[aspect-devel] S40RTS and thermal boundary layers

[Wolfgang Bangerth]

On 07/22/2016 05:31 PM, Marie Kajan wrote:
> Yes, they do vanish when I remove the 'Fixed temperature boundary indicators'
> parameter as suggested by Rene... but I guess I'm trying to wrap my head
> around how the ASPECT code takes those initial conditions and fixed
> temperatures and generates "boundary layers" in a single time step? (I'll
> include a picture of both versions for clarity.)

Before the very first time step, we use the initial conditions to initialize 
the temperature field everywhere. We then solve a time step with length zero 
(we solve for the temperature in every time step, so we also do in the first 
time step -- but because we don't know the flow velocity yet, we solve with a 
zero time step). During this solve step, we impose boundary conditions as we 
always do, on the temperature, and at the end of the time step we create 
graphical output. In other words, what you get to see is not actually the 
initial temperature, but the temperature after solving one zero-length time 
step with the boundary values you describe in your input file.

Does this explain what you are seeing?

Best
  W.


-- 
------------------------------------------------------------------------
Wolfgang Bangerth               email:            bangerth at math.tamu.edu
                                 www: http://www.math.tamu.edu/~bangerth/