Moved all fields and field reading into the DsmcCloud, all calculation and
resetting to single functions for all fields.
Changed constructors so that no fields are supplied to the solver called from
dsmcFoam and an initialisation dictionary is supplied by dsmcInitialise.
Adding check when inseting a new vertex on a long Delaunay edge that the edge
does not pierce the surface, causing a span between thin parts of the surface.
Contains temporary checking code for internalPoints that are not inside the
surface.
Was passing points by copy rather than const reference to conformationSurfaces
...Intersection queries, rectified.
Modification to relaxation function produce the desired first and last values.
Moved writeInternalDelaunayVertices back to being called from writeMesh, but
option to write to constant.
Added function to limit vertex displacements that penetrate the surface or come
to close to it (closer than twice the local point pair distance) by recursively
halving the displacement vector.
Using runTime.loop() construct in app for time loop - with runTime++ at the end
of the loop - wasn't getting the polyMesh written to the final timestep. Moving
targetCellSize and internal delaunay vertex writing out of writeDual function
and calling them from the move() function.
- first stab at a simple runtime selection for extra turbulence-related
features
- the code duplication seems to indicate that a common ancestor
for compressible and incompressible turbulence models could be good
- thermalDissipation()
corresponds to the energy lost due to viscous efffects and
what exits the energy cascade via dissipation.
- thermalDissipationEff()
corresponds to the energy lost due to effective viscous efffects.
Everything that is lost from momentum. Thus essentially assumes
turbulent equilibrium, but is what STAR-CD and Fluent seem to be using.
Thus even if it's wrong, provide it anyhow.
- minor consistency update in comments
