[jet_engine]

I was trying to run the following test case: 2D flow over flat plate laminar.

The case is pretty simple, it's just a 1x1m flat plate with laminar flow, which I was planning to use to show how the velocity profile matches with the Blasius solution (by plotting U/U_inf vs. eta).

The thing is, when I run it, there is a region of the flow which has a UMagnitude over the freestream velocity. The solver used is simpleFoam, but I've also tried using icoFoam with this other test case: Blasius Flat-Plate Flow Benchmark. The results are the same, UMagnitude goes over the freestream velocity.

The only change I made in the files is updating the transportProperties file as I'm using OF v2112:

turbulenceProperties:

FoamFile
{
    version     2.0;
    format      ascii;
    class       dictionary;
    location    "constant";
    object      turbulenceProperties;
}
// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * //

simulationType        laminar;

turbulence      off;

printCoeffs     off;

I have attached two images to show the results, one for Re 100 and the other one for Re 200.

Does anyone know what I'm doing wrong? Thanks in advance.

[piu58]

I think it is normal what you get.

If your inlet velocity is constant, and the velocity reduces at the boundary, it has to increase somewhere else.

The test case asumes a plate in an very large volume. You should come closer to that if you expand your area in y direction.

[jet_engine]

Originally Posted by piu58

I think it is normal what you get.

If your inlet velocity is constant, and the velocity reduces at the boundary, it has to increase somewhere else.

The test case asumes a plate in an very large volume. You should come closer to that if you expand your area in y direction.

Thanks for the suggestion! I will try to expand the domain in the y direction then.

Is there any better way or case setup to compare the results with the similarity solution besides increasing the Reynolds number a bit?

Thanks.

[piu58]

Originally Posted by jet_engine

Thanks for the suggestion! I will try to expand the domain in the y direction then.

Is there any better way or case setup to compare the results with the similarity solution besides increasing the Reynolds number a bit?

Thanks.

You me set the upper side as an outlet and not as a wall.

[Kyouma]

Hi,

I am new to using OpenFOAM and wanted to add my experience and concerns so far in carrying out a verification and validation study on the solvers ability to reproduce the Blasius analytical solution. I have a 35 mm plate in a 500 mm high domain with an entrance length (before plate leading edge) of 20 mm. I am using a velocity of 0.2 m/s and a viscosity of nu = 1.0E-06 and 1.6E-05. I noticed that for the lower Reynolds case (using the higher viscosity) that the magnitude of acceleration seen over the boundary layer is larger. I thought maybe I could adjust the flow domain to increase in height towards the right patch to as a slant to account for the displaced mass from the displacement thickness layer which is causing some level of acceleration. However after doing this, I did not see any change in my sampled velocity results at various stream wise locations on the plate, nor in ParaView, the acceleration was still there and on the same magnitude. My advisor insists that the issue is either my boundary conditions and the displacement thickness. I checked and applied a significant amount of combinations of Boundary Conditions that fit the flow analysis logically (fixedValue, slip, noSlip, freestreamVelocity, freestreamPressure, zeroGradient). All of them returned the same results from the sampled data, and in ParaView. I see great agreement with the flow results with Re > 1500 to 10000 when comparing non dimensional velocity profile vs Eta, the shape factor and wall shear stress. But for the lower Re case which is arround 110 to 330 shows velocity acceleration up to 0.212 m/s compared to 0.2 which skews my plots if i calculated u/Uo based on 0.2 m/s . I just wanted to ask if this is normal and even accounting for the displacement thickness why would the acceleration still occur?. A reference to literature or a reputable resource would be greatly appreciate if you have any on hand to explain this phenomena rather than having you type up something that requires significant amount of your time. I have attached my case setup files below for reference. I appreciate your time and consideration.

Cheers,
Kyouma