boundary conditions for cylindrical grid
 

Hi all! I am trying to implemend the boundary conditions for a 3D grid in cylindrical coordinates.For the z direction there is no problem.For the r direction, when r=0 do I make any special treatment? And for the thita direction I use the symmetry condition or periodic? Thank you in advance.

Re: boundary conditions for cylindrical grid
 

U can apply axis boundary conditions at r=0 i.e. apply zero gradients (or flux) there. As far as applying boundary condition in theta direction is concerned, u can apply symmetry or periodic BC based on the nature of ur problem.

Re: boundary conditions for cylindrical grid
 

For a full 3D problem, you should leave it free....so no condition...

But please create a grid which respects the geometrical symmetry of your system....otherwise you will have problems of convergence....

As said Salah, you can also impose a condition at the axis of symmetry, but then you will be able to find only "axysimmetric solutions" or solutions periodic in the theta direction.

But this condition will not allow you to detect full 3D behavior, i.e an oscillation which totally breaks the axial-symmetry....

So it depends on what is your aim. Steady state or unsteady state?, harmonic or chaotic regime ?, laminar or turbulent regime ?

Re: boundary conditions for cylindrical grid
 

Thank you for your answers. My problem is fully 3D and I am interesting in the unsteady and turbulent flow.At the axis of symmetry I do not impose any condition except from the symmetry one.Is this not correct? And at the thita direction I impose again the symmetry only when, thita=0 degrees and thita=360 degrees , where I=1 and I=NI.Correct?

Re: boundary conditions for cylindrical grid
 

What kind of turbulence model will you use ?

you said "from the symmetry one"....So I do not understand, do you impose or not the axial symmetry at the axis ?

In addition, at theta=0 the solution should be equal to the solution at theta=2Pi....so there is no symmetry there, but a periodic condition should be used.

Re: boundary conditions for cylindrical grid
 

I use the k-epsilon turbulence model and maybe in the future the RNG version of it. At the axis I impose the symmetric condition.In the beginning I have written a code for axisymmetric flow.I have the same treatment of the r-component of the velocity in the axisymmetric and in the full 3D flow. And one more question: at the periodic boundary condition, I set the coefficient Ai=0 and what else? Thank you

Re: boundary conditions for cylindrical grid
 

My personal feeling is that if you impose the symmetric condition on a full 3D domain, and if you use a turbulent model, you will find only the axisymmetric solution....but I am not sure, maybe you will have some phenomena travelling in the theta direction....

So please try and make me aware about your results.

I do not understand the meaning of "Ai=0" in your question:

"And one more question: at the periodic boundary condition, I set the coefficient Ai=0 and what else? "

Re: boundary conditions for cylindrical grid
 

I mean the symmetric condition only at r=0. The coefficient Ai=0 mean the coefficients of the algebraic equation, for example A(west), A(north) etc. When I have some results I will let you know. Thank you for your response