CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Main CFD Forum (http://www.cfd-online.com/Forums/main/)
-   -   which way to solve a moving airfoil? (http://www.cfd-online.com/Forums/main/13090-way-solve-moving-airfoil.html)

 CH March 6, 2007 01:20

which way to solve a moving airfoil?

hi,

my simulation involves moving airfoil in a fluid using SIMLPER FVM mtd. i'm thinking of implemented the ALE formulation. however i'm not sure how should i move the grid...

there's 2 ways - 1. deform the grid, keeping the outer boundary fixed. 2. move the whole grid, outer boundary move.

it seems that the 2nd mtd is much easier & it also preserve the grid quality. but is there any problem with it? i saw quite a few papers where ppl deform the grid instead of moving it...

i'll be very grateful if someone can comment on it. thank you.

 H. Yagubizade March 6, 2007 03:07

Re: which way to solve a moving airfoil? *NM*

 Harish March 6, 2007 09:38

Re: which way to solve a moving airfoil?

Are you studying the surging motion of airfoil?. In that case the body fitted moving grid is easier to implement than the grid deforming formulation.You might want to look up in school.google.com for papers by the group of sengupta.They have solved the flapping airfoil problem using a non inertial coordinates and obtained pretty good results.If you are interested in deforming grid case,look for the phd dissertation of young from some australian university.The dissertation is available online.He used deforming grids,I think to study the flapping motion.

 ag March 6, 2007 09:46

Re: which way to solve a moving airfoil?

Unless the airfoil is deforming, moving the entire grid is easier and eliminates the possibility of re-gridding issues. The downside is that for large motions, regions that were initially gridded in a fine manner to capture separation regions, etc. move into regions of more uniform flow, while regions possibly gridded more coarsely end up (at least momentarily) representing the solution region in areas of large gradients. For angular displacements less than 15-20 deg, I have found that moving the grid is usually sufficient if you start out with a good grid.

 All times are GMT -4. The time now is 00:18.