hi,

I am trying to simulate flow over a gas turbine stator blade.when I read the grid generated using gambit,it recognizes shadow pairs.Can someone tell me what is this concept?

Also,when fluent tries to build the grid after reading it,it gives a message "skipping zone periodic-shadow(not referenced by grid).Is this a grid error?because when I give grid check it does not report any error.

whenever you have a surface that divide two different material, you have shadow wall. E.g. you have a surface ( face ) that divide a solid zone from a fluid zone : the surface dividing has "two side" : first side is wall-solid , its whadow is wall-fluid. If you couple the walls , you have real heat transfer and termodinamic interaction between solid and fluid,but you can either choose to not couple the walls, and fix the temperature of wall-solid to a value A and the temperature of the wall-fluid to value B. Sometimes you need this.

the other message you receive is not an error.

hi,

shadow pair is a link between two periodic faces which ensure that these face has exactly the same meshing patern.

This is necessary to have a conformal periodic condition which is the only one that fluent can handle.

The solver itself treat the periodic condition differently and that's why you have a message "skipping zone periodic shadow". It isn't an error.

Try to run your case and verify that the periodicity is well handle by looking at pressure and velocity field.

Best regards

Alain

thank u for the reply,one more thing.when I read tutorial mesh for stator blade,both min. and max.of x-translation was exactly zero and for y-translation min.and max. was matching upto 6 decimal places. whereas,when I read the my own mesh,min. of x-translation was zero but max was around 2.3e-05.y-translation min and max values were matching only till 4 decimal places.

will this make a difference in the results?

It depends of the size of your models. If the spatial discrepancy is under the precision of the machine that won't change the result.

If you want a more precise mesh make a bigger mesh in gambit and then scale it in fluent.

best regards.

[myaccount]

Hi I am using periodic boundary conditions in GAMBIT.I have run the case in fluent.After convergence I changed the grid(finer/coarser), maintaining the same convergence criterion,I have obtained the same results.Velocity profile is exactly overlapping for the coarser and finest meshes.Where lies the problem?