Update internal mesh with a predicting equation instead of motion solver with diffus?
 

Update internal mesh with a predicting equation instead of motion solver with diffusivity?

Hi, everybody!

My channel has a wavy bottom wall, which can be described as y_wall(x,t)=a*sin(k*(x-c*t), i.e. a travelling sine wave. This could be achieved by adapting "waveDisplacement".

However, I have trouble with the diffusivity. None of the available diffusivity can satisfy my need. For it squeeze grids vertically above crests and stretch over troughs, squeeze makes the first grid spacing much too small, and stretch makes the first grid spacing even larger than the second. Any given point at the bottom wall can be crest at one time, but becomes trough the next time. So it's tricky to give a proper diffusivity. See attached pictures.

Yet, the grid points can be predicted only if the bottom wall position is known. The i_th point elevation y_i = y_0 * r^(i-1) + y_wall, where r is the expansion ratio, y_0 the first grid height.

Does anyone known how to update the internal points using a predicting equation rather the "diffusivity"? Where should I add it?

Best regards.
Nick

dear Nick
i am a new of fluent , i need the predicting equatio like sine wave,but i cannot work out in udf,plz help me ,can you show part of your udf code?any advice will be great appreciated,thanks a lot.
jinsong