Software Pages For CFRFS
One of the deliverables of the CFRFS project is a component-based toolkit for
flames using detailed chemical mechanisms. However, as various parts of the
Toolkit get completed, we will release them to the general public. Since the
validation of these software pieces are time-consuming, as are the peer-review of
the associated papers (to say nothing of Sandia's and DoE's software licensing
process!), these releases will lag the actual state-of-affairs.
Each of the pieces released will have their separate pages, which will also
contain their subsequent releases / bug fixes. Those pages will also contain
the listing of papers (algorithms and validations) associated with the software.
This page will simply serve as a listing of these separate pieces.
HODIF
HODIF (High Order Derivatives, Interpolations and Filters) is a software library
of stencils that can be used with block-structured adaptively refined meshes.
It will contain finite-difference stencils in 2D and 3D for evaluating derivatives,
as well as for inter-level interpolations (prolongations and restrictions). These
are required if one intends to perform high-order (fourth and higher) computations
on AMR meshes. Note that the high-order is spatial only. Also, we stipulate that
the level-to-level refinement ratio is 2. Filters provided here are
sometimes required to remove the high (unresolved) wavenumber content from the
solution; the lack of numerical dissipation makes high-order stencils rather
susceptible to instability and/or Runge phenomenon.
-
Release 0.1 , April 2006 : This release contains software that is
applicable to vertex-centered meshes (variables are defined at cell
corners). Upwinded derivative stencils are included. The software is
limited to 2D (3D has yet to be tested) and collocated derivatives i.e.
derivative are evaluated at the points where the variables are defined.
Staggered derivatives are being exercised now.
CFRFS Toolkit
A very preliminary version of the toolkit is nearing completion. While some of its
capabilities will not be advertised (validation not complete), it should be able
to perform on uniform meshes reliably, solve reaction-diffusion problems
with detailed chemistry (at least GRI 1.2) using high-order spatial discretizations,
extended stability Runge-Kutta methods and be parallel. We are currently working
on replacing CHEMKIN (used today) with
CANTERA so that we may freely distribute the software.
Any help with this will be very appreciated !!!
Created by Jaideep Ray