[Cleavily]

Hi guys, I'm currently working on simulating the pressurization/depressurization of a large test chamber (from about 15 torr to 760 torr). Inside the chamber is a small model that is off-gassing, and we want to see if this pressurization (and subsequent depressurization) is sufficient in diluting and removing this off-gas from the chamber. The test chamber is pressurized and depressurized a total of 3 times. I figured each pressurization and each depressurization would be a separate simulation. In real life, this entire procedure takes about 15 minutes, and a pressure sensor is used to see when the chamber pressure reaches 15 torr or 760 torr so that a valve can be closed/opened to resume pressurization/depressurization.

I'm very new to FLUENT, but what I have so far is this: I've established a mass-flow-inlet to represent the model off-gassing, and a pressure-inlet for the valve to atmosphere that is opened. Since the test box starts at 15 torr, I've set the operating conditions to be 15 torr, and the pressure-inlet to be 745 torr (gauge pressure). I'm using the pressure-based solver, and I have species transport on since the model off-gas is benzene vapor. This subsequently turns on the energy equation, and when I try to run the simulation for a few time steps it always gives a "divergence detected in the AMG solver: temperature" error after a few iterations. Overall, I believe my mesh is okay, and I'm not sure what else to do.

If anyone has any suggestions on how to model this scenario, please let me know! And thank you for your help!

[CeesH]

Can you post an image of the geometry with the different sections to clarify the setup a bit?

Some other questions: are there two inlets in real life, since you have specified two? If there is only one valve, you need either a mass flow inlet (with fixed flowrate), or a pressure inlet, but not both.

Did you set up the simulation to be transient? What's the timestep (and why?)

Is species transport relevant? From what I gather, you model a box being pressurized; there does not seem to be an outlet or requirement to observe the mixing of different species - so why include it?

How do you treat density?

[Cleavily]

Originally Posted by CeesH

Can you post an image of the geometry with the different sections to clarify the setup a bit?

Some other questions: are there two inlets in real life, since you have specified two? If there is only one valve, you need either a mass flow inlet (with fixed flowrate), or a pressure inlet, but not both.

Did you set up the simulation to be transient? What's the timestep (and why?)

Is species transport relevant? From what I gather, you model a box being pressurized; there does not seem to be an outlet or requirement to observe the mixing of different species - so why include it?

How do you treat density?

Thanks for the response! I'll try to answer your questions. Unfortunately I can't really include pictures of the geometry, but it consists of a square test chamber (approx. 2m x 2m x 3m), with a long cylindrical pipe attached to one face (approx. 1m diameter).

The only valve open would be a pressure inlet, at 760 torr. The test chamber is at 15 torr. The second inlet would be a test article inside the chamber which is off-gassing, and I've assigned it as a mass-flow-inlet.

Yes the simulation is transient, and I've set up a 0.01s timestep. I'm not sure if this is a good number to use, as I have nothing to compare to an application this unique.

There is a mixing of air (in the test chamber at the start) and benzene vapor coming from the mass-flow-inlet. The goal is to see if after the chamber is pressurized, if there are any pockets of benzene.

Please let me know if you have any more questions!

[CeesH]

A simplified drawing of the domain might do too. But anyway, if you are new to FLUENT, I'd advice to start off a bit more simple than throwing everything in there, all at once.

So maybe first, just start with a simple box. 1 inlet, no outlet. Just one phase - so no species transport. Make a list of the relevant physics, and decide how to include them. Will there be turbulence or not? That decides your viscous model. I suppose with pressurizing a gas, you will need to have an equation of state as well; will the ideal gas model suffice? Will there be temperature changes in your process that need to be accounted for? You can also test the influence of timestep sizing here. With implicit time integration, you won't need a Courant number of <1, but at least the timestep should be sufficiently finer than the relevant timescales in the problem. You can make some guesstimates there, and try varying the stepsize to see its influence.