Learning Forum

[S.Shrestha-4]

Hi! 

I am trying to run a multiphase simulation in 2D rectangular mesh using VOF method with velocity inlet, pressure outlet and heat flux input at bottom side. At outlet gauge pressure = 79997.66972 [Pa], operating pressure= 400000 [Pa], so that I have an absolute pressure of 479997.6697 [Pa]. 

I get following error and the simulation stops.

Divergence detected in AMG solver: pressure correction temperature limited to 5.000000e+03 in 15283 cells on zone 8 in domain 1

Divergence detected in AMG solver: pressure correction

Divergence detected in AMG solver: pressure correction

Reversed flow on 15 faces of pressure-outlet 11.

Divergence detected in AMG solver: pressure correction

Error at host: floating point exception

Could someone help me figure out why it happens? Thanks a lot!

[Rob]

Have a look at the flow field - what is going on? 

[S.Shrestha-4]

Looking at the temperature field, it quickly goes to 5000 K while max temperature should be around 400 K, and the pressure field is all messed up. 

[Rob]

And how are you adding the two phases?

[S.Shrestha-4]

 

I added air as primary phase and, water and vapor as secondary phases. I saw in one of the tutorials online that the reason for including a phase with air is that if only water liquid and vapor are defined, the calculation starts as if there is already water vapor inside the pipe (don’t know how true is this?)  I used explicit vof formulation. Also, the geometry dimension is in order of micrometer.  

 

[Rob]

The calculation will start based on what you set for the initial conditions: that's something you need to check. With VOF you can only specify one or other phase on the inlet boundary.  Micron scale may be OK, it depends on how many microns you're using. However, the mesh resolution necessary may be far more than you have set. 

[S.Shrestha-4]

I did hybrid initialization and did a patch afterward with water and vapor volume fraction = 0. At the inlet I set, water volume faction = 1 and vapor = 0. Could it be that I set pressure outlet at 5 bar? What pressure inlet would fluent set? I used velocity inlet, so there was no pressure inlet input. 

What other initial conditions should I set and where?

[Rob]

And vapour is being added where? If the outlet is 5 bar (gauge) the inlet will be a higher pressure to push the set flow in, but that may take an iteration to pick up. 

[S.Shrestha-4]

Vapor is not being added, water has to boil to form vapor and only vapor is allowed to exit the outlet. 

[Rob]

Where does the liquid go that's added from the inlet? 

[S.Shrestha-4]

The idea is that liquid vaporizes completely, so no liquid comes out and only vapor comes out the outlet. 

[Rob]

And you've fixed the wall temperature, or fixed the heat flux? What happens in the early stages if you initialise the domain to be full of water (think about the physical aspect as opposed to the solver)? 

[S.Shrestha-4]

Yes, there is constant heat flux at the bottom wall. How do you initialize the domain with water when it's running through the channel? Do I patch domain after initialization with water volume fraction = 1? I will try with this. 

And since no water is allowed to exit the outlet, would it completely vaporize after some time? 

[Rob]

The problem is liquid enters at a set rate and if it doesn't evaporate can't leave. The solver then struggles and fails. You need to rethink the set up. 

[S.Shrestha-4]

 

It’s not clear how I can set it up differently. My inlet velocity is around 0.004 m/s and there is no restriction set at outlet which says liquid can not come out. It would be ideal to have a vapor at outlet, however. As there is a nozzle downstream the channel. I have only defined static pressure at the outlet.

 

[S.Shrestha-4]

 

Hi! I set up the simulation with two phases, with water at the beginning filling the domain, like you suggested. However, even after around 335k iterations, and the wall temperature having passed beyond the saturation temperature of water (423 K at 5 bar), evaporation is not taking place. Heat flux at lower wall is around 90 kW/m2, other wall is adiabatic. I used "to phase frequency" =10 s^-1.  There seems to start some nucleation at the bottom surface, but it’s taking way long to form bubbles, or is this normal? 

Also, I keep getting this:

 Reversed flow on 9 faces of pressure-outlet 11.
 344621  1.2086e-04  5.2476e-05  4.2446e-05  1.2394e-08  0:00:01  149
!344621 solution is converged
  step  flow-time report-def-1
 328052  7.7101e-01  4.1939e+02
Flow time = 0.7710071859378018s, time step = 328052
299948 more time steps

Updating solution at time level N…

Global Courant Number [Explicit VOF Criteria] : 0.000153432
 done.

It does not iterate and says converged, but the residuals look fine.

 

[Rob]

Zoom in around one of the vapour regions and replot with node values off and mesh lines on.