A Multi-Mechanics Approach to Computational Heat Transfer

Abstract

We present a turbulent combustion code for modeling heat transfer in fires that arise in accident scenarios. The code is a component of a multi-mechanics framework and is based on a domain-decomposition, message-passing approach to parallel computing. The turbulent combustion code is based on a vertex-centered, finite-volume scheme for 3D unstructured meshes. The multi-mechanics nature of the frameworks allows us to couple to a conduction heat transfer code for conjugate heat transfer problems or a participating media radiation code for radiation transport in soot-laden flows. We describe our numerical methods, our approach to parallel computing, and the multi-mechanics frameworks. We demonstrate parallel performance using some example verification problems.

Diagnostic development for determining the joint temperature/soot statistics in hydrocarbon-fueled pool fires : LDRD final report.

“…The soot radiation flux is a key threat to the safety and security of personnel and infrastructure and is a dominant concern for risk assessment and analysis. New computational models are currently being developed for the Advanced Simulation and Computing (ASC) code FUEGO [2][3][4] at Sandia National…”

Analysis of jet flames and unignited jets from unintended releases of hydrogen

“…The computational domain was discretized into 6x10 5 elements with nonuniform size (finer mesh at end of tube, z=0). The boundary condition for the jet exit was a uniform velocity of 2.96x10 4 cm/s at the end of the tube. The outer boundaries of the rectangular domain were defined as open boundaries with a constant pressure boundary condition.…”

“…A constant pressure boundary condition was applied around the circumference and at the open end of the cylinder. Results were computed on two hexagonal element meshes (5.9x10 4 and 4.8x10 5 elements where the refined mesh is obtained by dividing each coarse mesh element into 8 smaller elements) that expanded in both the axial and radial directions from the jet. Results were obtained for the standard k-ε and RNG k-ε turbulence models and for two convection operators, the 1st order upwind scheme and MUSCL scheme [4].…”

“…Results were computed on two hexagonal element meshes (5.9x10 4 and 4.8x10 5 elements where the refined mesh is obtained by dividing each coarse mesh element into 8 smaller elements) that expanded in both the axial and radial directions from the jet. Results were obtained for the standard k-ε and RNG k-ε turbulence models and for two convection operators, the 1st order upwind scheme and MUSCL scheme [4]. The figure of merit for the accuracy of the simulation is based on the centerline velocity decay, which in the self-similar region of a free turbulent jet is correlated by an expression of the form ( )…”

“…The FUEGO code was designed to simulate turbulent, reacting flow and heat transfer [4] on massively parallel computers, with a primary focus on heat transfer to objects in pool fires. More recently the code has been adapted for compressible flow and hydrogen combustion.…”