Note: This discussion is about an older version of the COMSOL Multiphysics® software. The information provided may be out of date.
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.
Cfd module, hull viscous resistance
Posted 16.11.2010, 10:29 GMT-5 Fluid & Heat, Computational Fluid Dynamics (CFD) Version 4.0a 9 Replies
Please login with a confirmed email address before reporting spam
i'm trying to solve the viscous resistance of a simplyfied hull geometry: i have modelled the boundary layer with the proper instruments an i've also checked the maximum wall lift off in viscous units as described in the Comsol cfd user manual. my problem is that the solution doesn't converge. Can you suggest me a solution to this problem or some guide lines to solve it?
Geometry and boundary conditions
The hull lenght is about of 10 m, the beam is of 3 m, the maximum draft is of 0.57 m . i've done such model using an ellipsoid with those charachteristics (a=5m,b=1.5m,c=0.57m) and i put it in a rectangular channel (L=40m, B=5m,C=2m) with the stern at a distance of 10m from the inlet. the model is only half because of the symmetry of the hull, the waterline plane is also a boundary of my rectangular domain and i chose for it a symmetry condition. At the inlet i have a normal velocity of 3.5 m/s; at the outlet i have pressure, no viscous stress, on the hull i have wall functions (k-eps turb. model with 0.05 turbolence intensity and 0.1m as turb lenght scale), on the other walls i have allowed the slip condition.
As a result of my calculations the height of the first cell in the boundary near the hull, should be of about 1e-5 m to satisfy the condition on the wall lift off in viscous units parameter.
Thank you very much for the attentions, Edoardo
Please login with a confirmed email address before reporting spam
i'm sorry i made a mistake writing "stern" instead of "bow" when i say: "i put it in a rectangular channel (L=40m, B=5m,C=2m) with the BOW at a distance of 10m from the inlet. "
Thank you for your help, Edoardo Faiella.
Please login with a confirmed email address before reporting spam
Thank you, Edoardo Faiella
Please login with a confirmed email address before reporting spam
thank you, Edoardo Faiella
Please login with a confirmed email address before reporting spam
Please login with a confirmed email address before reporting spam
I'm using Comsol 4.0a.
My model is an exemplification of a sailboat hull (without appendages). I have the original hull but i just wanted to make sure that my problems weren't caused by any geometry error.
I know that viscous resistance calculation is much expensive in terms of hardware and time (i use an intel xeon workstation with 16GB of RAM).
I would be also very interested if it would be possible to evaluate the hull wave resistance using the multiphase cfd module (level set or phase field): for example it should be done a benchmark on the wigley hull.
The purpose of my questions is that of using Comsol for marine applications: first just CFD and then fluid-structure interaction.
Thank you very much, Edoardo Faiella
to download files (20mb):
www.easy-share.com/1913145788/hull geom.zip
Please login with a confirmed email address before reporting spam
Did you take a look to my model? Any suggestion?
Thank you very much, Edoardo Faiella
Please login with a confirmed email address before reporting spam
Thank you, Edoardo Faiella
Please login with a confirmed email address before reporting spam
Any suggestion is welcome.
Thank you, Edoardo Faiella
Please login with a confirmed email address before reporting spam
Your model looks very good to me. I would only change on thing. I would change the slip wall to open boundary. There is not really a wall there is there ?
i think the solver set up is the problem. The default for 3D flow is the iterative solver. You have the geometric multigrid specified. It will automatically create coaser meshes for you using free mesh. However, you have a boundary layer mesh also there. I have not had much success with the automatic feature for the multigrid for cases where a boundary layer exists.
Instead, I have always used the "manual" option and specified several coarser meshes myself. You will need the coarsest mesh to fit in a direct solver. So, test that first, then create successively finer meshes testing the multigrid each time. I use linear elements throughout except , if you want, quadratic elements for the finest case if you have room in your machine to do that.