[gdbb89]

Hello everyone!

I am trying to learn how to do aeroacoustics analysis in COMSOL, and I found this tutorial to follow: https://www.comsol.com/model/downloa..._with_flow.pdf

On page 26, there is a plot of turbulent viscosity along the centerline of the lower pipe:



When I attempt to replicate this (by following the tutorial), I get this:



One slight issue is that I'm using v5.1, and this tutorial was made for v5.3. However, I was able to work my way through the tutorial using my version. There were a couple extra steps I had to take in the preparation of the geometry and mesh, but they were rather minor, so I don't believe I messed anything up. For instance, here is my full mesh:



I am using hex elements in the fore and aft sections of the pipe/tube, and tet elements in the chamber and the T-section of the pipe. A closer look at my mesh:





My question here is less "how do I fix this" and more so "what physical mechanism is this representing?" Physically speaking, what is the mechanism that drives an increase in turbulent viscosity along the centerline of the pipe in a situation like this? I should note that there is a slight velocity increase along the centerline as well. Is it an issue with my mesh?

I first guessed that the velocity increase indicated an acceleration along the centerline and that maybe my outlet pressure of 0 atm was absolute and not gage, thereby "sucking" the air out of the tube. But I ran another case with 1 atm outlet pressure, and that only served to further increase the turbulent viscosity away from the tutorial's results.

Any insight is appreciated.

Thanks!

[LuckyTran]

It looks to me like it is happening at the interface between the hex and tet mesh. You can try moving this border around to see how it influences the result. I.e move the interface upstream and downstream by decreasing the size of the hex region and increasing the size of the tet region and vice-versa.


Other than that little kink, I think you reproduced the results of the tutorial quite well.

[gdbb89]

Originally Posted by LuckyTran

It looks to me like it is happening at the interface between the hex and tet mesh. You can try moving this border around to see how it influences the result. I.e move the interface upstream and downstream by decreasing the size of the hex region and increasing the size of the tet region and vice-versa.


Other than that little kink, I think you reproduced the results of the tutorial quite well.

Thanks for the response. Is it not an issue that the turbulent viscosity increases in my simulation as opposed to decreasing in the tutorial?