Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.
Turek's benchmark
Posted 18 nov 2017, 05:45 GMT-5 Fluid & Heat Version 5.3 2 Replies
Please login with a confirmed email address before reporting spam
Hello everybody, I have a question for you, hoping somebody can help. I'm using Comsol 5.3 to analyze Turek's benchmark (following the instructions found here: https://www.comsol.it/model/vibrating-beam-in-fluid-flow-9408), and it works fine. But I'm interested in change the velocity profile by assuming an average velocity U=10m/s instead of 2m/s, and what I get is that the solution doesn't converge.
"Nonlinear solver did not converge. Time : 0.7432944384400323 Maximum number of Newton iterations reached. There was an error message from the linear solver. The relative residual (2.1) is greater than the relative tolerance. Last time step is not converged." And some warnings about inverted mesh in some points.
Can you please help in some way? I tried to use a finer mesh, or the generalized alpha solver insted of the BDF, and I don't use any limits about minimum and maximum step (so that the program itself can choose), but nothing works.
I attach the deformed beam in the last time-step, it also seems that the condition about velocity doesn't work, as the blue color corresponds to 0m/s.
I hope you can help, and I thank you for any suggestions.
Attachments:
Please login with a confirmed email address before reporting spam
Hello Nikolas,
With fluid structure interaction models, one thing to keep in mind is that if the structure deforms too much the mesh may become inverted. Refining the mesh rarely solves the problem, in fact it can make it worse because it's easier for a small element to get inverted than a bigger one.
The following two blog posts present various strategies that can help:
https://www.comsol.com/blogs/model-translational-motion-with-the-deformed-mesh-interfaces/
https://www.comsol.com/blogs/deformed-mesh-interfaces-rotations-and-linear-translations/
Best,
Jeff
-------------------Jeff Hiller
Please login with a confirmed email address before reporting spam
Hi,
It may also be so that no stable solution exists. You could try with a smaller increase in velocity, and see what happens.
In this example, laminar flow is assumed. This may not be the case when the velocity is higher. If that is the problem, you may need to incorporate a turbulence model.
Regards,
Henrik
-------------------Henrik Sönnerlind
COMSOL