open TELEMAC-MASCARET The mathematically superior suite of solvers

Dear all,

While performing a coupled t2d simulation in a curved channel, I wanted to have the re-circulating flow in the inner bend which is shown in the attached image.

As far as I know, in the case of zero viscosity model, velocity diffusivity can be changed.

I am using the K-epsilon model for turbulence modeling. In this regard, which parameter should I modify to obtain the recirculating flow clearly?

My t2d and sisyphe steering files are also attached.

Thanks in advance.

Saroj

Hi,

The following (main) parameters can influence presence and / or size of recirculations:
- actual flow conditions (velocity magnitude and distribution in this case)
- mesh size
- turbulence model
- advection schemes

The images you posted indicate that you have a rather large mesh size along the banks making it very difficult to reproduce possible recirculations. Keep in mind that your mesh size is very important for the phenomena you want to reproduce!

Another very important thing is: are you sure that recirculation should occur there? You need calibration data to assess that. As you seem to be working on an idealized case, you should make a sensitivity analysis highlighting the variations that can be obtained with different settings.

My experience shows that k-epsilon is suitable for reproducing recirculations, but other models should work fine as well (Smagorinsky, Spallart-Almaras?).

I remember a nice paper from a past User Conference (2014, 2015?) that was highlighting the influence of advection schemes on circulation patterns between groynes.

Good luck!

Best regards
PL

Dear PL,

Thank you for your comprehensive suggestions.

I have performed laboratory experiments which shows that the re-circulation occurs in that region.

Talking about the mesh size, I once achieved the re-circulation (as shown in the attached image) with the same mesh size but at that time I used zero viscosity for turbulence while currently, I am trying with K-Epsilon. (The steering files are attached)
However, I have not considered much about the advection schemes which you have pointed out. Any suggestions on which scheme to consider?

Also, I have performed some cases with groynes. In that case, I can clearly see the re-circulation between the groynes. (as seen in the attached image)

I am still trying to achieve the re-circulation in those regions. Let's see.

Thank you.

Regards,
Saroj

Hello,

It will be interesting to see your results!

I still think that your mesh size is pretty coarse if you plan to study recirculations. Even if you manage to have some in your simulations there are great chances that they would be (heavily) mesh-dependent!

Regarding advection schemes I cannot give accurate advices since I very seldom work on cases with extended field and calibration data (read: consulting!). Based on the paper I mentionned in my previous post I chose to use the MURD scheme in my models (4/14). But since then new schemes (and even turbulence models) were added so the best would be to test them and see what works the best for your case.

Good luck,
PL