(Wind-US Documentation Home Page)
(Wind-US User's Guide)
(GMAN User's Guide)
(MADCAP User's Guide)
(CFPOST User's Guide)
(Common File User's Guide)
(Wind-US Installation Guide)
(Wind-US Developer's Reference)
Wind 4.0 --> Wind 5.0 Changes
The following list summarizes the principal changes made between
Wind Versions 4.0 and 5.0.
The number in parentheses following the description (e.g., 5.16) is
the specific version in which the change was introduced.
The following changes have been made to the Wind code.
- The RHS keyword
has been restructured, and some new combinations of differencing
scheme and procedure are now available.
In particular, the van Leer scheme can now use the higher order
coupling methods which were previously limited to the Roe scheme
The following table lists the various combinations of RHS
keyword options available using the old and new syntax.
Following the usual
for displaying keyword syntax, optional keyword parameters
are inside square brackets.
||ROE FIRST [UPWIND]
||ROE FIRST PHYSICAL
||ROE [SECOND [PHYSICAL]]
||ROE SECOND CENTRAL
||ROE SECOND UPWINDBIASED
||ROE SECOND UPWIND
||ROE THIRD [UPWINDBIASED]
||ROE THIRD UPWIND
||ROE FOURTH [UPWINDBIASED]
||ROE FOURTH CENTRAL
||ROE FIFTH [UPWINDBIASED]
||VAN LEER FIRST||
||VAN LEER FIRST [UPWIND]
||VAN LEER CENTRAL||
||VAN LEER SECOND CENTRAL
||VAN LEER MIXED||
||VAN LEER SECOND UPWINDBIASED
||VAN LEER SECOND||
||VAN LEER SECOND UPWIND
||VAN LEER THIRD||
||VAN LEER THIRD [UPWINDBIASED]
||VAN LEER THIRDU||
||VAN LEER THIRD UPWIND
||VAN LEER FOURTHU||
||VAN LEER FOURTH [UPWINDBIASED]
||VAN LEER FOURTHC||
||VAN LEER FOURTH CENTRAL
||VAN LEER FIFTH||
||VAN LEER FIFTH [UPWINDBIASED]
With two exceptions, the syntax used in Wind 4.0 will still
work, but is no longer recommended or documented.
The exceptions are ROE SECOND and VAN LEER SECOND.
In Wind 4.0, these give second-order Roe and Van Leer upwind,
They now give second-order Roe and Van Leer upwind-biased,
with modifications for stretched grids. (5.16)
- The HLLE explicit operator is now available as one of the choices
with the RHS
With the HLLE scheme, the
may be used to control the behavior of the anti-diffusion term.
- The restriction on number of grid points required for
sequencing has been relaxed. (5.18)
- A HOLD keyword
has been added, allowing the user to specify whether total
conditions or characteristics should be held at
The default for ideal gases is to hold total conditions, which
is different from earlier versions. (5.30)
In addition, the default option for ideal gases at
boundaries is now HOLD_TOTALS. (5.82)
- The time step is now recomputed at the start of each iteration,
unless Newton time stepping is being used, in which case it is
computed at the beginning of each Newton time level.
The previous behavior, in which the time step is recomputed at
the beginning of every cycle (even with Newton time stepping)
may be obtained using
option 106. (5.45)
- An underflow limiter has been added to the algorithm used with
to compute the new temperature at boundaries where total
conditions are being held fixed. (5.50)
- The time history
tracking capability has been improved.
- Time history data may now be written into the .cth
file when running in parallel mode.
The TEST 125 option, formerly used to write time
history data into the .lis file in parallel mode,
has therefore been eliminated.
- Many additional variables may now be tracked.
- Restart cases will now append to an existing .cth
file, unless changes have been made to the
keyword input parameters or to the reference conditions.
- Because of changes in the .cth file format, the new
must be used when post-processing a new .cth file.
The improved capability required some changes to the shell
scripts used to run the Wind code.
The modified scripts are wind.sh,
wind_input.sh, wind_jobfile.sh, and
wind_miscfns.sh, and the new versions were uploaded
to IVMS on May 30, 2001.
If you're currently using older versions of these scripts,
updated versions must be installed in order to use the improved
time history tracking.
For someone with user-level access to IVMS, this can be done
by downloading all the scripts and copying these four into
your CFDROOT/bin directory, or by downloading and
installing the current Wind application distribution. (5.52)
- Newton iteration
now works for parallel runs with
mode shared and
- A relaxation factor may now be used with the
and IMPLICIT GAUSS keywords. (5.64)
- A problem with the use of the IJK_RANGE keyword in the
keyword block, for flows with reacting or frozen chemistry, has
been fixed. (5.66)
- In the .lis file, headers are now printed only when
residuals are output. (5.74)
- A DIRECTION keyword has been added to the
keyword block, allowing the user to easily specify that the
incoming flow is normal to the boundary face, or along the
grid lines intersecting the boundary face.
This replaces TEST option 176, which has therefore
been removed. (5.80)
- A BLOW SURFACE capability has been added to the
allowing the blowing velocity to be computed, or set to zero,
on a point-by-point basis, based on the specified plenum total
pressure and the local flowfield static pressure.
In addition, the BLOW VALVE capability has been fixed.
- The memory allocation procedure for the flow solution phase
has been restructured. (5.87)
- The DOWNSTREAM
PRESSURE keyword now allows specification of either
zeroth- or first-order extrapolation, and a relaxation option
was added for spatially varying downstream pressures.
FLOW keyword also now allows specification of the
extrapolation order, and an option was added to modify the mass
flow directly instead of iterating the downstream pressure.
- Bleed boundaries are now treated as no-slip walls for the
solution of the turbulence modeling equations, as well for the
The previous behavior, treating bleed boundaries as slip walls
when solving the turbulence model equations, may be obtained using
option 67. (5.101)
- Several fixes have been made involving the
use of CGNS files
in Wind. (5.25, 5.83, 5.105, 5.137, 5.140)
- Options to use the Koren and van Albada limiters have been added
to the TVD
- The time step calculation procedure from the OVERFLOW code has
been added as an option with the
- An alternative implementation of the Roe scheme, from the
OVERFLOW code, has been added as an option with the
- The ARC3D 3-factor diagonal scheme, as implemented in OVERFLOW,
has been added as an option with the
- A VORTEX GENERATOR
keyword has been added, for modeling the effects of vortex
generator arrays on the flowfield. (5.112)
- A TDA_INVALID
keyword has been added, allowing the user to reset boundary data
in the .tda file. (5.122)
- Various bugs in the singular
axis and pinwheel
boundary conditions have been fixed.
The algorithm was also modified slightly to always average the
pressure and not the total energy (in addition to the density,
and momentum or velocity components). (5.127)
- The algorithm used with the
keyword has been fixed, and the differencing techniques now
match what was in the old NPARC code, which seems to work better
near boundaries. (5.128)
- Various tweaks have been made to get
working properly, although there is apparently still a problem
when the ARBITRARY
INFLOW keyword is used with Liu-Vinokur chemistry. (5.134)
The following changes have been made to the Makefiles.
- Added and/or modified makefiles for IBM RS6K and Compaq alpha
The following changes have been made to the scripts.
- Modifications to allow use under Cygwin on a Windows system
- Added support for Compaq alpha systems (2.0.44)
The following changes have been made to the utilities.
- Added support to CFPOST for Compaq alpha systems
- Added new thplt
utility for post-processing the new time history files
- The download, build, and installation process for the
utilities has been improved
Last updated 4 Nov 2002