The hRefConst and eRefConst thermos that were local to
reacting*EulerFoam have been removed and the reference state that they
used has been incorporated into the standard hConst and eConst thermos.
The hConst thermo model now evaluates the enthalpy like so:
Ha = Hf + Hs
= Hf + Cp*(T - Tref) + Hsref (+ equation of state terms)
Where Ha is absolute enthalpy, Hs is sensible enthalpy, Cp is specific
heat at constant pressure, T is temperature, Tref is a reference
temperature and Hsref is a reference sensible enthalpy. Hf, Cp, Tref and
Hsref are user inputs. Of these, Tref and Hsref are new. An example
specification is as follows:
thermodynamics
{
Hf -1.34229e+07;
Cp 2078.4;
Tref 372.76;
Hsref 128652;
}
The ref quantities allows the user to specify a state around which to
linearise the relationship between temperature and enthalpy. This is
useful if the temperature range of the simulation is small enough to
consider the relationship linear, but linearity does not hold all the
way to standard conditions.
To maintain backwards compatibility, Tref defaults to standard
temperature, and Hsref defaults to zero, so a case using hConst thermo
requires no modification as a result of this change.
The only change to the default operation is that to calculate sensible
enthalpy Cp is multiplied by the difference between the current
temperature and the standard temperature, whether as previously Cp was
multiplied by the current temperature only. This means that at standard
conditions sensible enthalpy is now zero, and absolute enthalpy equals
the formation enthalpy. This is more consistent with the definitions of
the various enthalpies, and with other thermo models such as janaf. This
change should only affect reacting cases that use constant thermo
models.
The total enthalpy is calculated as
Ha = ha + K
where
ha is absolute enthalpy
K is the kinetic energy: 1/2*magSqr(U)
The total enthalpy or a particular phase can be calculated by specifying the
optional "phase" name, e.g.
#includeFunc totalEnthalpy(phase = liquid)
The ThermalPhaseChangePhaseSystem stores the thermal phase change dmdt
used in the previous continuity error update and uses that to stabilize
the interfacial heat transfer calculations when phase fractions approach
zero.
Patch contributed by Juho Peltola, VTT.
continuityError is now just the transport inconsistency. Mass sources,
whether as a result of fvOptions or phase-transfer/change processes, are
not included.
This is particularly useful for multiphase simulations for which integrating the
density weighted phase properties also requires the phase fraction to be
including in the weighting.
A single weight field can be specified as before:
weightField rho;
or a list specified by:
weightFields (alpha.water rho.water);
A number of file name patterns have been removed from the list of things
that cleanCase deletes. Some patterns related to obsolete files that
OpenFOAM no longer generates, and some were deemed too generic to
delete as they might contain important persistent information.
Face centres are calculated by area-weighting the centres of triangles
that are formed by connecting each edge to a common point. The common
point is taken to be the average of all the face vertices, though in
principle the choice is arbitrary.
The areas used to perform the weighting are now taken to be the
projected areas of the triangles in the direction of the face normal
(i.e., the dot product between the triangle area and the face normal).
Previously the triangle area-magnitudes were used.
The new approach results in a centre that for flat faces is independent
of the choice of common point. It also means that concave faces have
contributions to the weighted sum from reversed triangles correctly
subtracted from the total. The centre of warped faces still changes with
the choice of common point, but that variation is now reduced to be
only in the direction of the face normal.
Preliminary results suggest a positive effect of this change on the
convergence of simulations on meshes with significantly distorted faces
and cells. The simpleFoam motorBike tutorial now converges with
residuals approximately half that previously observed.
Resolves bug report https://bugs.openfoam.org/view.php?id=1993
Updated the continuity error compensation term in the face momentum formulation
so that separate flow and source continuity errors are combined into a single
term.
Rotational transforms can now be automatically calculated from the
coupled patch geometries even when a) the patches are non-planar, b)
the patch normals are not perpendicular to the rotation axis, and/or c)
the rotation angle is 180 degrees.
This case is an updated version of
tutorials/multiphase/multiphaseEulerFoam/damBreak4phase using the latest models
available in reactingMultiphaseEulerFoam for interface capturing.
The utilised static parts of polyMeshGeometry are now part of a
polyMeshCheck namespace. Everything else has been removed, as they were
unused, out of date, and/or duplicated elsewhere.
This operator was defined with an incorrect (or at least
counter-intuitive) ordering. Given the possibility for ambiguity
associated with transformation ordering it is considered preferable for
the transformer class not to have compound assigment operators. Forcing
the full syntax (e.g., "a = a & b" or "a = b & a") makes it clear in
which order the transformations are intended to be applied.
There is now an optional tauy entry in the thixotropicCoeffs to specify the
Bingham yield stress, e.g.:
filmViscosityModel thixotropic;
thixotropicCoeffs
{
mu0 100;
muInf 0.1;
a 0.1;
b 3;
c 0.005;
d 0.9;
tauy 1.1; // Yield stress in Pa
}
