Problem in simulations
Posted: Thu Mar 03, 2011 5:01 am
Here is where I currently am and the issues I am working to resolve. Any advice, correction, or conclusions are welcomed.
Issue 1)
While working with turbulent channel flow problems I found out that the no slip wall condition was not present in my results (ie Velocity at wall was not zero). This is my current problem which I am trying to resolve. To investigate I scaled back to a pipe flow problem. Running the simulation as laminar with a Reynolds number around 100 yields the expected results and velociy profiles for classical pipe flow. However running the simulation as turbulent, Reynolds number of 30,000 using the K-epsilon turbulence model does not show zero velocity at the walls. Reading about the K-epsilon turbulence model online I got the impression it may not be the best set of assumptions to use near the wall/boundary layer. Is this correct? Indeed viewing my results showed the velocity gradent dv/dy near the wall was near infinite. What turbulence model, might you recommend.
Issue 2)
I am trying to enforce periodic boundary conditions in the stream wise direction. To do this I THINK? I must proscribe a pressure drop to drive the flow. To do this I THINK? I must edit the user subroutine ustsns.f90 to change the source term in the momentum equation. I have never used python before, but it seems pretty straight forward to read and edit the user routine file. Do you think I am headed in the right direction hear or do you know of a better way to proceed or am I totally wrong.
Issue 3)
Calculation of the Mean Velocity. To get this from Code Saturne I will have to edit the user subroutine usproj.f90. which is user defined output subroutine. Am I in the right track ?
I have attached some files as zipped that I thought might help you understanding my problem.
Thanks and Regards,
Debashis
Issue 1)
While working with turbulent channel flow problems I found out that the no slip wall condition was not present in my results (ie Velocity at wall was not zero). This is my current problem which I am trying to resolve. To investigate I scaled back to a pipe flow problem. Running the simulation as laminar with a Reynolds number around 100 yields the expected results and velociy profiles for classical pipe flow. However running the simulation as turbulent, Reynolds number of 30,000 using the K-epsilon turbulence model does not show zero velocity at the walls. Reading about the K-epsilon turbulence model online I got the impression it may not be the best set of assumptions to use near the wall/boundary layer. Is this correct? Indeed viewing my results showed the velocity gradent dv/dy near the wall was near infinite. What turbulence model, might you recommend.
Issue 2)
I am trying to enforce periodic boundary conditions in the stream wise direction. To do this I THINK? I must proscribe a pressure drop to drive the flow. To do this I THINK? I must edit the user subroutine ustsns.f90 to change the source term in the momentum equation. I have never used python before, but it seems pretty straight forward to read and edit the user routine file. Do you think I am headed in the right direction hear or do you know of a better way to proceed or am I totally wrong.
Issue 3)
Calculation of the Mean Velocity. To get this from Code Saturne I will have to edit the user subroutine usproj.f90. which is user defined output subroutine. Am I in the right track ?
I have attached some files as zipped that I thought might help you understanding my problem.
Thanks and Regards,
Debashis