I just submitted this to CFX customer support:

1.) I found that I cannot define an expression in the boundary condition form; For wall heat flux I inserted the following expression: qW*x/pX (where qW and pX are constants, so only x is in a variable). This caused solver to end before 1st time step and to write : "Error #001100279 has occurred in subroutine ErrAction. Message: Stopped in routine Cound not find phase directory. Also any other kind of variable expression ended the solver execussion.

As the example is the most simple one and very similar to the one in CFX tutorial (inlet profile), I would like to know what is wrong with my model or with the code.

2.) I am trying to make a periodic heat transfer calculation, which force me to imposed periodic and Neumann kind of BC. As the temperature is not fixed in any of numerical points, the temperature in my simulation domain is raising indefinitely. What I would like to do is normalize the temperature field in each time step like this: T=T-min(T)+23 (for 23 degree of Censius). Is this kind of normalization feasible in CFX 5.5 and how to make it.

Did anyone experience something like this (1). Does anyone have an answer on these to questions.

Andrej

Hi Andrej,

1) I guess one would need the case to check it out; so CFX support is your best bet; is it multiphase?

2) I guess for this you would need some fortran to be able to access the variable structure at the end of the previous iteration to set it up at the next; that is not trivial in such an unstructured code and because of the solver functionning; you would need to enquire to your support again on how to access the variable structure and use the fortran hooks to code such a condition. Periodicity is nice but difficult to manage as you continuously build up your momentum or energy and can get of course very easily...