where fluid flow and heat transfer are of importance. Computational fidelity, reliability, uncertainty, and limitation of CFD codes have gained more and more attentions from researchers for both complex and emerging problems in recent years. In this talk, Dr. Lin will give a look at the available mathematical models for flows in continuum and non-continuum regimes, and a discussion about CFDs capability, limit, verification and validation for modeling flows in both macro and micro scales. Several previously funded research projects will be used to illustrate the applications of CFD in aircraft icing, biomedical flows, nuclear waste transport, electronic cooling, and micropropulsion systems.

Back to Agenda


Numerical Simulations of Turbulent Combustion
Ramanan Sankaran, NCCS/ORNL

I will describe my research interests - DNS of turbulent combustion, combustion modeling for RANS/LES applications. I will also discuss my work at NCCS as part of/for the combustion application area.

Back to Agenda


Simulation of Biodiesel Performance and Emissions in Advanced Combustion Modes
Joanna McFarlane, Oak Ridge National Laboratory

Advanced combustion modes are being developed for the reduction of soot and NOx production from the burning of fuel in automobile engines. The impact of the use of bio-based fuels on advanced combustion is particulary pertinent given their popularity as alternatives to petroleum-derived fuels. The combustion of biodiesel is being simulated in an automobile engine, the results of which will be compared to actual engine tests on biodiesel and biodiesel blends. This study is in progress, but recent efforts have established a reaction mechanism for soy biodiesel, and have explored methods of mechanism reduction. In parallel, spray model development using interface tracking has the potential to accurately predict the effect of fuel physical properties on droplets and sprays.

Back to Agenda


Interdendritic Flows during Metal Casting Solidification
Adrian Sabau, ORNL

Models for the prediction of interdendritic cavity defects, from which cracks can easily propagate, are necessary for the advancement of metal casting processes. It is shown that the location and severity of interdendritic cavity defects can be predicted using methods based on thresholds for interdendritic liquid metal flows. A short review on cavitation phenomena is presented and future directions are identified.

Back to Agenda


Computer Fire Modeling
David Icove, University of Tennessee

A case is reported in which computer fire modeling was used to reevaluate a fire that killed three fire fighters. The National Institute of Standards and Technology (NIST) Fire Dynamics Simulator (FDS) was employed to model the fire in order to estimate the concentration of carbon monoxide present in the dwelling. The carbon monoxide was the immediate cause of death of two of the fire fighters, who appear to have removed their face pieces in order to share available air. This estimate, along with an assumed respiration volume and known blood carboxyhemoglobin, was used to estimate the time of exposure.

Back to Agenda