[soumitra2102]

I have to simulate sub cooled water boiling (using RPI wall boiling) in a tank open to atmosphere. I have used outlet boundary for the top surface of the tank. The pressure in tank remains reasonable (~atmospheric) and temperature of liquid water increases with time. But as soon as temperature comes near to saturation (365K -373K), the solver stops with a fatal overflow error (Maybe increased agitation in the tank would have caused backflow at the boundary). Of course, I had overflow warnings at every single iteration (along with very unstable residuals for all monitoring fields), but I let the solver continue solving for as much iterations as it can.


When I changed the boundary to degassing, the pressure in the tank shoots to a very high value and keeps on increasing with time. (due to this the fluid may not boil ever). I was expecting the dispersed phase- gas water to leave the domain, but the reason behind pressure rise is not understandable. One good this was that the residual plots were very smooth although there was overflow warning for only liquid phase.


I can not use opening boundary because I do not have the 'Opening Temperature' field value prior to simulation.


What can I do if I want to successfully simulate the boiling multiphase process?


P.S.: I did not use Homogeneous binary mixture to link both the phases; I rather defined two separate fluids, 1st one as liquid-Continuous phase and 2nd one being Vapour- Dispersed phase in the domain definition, as per one ANSYS Workshop tutorial on RPI Wall boiling and Thermal Phase Change, where special CEL is used for parameters like BubbleDiam, etc. and where simulation results were compared with Bartalomej test case.


Should I have used the Homogeneous Binary Mixture definition approach?

[Darko]

I also have faced the same problem in a multiphase reactor.
Have you found a solution so far?

[ghorrocks]

You have to be careful that the pressure is correctly handled. I don't think the initial post did that, he/she should have placed a pressure boundary somewhere to keep the pressure realistic. (unless this is a sealed tank and the pressure actually does rise)

[soumitra2102]

Yes, I think ghorrocks is right. In my case I think the tank rather acted as a sealed tank. Hence, obviously the pressure kept increasing with the increase in the temperature. And hence boiling could never actually be achieved.

But I still don’t know how I could have solved this problem.

[Darko]

Originally Posted by ghorrocks

You have to be careful that the pressure is correctly handled. I don't think the initial post did that, he/she should have placed a pressure boundary somewhere to keep the pressure realistic. (unless this is a sealed tank and the pressure actually does rise)

This does make sense to me in some way.
But to make sure let me explain the case I am interested in.
I am simulating an aerated stirred tank and from the experiments the mixture height (liquid + gas) in the tank is clear, this is where I have defined the degassing boundary.
I used a mass flow inlet for the gas sparged at the bottom.
I also know that the headspace pressure is around 1.29 atm and I tried to mimic that by setting the operating pressure to 1.29 atm a few centimeters below the degassing boundary.
It might be worth noticing that my gas phase is modeled as an ideal gas.
Now what ghorrocks said makes sense because when I use pressure output the simulation gives the gas flow regime that was observed in the experiment for most parts but then the top part is not simulated properly as expected.
My take from our discussion was that I need to provide some pressure boundary to achieve proper pressure profile, right?
Will it work if I introduce a pressure inlet?

Regards,
Chris

[ghorrocks]

Inlets only allow inflow, outlets only allow outflow. So only use them if you know the flow direction. If you don't know to direction, or it could be either, then use an opening.