CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   How do I avoid calculate pressure on 1st itr (http://www.cfd-online.com/Forums/cfx/80167-how-do-i-avoid-calculate-pressure-1st-itr.html)

 hts# September 17, 2010 02:06

How do I avoid calculate pressure on 1st itr

Hi, I am trying to use an expression to calculate mass flow (to be given as an inlet and outlet boundary conditions) in a small cavity attached through frozen rotor to the bigger domain (having the actual inlet and outlet of the model). This expression involves a constant clearance value and the pressure calculation across the frozen rotor patches between the cavity and main domain. The solver fails at first itr because pressure values are not available at that time.. I need an expression to give a contant pressure value at iteration 1 and then later start calculating from the simulation itself. I tried to use if statement.. but in that it agains tries to calculate the pressure at first itr.

Thanks
#

 ghorrocks September 17, 2010 06:04

I take it from your question you have a narrow clearance slit which is leaking fluid from a high pressure region to a low pressure region. I assume the fluid is air or some form of compressible gas.

There are often large pressure gradients across this type of flow, meaning the flow is often sonic and convergence of sonic flows with greatly varying length scales is tricky.

Your idea of locking pressure at iteration 1 is not a wise approach. Better to understand what is actually the problem first.

Can you define: What is the inlet and outlet pressures? What is the size of the chambers? The gaps? What is the fluid? Any other relevant physics?

But if my guesses are right you should consider local timestepping for a while to start things off, then move to physical timesteps with small time steps. When they are running nicely increase the size of them to get convergence.

This approach is summarised here http://www.cfd-online.com/Wiki/Ansys...gence_criteria

 ghorrocks September 17, 2010 06:06

Oh yes - and the initial pressure level is set in the initial guess. It is never undefined. And it is a coupled momentum/pressure solver so you can't fix the pressure values anyway.

 All times are GMT -4. The time now is 10:26.