[ganeshv]

The fixedMeanValue BC doesn't work for me in parallel mode. It works fine in serial mode though. Has anyone got it to work in parallel mode?

I'm trying to implement a wall function for eddy viscosity that requires averaging of the velocity vector over the set of cells adjacent to a boundary. The BC I have implemented works fine in serial mode. But once I go parallel, the gAverage function fails on me. I tried implementing the rmap and autoMap functions that are present in the fixedMeanValueFvPatchField BC and also changed the evaluate function to an updateCoeffs function...but to no avail.


Could some one explain to me what is the difference between evaluate and updateCoeffs functions. Does updateCoeffs do evaluate + setting the boundary, value, gradient coeffs etc?

And what does autoMap and rmap do and is it relevant to the gAverage function working properly?

[chegdan]

Originally Posted by ganeshv

The fixedMeanValue BC doesn't work for me in parallel mode. It works fine in serial mode though. Has anyone got it to work in parallel mode?

I'm trying to implement a wall function for eddy viscosity that requires averaging of the velocity vector over the set of cells adjacent to a boundary. The BC I have implemented works fine in serial mode. But once I go parallel, the gAverage function fails on me. I tried implementing the rmap and autoMap functions that are present in the fixedMeanValueFvPatchField BC and also changed the evaluate function to an updateCoeffs function...but to no avail.


Could some one explain to me what is the difference between evaluate and updateCoeffs functions. Does updateCoeffs do evaluate + setting the boundary, value, gradient coeffs etc?

And what does autoMap and rmap do and is it relevant to the gAverage function working properly?

I have a simulation running right now with the fixedMeanValue BC in parallel. The code was from the thread (http://www.cfd-online.com/Forums/ope...condition.html) and I have recompiled (with no changes to the code) it to work with the 2.0.x version of OF. So far, everything works fine.