*** WARNING: Diss.-rate fully specified by two-layer wall treatment problem

Hi,

I am simulating a flow through a pipe with oil at different temperatures. I am using the k-epsilon model, but I have also tried other turbulence models (RNG k-epsilon, S-A, Realizable k-epsilon, k-omega, SST). The result of interest is the distribution of the mass flow rate coming out of any outlet as the inlet mass flow rate varies. The results are nearly identical for all turbulence models used, differing no more than +-1%.

For lower temperatures (such as -20 °C, 0 °C, and 30 °C) with any k-epsilon model the residual dissipation rate and solution ratio are for the entire simulation equal to 0, and I get the following warnings, even changing the mesh.

acuSolve: staggered dissipation rate "dissipation_rate": FORM-LHS

acuSolve: *** WARNING: dissipation_rate is fully specified by the two-layer wall treatment

acuSolve: res diss.-rate ratio = 0.000000e+000

acuSolve: GMRES No iteration = 0 (0.00)

acuSolve: GMRES 0/1/n norms = 0.000000e+000 0.000000e+000 0.000000e+000

acuSolve: GMRES Iter. CPU/Elapse = 0.000000e+000 0.000000e+000

acuSolve: diss.-rate sol ratio = 0.000000e+000

How can I solve this problem? Is it correct to consider the results as correct with this warning?
I attach the .inp model and the .log file of one of these simulations.

Thanks in advance

example.inp

example.Log

Find more posts tagged with

AcuSolve

English

HyperMesh CFD

v2021.1

Accepted answers

All comments

acupro

Do you get the same behavior with newer versions? (The current version is 2022.3.)

Can you post an image (or images) of the overall domain - indicating what BCs are applied where? One should only use mass-flux, flow-rate, average-velocity inflow types if the boundaries neighboring the inflow are type = wall. Also, the inflow should not be in-plane with a neighboring wall or other boundary.

Fededea

Do you get the same behavior with newer versions? (The current version is 2022.3.)

Can you post an image (or images) of the overall domain - indicating what BCs are applied where? One should only use mass-flux, flow-rate, average-velocity inflow types if the boundaries neighboring the inflow are type = wall. Also, the inflow should not be in-plane with a neighboring wall or other boundary.

I have not tryed with the newer versions.

In attachment an image, there is 1 inlet that you can see in the image and many outlet (9), not all in the image.

The used boundary condition are:

inlet: volumetric flow rate
outlet: at environment pressure = 0 Pa

Thanks

acupro

I have not tryed with the newer versions.

In attachment an image, there is 1 inlet that you can see in the image and many outlet (9), not all in the image.

The used boundary condition are:

inlet: volumetric flow rate

outlet: at environment pressure = 0 Pa

Thanks

I just tested with a simple pipe model - and at the start of the first time step I got the same warning, but then it worked itself out and subsequent values were non-zero as expected.

Looking at that model geometry, and the Log file indicating only 1.3 Million nodes in the meshed model, my guess is you don't have enough mesh resolution, and that is likely the cause. Make sure you have boundary layers defined off the walls to get the appropriate near-wall resolution.

Can you also post a zoomed-in image of the inlet area - showing the mesh?

Fededea

I just tested with a simple pipe model - and at the start of the first time step I got the same warning, but then it worked itself out and subsequent values were non-zero as expected.

Looking at that model geometry, and the Log file indicating only 1.3 Million nodes in the meshed model, my guess is you don't have enough mesh resolution, and that is likely the cause. Make sure you have boundary layers defined off the walls to get the appropriate near-wall resolution.

Can you also post a zoomed-in image of the inlet area - showing the mesh?

Here the image. I also tryed with a finer mesh but the problem remain.

acupro

Here the image. I also tryed with a finer mesh but the problem remain.

Can you attach the latest .inp and .Log files for that finer mesh?

Next step - please install and try with latest version 2022.3.

Fededea

Can you attach the latest .inp and .Log files for that finer mesh?

Next step - please install and try with latest version 2022.3.

Hello, here the .inp and .log files for finer mesh

fine.inp

fine.Log

acupro

Hello, here the .inp and .log files for finer mesh

I see the finer mesh didn't change the behavior. Any difference with version 2022.3 - the current version?