Hold up in a closed vessel
 

Hi all,

I am simulating (steady state and transient) the flow in a closed vessel, meaning that there are neither inlets nor outlets. The initial total amount of mass (hold up) is set by the reference pressure and the initial guess for the pressure. During the simulation the hold up is far from constant. Mass does not seem to be conserved in CFX 5.5.1/5.6!!!!!! Does any one no why? How can I fix this problem?

Pascale

Re: Hold up in a closed vessel
 

Hi Pascale,

How many coefficient loops are you taking per timestep, and is the solver converging to you linear solver target?

Regards, Robin

Re: Hold up in a closed vessel
 

If you check my earlier posting, ( http://www.cfd-online.com/Forum/cfx.cgi?read=4987 ) you will notice some similrities. In the initial system (of that posting) I managed to solve the problem. I have not yet succeded in a different system i am dealing with presently. However, the starting point is what Robin has mentioned.

Regards Nyatoto

Re: Hold up in a closed vessel
 

- 3 loops

- after the first five time steps, the residuals fall down to approx 1e-5. Isn't that sufficient?

Pascale

Re: Hold up in a closed vessel
 

Pascale,

The inner loop residual should be at least a MAX of 1e-3. You should require 3 to 5 coefficient loops to get to this level (note that this is not the same as the outer loop residual). I will typically allow the solver to use up to 7 inner loops and then adjust my timestep so it only uses 3 to 5 of them. If you are using all seven, you know the linear solver is not converging sufficiently.

The code does not explicitely conserve volume fraction. There is a beta parameter in CFX-5.6 which will do this, although it is really intended for steady state calculations. With a transient simulation you should be conserving the volume fraction anyway, if you are converging at each timestep. For the beta parameter, contact your local technical support.

Regards, Robin