[fmerk]

Hello foamers,

I'm working on a simulation where the velocity upstream at the inlet is unknown. To solve the incompressible flow, I'm working with interFoam using Openfoam version v1612+. I'm applying the totalPressure boundary condition. It's defined as:

p_p = p_0 - 0.5*u^2

where I insert rho*g* z coordinate of flow height at inlet for p_0.

Now my question is:

- Openfoam doesn't complain about my numerical setup, yet from a physical point of view, using the hydrostatic pressure rho*g*z for p_0 in the totalPressure boundary (at the inlet) doesn't make sense, does it? I'd say the pressure is underestimated.

- If I impose the totalPressure BC at the inlet patch, is the p_rgh field for this patch then representing the total pressure (as defined by e.g. Bernoulli)? Usually, p_rgh is defined as p_rgh = p - rho*g*h(z), so not necessarily the total pressure.

- Displaying the results in paraview implies another complication for me:
the mean p_rgh(inlet patch value) is 19033 for one particular test case. If I add the dynamic pressure (0.5*rho*u^2), I end up with 21091.5 which I initially set p_0 for the totalPressure BC (as hydrostatic pressure with 9810*2.15m). So, is the p_rgh field representing the hydrostatic share of the pressure distribution?

- The resulting velocity, which then determines the dynamic pressure, is also mysteriously arising for me. Where does the U come from? I'm using a pressureInletOutletVelocity boundary condition for my U field. There, the velocity is calculated via the flux phi.

How are the totalPressure BC and the velocity connected? If someone could illuminate this context to me, I'd really be happy. If i missed some obvious explanation somewhere in the forum or in one of the comprehensive theses by jasak or ruschke, just point it out!

kind regards

[student666]

Hi,

for a solver like simpleFoam, when you set a totalPressure bc on a patch (outlet) you should set a velocity bc as well.
The other patch (inlet) should be set with a fixed pressure and a outletInlet bc that could be:

    {
        type            outletInlet;
        phi             phi;
        outletValue     uniform 0;
        value           uniform 0;
    }

This closes your problem.
The solver uses the velocity to calculate p(static) field at the outlet and with the p fixed at inlet it can solves for the velocity at the inlet and satisfy the Bernoulli equation.
(truly speaking i don't think it does this way, but by hand calculation it is what I do)

I think that interFoam does the same, but it's supposed to take a further step as for g field that is a "sort of source" (body forces) for the p field as for different density of the different phases....I think that phisically speaking it is what it does...and it should behave like simpleFoam + further corrections...

You can test totalPRessure bc and better understand its behaviour by testing a simple channel case with an easier solver as icoFOAM.

Regards.