[student666]

Hi all,

I’m trying to better understand how to set properly bcs for pressure and velocity on a 2D laminar case (rectangular channel 3x0.3x0.1 ), using OF v4.1, with fanPressure bc on the outlet patch.
Test case is as for enclosed picture.

fanPressure.C source code is

// Average volumetric flow rate
    scalar volFlowRate = 0;
    if (phi.dimensions() == dimVelocity*dimArea)
    {
        volFlowRate = dir*gSum(phip);
    }
    else if (phi.dimensions() == dimVelocity*dimArea*dimDensity)
    {
        const scalarField& rhop =
            patch().lookupPatchField<volScalarField, scalar>(rhoName());
        volFlowRate = dir*gSum(phip/rhop);
  }

I set the following flowrate/pressureDrop fanCurve.

5((0.0005 6)(0.001 5)(0.0017 4.5)(0.0022 4)(0.0028 1))

As fanPressureBC sets a pressure condition on outlet, I set for U the following BC:

    inlet
    {
        type           outletInlet;
        phi            phi;
        outletValue    uniform (0 0 0);
        value          uniform (0 0 0);
    }
    outlet
    {
	type            pressureInletOutletVelocity;
	value		uniform (0 0 0);
    }

and this is p file:

    inlet
    {
        type            fixedValue;
	value		uniform 0;
    }

    outlet
    {
        type            fanPressure;
        fileName        "fanCurve";
        outOfBounds     clamp;
        direction       out;
	U		U;
	phi		phi;
//	psi		none;
//	rho		rho;
        p0              uniform 0;
        value           uniform 0;
    }

when I run simpleFoam for 1000iterations I have the following last log:

Time = 1000

smoothSolver:  Solving for Ux, Initial residual = 1.512615e-05, Final residual = 5.2722318e-08, No Iterations 1
smoothSolver:  Solving for Uy, Initial residual = 8.228806e-05, Final residual = 2.0212224e-07, No Iterations 1
GAMG:  Solving for p, Initial residual = 0.00017614954, Final residual = 1.6843817e-05, No Iterations 1
time step continuity errors : sum local = 7.2177518e-06, global = 1.5811468e-07, cumulative = -15.385409
ExecutionTime = 295.05 s  ClockTime = 308 s

and the residual is attached (run1)

this is the postProcess -func 'patchAverage(name=outlet,U)' output for latestTime:

Time = 1000
Reading fields:
    volVectorFields: U
Executing functionObjects
surfaceRegion patchAverage(name=outlet,U) write:
    average(outlet) of U = (50.051912 -6.985344e-05 0)

surfaceRegion patchAverage(name=outlet,U) write:
    average(outlet) of U = (50.051912 -6.985344e-05 0)

Impossible: Out Of Bound Solution!

If I reverse outlet with inlet, so in order to have the fan to blow inside the channel with following BC:
U

    inlet
    {
        type           inletOutlet;
        phi            phi;
        inletValue     uniform (0 0 0);
        value          uniform (0 0 0);
    }
    outlet
    {
	type            pressureInletOutletVelocity;
	value		uniform (0 0 0);
    }

and for p:

    inlet
    {
        type            fixedValue;
	value		uniform 0;
    }

    outlet
    {
        type            fanPressure;
        fileName        "fanCurve";
        outOfBounds     clamp;
        direction       in;
	U		U;
	phi		phi;
//	psi		none;
//	rho		rho;
        p0              uniform 0;
        value           uniform 0;
    }

I get following outoput for postProcess -func 'patchAverage(name=inlet,U)':

Time = 1000

Reading fields:
    volVectorFields: U

Executing functionObjects
surfaceRegion patchAverage(name=inlet,U) write:
    average(inlet) of U = (-1.3472552 -1.511e-11 0)

surfaceRegion patchAverage(name=inlet,U) write:
    average(inlet) of U = (-1.3472552 -1.511e-11 0)

Reasonable...even if you calculate 1.347*0.3*0.1=4,02×10⁻² m3/s (still out of bound, but less...)

So the question is, how to make this BC works for a suction fan? what am I missing?

Thanks.

Michele

[student666]

Hi all,

I performed some test.

In summary, I increased the mesh resolution, but my opinion is that average velocity is still out of bound.
As according to the fan curve I set, the resuting velocity cannot exceed 0.0028 [cm/s] /0.3 [m]/0.1 [m] = 0.093 m/s.

I also tried to change fvScheme, using Van Leer limited for U divScheme e.g. but I still have too high velocity.

Am I doing right?