View Issue Details

IDProjectCategoryView StatusLast Update
0002617OpenFOAMContributionpublic2017-07-17 16:49
Reporterhuberta Assigned Tohenry  
PrioritynormalSeverityminorReproducibilityalways
Status resolvedResolutionfixed 
PlatformGNU/LinuxOSUbuntuOS Version15.04
Product Versiondev 
Fixed in Versiondev 
Summary0002617: BuoyantPimpleFoam crashes when no pressure is fixed
DescriptionI tried to run the hotroom tutorial with buoyantPimpleFOAM just changing all the BC to walls but using Boussinesq linear approximation to calculate the density. This case would then look similar to the hotroom tutorial available for the solver buoyantBoussinesqPimpleFOAM with BC changed to wall everywhere. However, the simulation crashes when running it with buoyantPimpleFOAM while it does not with buoyantBoussinesqPimpleFOAM
Steps To ReproduceA way to solve the problem is to set a reference pressure in a fixed point for the buoyantPimpleFOAM solver in the createFields.H file, the same way as it is done for the buoyantBoussinesqPimpleFOAM i.e.:

label pRefCell = 0;
scalar pRefValue = 0.0;
setRefCell
(
    p,
    p_rgh,
    pimple.dict(),
    pRefCell,
    pRefValue
);

if (p_rgh.needReference())
{
    p += dimensionedScalar
    (
        "p",
        p.dimensions(),
        pRefValue - getRefCellValue(p, pRefCell)
    );
}
Additional InformationBy doing so, in the case the pressure in not fixed via a pressure inlet BC or outlet, it would help to stabilize the calculation and prevent it from crashing
TagsNo tags attached.

Activities

henry

2017-07-17 14:36

manager   ~0008403

Setting a reference pressure is only correct if the flow is incompressible as is the case with theBoussinesq equation of state. Have you tried running with your "fix" with the perfect gas or other compressible equations of state?

huberta

2017-07-17 15:17

reporter   ~0008404

you are right I was more focused on the incompressible case. When run with any other model (except polynomial) it does not need the "fix" suggested and the fix even makes the calculation crash in these cases.

So to sum up, the solver works fine with the set-up above without any modification. It is just when dealing with polynomial properties or Boussinesq approximation the "fix" suggested is needed

henry

2017-07-17 15:22

manager   ~0008405

Can you provide a patch will work for any equation of state?

henry

2017-07-17 16:49

manager   ~0008406

Resolved by commit 40ddc7204530d7724d31e04edf660707965e4c0e

Issue History

Date Modified Username Field Change
2017-07-17 14:25 huberta New Issue
2017-07-17 14:36 henry Note Added: 0008403
2017-07-17 15:17 huberta Note Added: 0008404
2017-07-17 15:22 henry Note Added: 0008405
2017-07-17 16:49 henry Assigned To => henry
2017-07-17 16:49 henry Status new => resolved
2017-07-17 16:49 henry Resolution open => fixed
2017-07-17 16:49 henry Fixed in Version => dev
2017-07-17 16:49 henry Note Added: 0008406