[liybzd]

Hi Foamers,

I'm using groovyBC for oscillatory flow in a simple rectangular box, it's only a test.
Six boundaries in this case, front, back, top and bottom are all slip.
inlet and outlet are set to groovyBC. The solver is pisoFoam.

The velocity filed on left and right are set as:

inlet
{
type groovyBC;
valueExpression "vector(Um*cos(ome*time()),0,0)";
variables "Um=1.95;ome=2*pi/5.5;";
timelines ();
value uniform (1.95 0 0);
}
outlet
{
type groovyBC;
valueExpression "vector(Um*cos(ome*time()),0,0)";
variables "Um=1.95;ome=2*pi/5.5;";
timelines ();
value uniform (1.95 0 0);

In this case, U=1.95*cos(ome*t).

Since -dp/dx=dU/dt=-1.95*ome*sin(ome*t), so the BC of pressure filed is set as:
inlet
{
type groovyBC;
gradientExpression "-Um*ome*sin(ome*time())";
variables "Um=1.95;ome=2*pi/5.5;";
timelines ();
fractionExpression "0";
value uniform 0;
}
outlet
{
type groovyBC;
gradientExpression "Um*ome*sin(ome*time())";
variables "Um=1.95;ome=2*pi/5.5;";
fractionExpression "0";
timelines ();
value uniform 0;
}


Both velocity and pressure filed blows up in a few seconds, I don't know if there is mistake in it...

[gschaider]

Originally Posted by liybzd

Hi Foamers,

I'm using groovyBC for oscillatory flow in a simple rectangular box, it's only a test.
Six boundaries in this case, front, back, top and bottom are all slip.
inlet and outlet are set to groovyBC. The solver is pisoFoam.

The velocity filed on left and right are set as:

inlet
{
type groovyBC;
valueExpression "vector(Um*cos(ome*time()),0,0)";
variables "Um=1.95;ome=2*pi/5.5;";
timelines ();
value uniform (1.95 0 0);
}
outlet
{
type groovyBC;
valueExpression "vector(Um*cos(ome*time()),0,0)";
variables "Um=1.95;ome=2*pi/5.5;";
timelines ();
value uniform (1.95 0 0);

In this case, U=1.95*cos(ome*t).

Since -dp/dx=dU/dt=-1.95*ome*sin(ome*t), so the BC of pressure filed is set as:
inlet
{
type groovyBC;
gradientExpression "-Um*ome*sin(ome*time())";
variables "Um=1.95;ome=2*pi/5.5;";
timelines ();
fractionExpression "0";
value uniform 0;
}
outlet
{
type groovyBC;
gradientExpression "Um*ome*sin(ome*time())";
variables "Um=1.95;ome=2*pi/5.5;";
fractionExpression "0";
timelines ();
value uniform 0;
}


Both velocity and pressure filed blows up in a few seconds, I don't know if there is mistake in it...

Usually setting the velocity on all patches (which you basically do) is not a good idea as the slightest inaccuracies will blow up an incompressible solver (compressible too, but it will take longer).

As it seems to take only a few timesteps to blow up I'd suggest that you write out every timestep and have a look at them. Usually then it becomes painfully clear what the problem is. If it takes a bit longer you can use purgeWrite in the controlDict to keep only the last few.

Another problem can be that you only specify Neuman-conditions for the pressure. Usually OF then needs a reference-pressure from you. I'm not 100% sure but as groovyBC is a mixed-BC it is possible that OF says "Ah. there's a Dirichlet-type BC here. Don't need a reference pressure". Remedy here would be to fix the pressure on one boundary

[dewey]

Hi, i want to simulate the pulsatile flow in the brain, and i took the syntaxis from here.


We define the vector velocity with his components, but I want to use a velocity normal to the inlet surface or flow rate, because my inlet geometry is not only in one plane.


could you help me please!







inlet

{
type groovyBC;
valueExpression "vector(R*(1+(A1*cos(2*1*pi*ome+B1))+(-A2*cos(2*2*pi*ome+B2))+(A3*cos(2*3*pi*ome+B3))+(A4 *cos(2*4*pi*ome+B4))+(A5*cos(2*5*pi*ome+B5))),0,0) ";
variables "A1=0.29244;A2=0.5908;A3=0.2726;A4=0.198;A5=0.1124 ;B1=4.027;B2=6.509;B3=1.913;B4=1.461;B5=0.074;Te=0 .75;R=0.3324;ome=time()/Te;";
timelines ();
value uniform (0.093 0 0);
}

[gschaider]

Originally Posted by dewey

Hi, i want to simulate the pulsatile flow in the brain, and i took the syntaxis from here.


We define the vector velocity with his components, but I want to use a velocity normal to the inlet surface or flow rate, because my inlet geometry is not only in one plane.


could you help me please!







inlet

{
type groovyBC;
valueExpression "vector(R*(1+(A1*cos(2*1*pi*ome+B1))+(-A2*cos(2*2*pi*ome+B2))+(A3*cos(2*3*pi*ome+B3))+(A4 *cos(2*4*pi*ome+B4))+(A5*cos(2*5*pi*ome+B5))),0,0) ";
variables "A1=0.29244;A2=0.5908;A3=0.2726;A4=0.198;A5=0.1124 ;B1=4.027;B2=6.509;B3=1.913;B4=1.461;B5=0.074;Te=0 .75;R=0.3324;ome=time()/Te;";
timelines ();
value uniform (0.093 0 0);
}

In groovyBC the function normal() gives you the unit vector normal to each patch face. So by multiplying it (watch the sign) with your velocity you should get what you need

[dewey]

Thrank you my friend, now I am trying this:


{
type groovyBC;
valueExpression "R*(1+(A1*cos(2*1*pi*ome+B1))+(-A2*cos(2*2*pi*ome+B2))+(A3*cos(2*3*pi*ome+B3))+(A4 *cos(2*4*pi*ome+B4))+(A5*cos(2*5*pi*ome+B5)))";
variables "A1=0.29244;A2=0.5908;A3=0.2726;A4=0.198;A5=0.1124 ;B1=4.027;B2=6.509;B3=1.913;B4=1.461;B5=0.074;Te=0 .75;R=0.3324*normal();ome=time()/Te;";
timelines ();
value uniform (0.093 0 0);
}



I am multiplying the R*normal(), but could you explainme what the sign mean?


What will be the result with (-) or (+), I will try both.

[gschaider]

Originally Posted by dewey

Thrank you my friend, now I am trying this:


{
type groovyBC;
valueExpression "R*(1+(A1*cos(2*1*pi*ome+B1))+(-A2*cos(2*2*pi*ome+B2))+(A3*cos(2*3*pi*ome+B3))+(A4 *cos(2*4*pi*ome+B4))+(A5*cos(2*5*pi*ome+B5)))";
variables "A1=0.29244;A2=0.5908;A3=0.2726;A4=0.198;A5=0.1124 ;B1=4.027;B2=6.509;B3=1.913;B4=1.461;B5=0.074;Te=0 .75;R=0.3324*normal();ome=time()/Te;";
timelines ();
value uniform (0.093 0 0);
}



I am multiplying the R*normal(), but could you explainme what the sign mean?


What will be the result with (-) or (+), I will try both.

Exactely. I keep forgetting whether normal() points in or out. So I leave finding the correct sign as "an exercise for the reader" (that is the usual formulation if a textbook author isn't sure about the details)