Don't save things at all, and use a ``writeObject`` FO to save only the variables you want at the interval you want?

Yeah but what's the one runtime setting to do "don't save anything at all"

Use a ``writeInterval`` bigger than your ``endTime`` in ``controlDict``?

Crazy hack

Until you regret not saving it because you wanted to restart your case!

Still possible to play it safe and still fully write only the last timestep

I'm currently simulating a gas-liquid stirred tank with two different impellers. I'm having trouble implementing a tracer for mixing time and calculating hold-up. Unfortunately, I'm using STAR-CCM+ and can't switch software. Does anyone here have expertise with STAR-CCM+, or know of another forum or server where I could get help?

Sir, this is a OpenFOAM server

If you paid for a STAR license, you can ask them for help

license? whats that?

Whats your point here?

Well, In Elementary FluidMech they will explain you that the general state of stress on a control volume at rest is one in which the pressure (the component of stress normal to a CVs surface) is a constant at equal heights and points towards the volume, by convention.

That's why your two forces act "against the surface"

when you are in thermodynamics and/or intermediate fluid mechanics they will explain you that perfect fluids cannot be in a state of tension

Man, download White's Solution Manual and stop spamming this channel

The whole chatgroup is called "OpenFOAM", not "White's Fluid Mechanics Solution Manual Group"

this isn't really a homework help server though

your classmates are probably a better resource

just go talk to them

Ok, well please could you tell me a better server to get help from? I'm a maths student, but the module is taught by bioengineering, which is on a completely different campus to the maths department (i.e. on the other side of London). Because of this, I don't have any classmates to ask so the internet is the best resource I have

You have teachers, TA's, 1000's of books, Solution Manuals, etc... If you rely on random chat servers for your education, I have some news for you: stack exchange

Hi, I am solving an equation with OpenFoam, however when i use lower values for 'k' which represents the thermal conductivity, it starts either not converging, or when k is even lower it immediatly crashes with a floating point error... any ideas for making this convection dominated problem converge? Either via changing my formulation of the equation or by tweaking the solvers/preconditioners?

const volVectorField& U = primalVars_.UInst();
surfaceScalarField phiBeta("phibeta", fvc::flux(U*(1-beta)));

// Adjoint energy equation
fvScalarMatrix TaEqn
(
     rho_fluid * c_p_fluid * fvm::div(-phiBeta, Ta)
      - fvm::laplacian(k, Ta) 
);

Current solvers:

"Ta.*"
    {
        solver          PBiCGStab;
        preconditioner  DILU; 
        tolerance       1e-09;
        relTol          0;
    }

If i use upwind phi instead of linear interpolation it doesnt crash any more but the solution still diverges via some weird checkerboard thing

Does anyone have a rough Idea of if a 5700X AMD CPU with 32 Gigs ram or Nvidia a200 laptop with 4gb Vram and 64 g memory what would run simple laminar flow high cell count simulations

Hi everyone, I’m running a wind turbine simulation in OpenFOAM v2312 using the k-omega SST turbulence model with steady-state (simpleFoam) and MRF. The issue I’m facing is that the residuals for omega and k remain unrealistically high or diverge, even though the velocity (U) and pressure (p) residuals seem reasonable. Here are the key setup details: • OpenFOAM version: v2312 • Solver: simpleFoam • Turbulence model: k-omega SST • Mesh: generated via snappyHexMesh; boundary layers included; checkMesh is mostly clean except for 1–2 faces with high non-orthogonality that I can’t remove even after further refinement • Wall treatment: using standard wall functions (omegaWallFunction, kqRWallFunction, nutkWallFunction) • Boundary conditions: • Inlet: fixed velocity; k and omega computed from 5% turbulence intensity and estimated length scale • Outlet: zeroGradient for all fields • Walls (blades, hub, etc.): noSlip with appropriate wall functions • Relaxation factors: • p: 0.4 • U: 0.8 • k, omega: 0.3 Despite trying different under-relaxation factors, and verifying initial and boundary conditions, the residuals for k and omega are not stabilizing. Any insights into what could be causing this? Has anyone dealt with similar issues, especially in wind turbine or rotating frame setups? I’m happy to share snippets of my case files or logs if needed. Thanks in advance for your help!

You might want to try playing around with some of the turbulence model parameters in the turbulenceProperties file, if it has any

I would not do that. These are constants that are calibrated so the model recreates certain "calibration" flow setups properly and once you mess with them (when you have no data to base it on) you loose the (assumed) validity of your turbulence model.

Are you properly resolving all the gradients? In case of localy strong gradients that werent resolved I have had the same issues . And when that is the case the standard approach of using more stabil numerics doesnt work since you further reduce the resolution of your numerics making the problem worse