## Fluids

#### Different heat transfer results when using compressive or geo-reconstruct

• Navsing
Subscriber

I'm simulating condensation of water on a 2D pipe. I have written a UDF using the DEFINE_MASS_TRANSFER macro to determine the mass transfer rate of the vapour via using an energy balance method which forces the interface temperature to stay at saturation temperature each time step using this law.

When using the compressive interface tracking method with a slope limiter value of 1, I've manage to validate the film thickness against the Nusselt theory. However my heat transfer coefficient/heat flux values are completly wrong. I obtain a value of around 2500 W/m2K when it should be 9000 W/m2K. I've even set the reference temperature to be equal to saturation. The same can be said for the temperature and velocity profiles. Also, the temperature at the interface does not equal saturation as it is meant to. This can be seen on the picture

Although, when using the geometric reconstruct, the film thickness is much larger than it should be but however, I obtain higher heat transfer coeffcients than I did with the compressive method (4500 W/m2K). And the interface temperature is closer to saturation (369K).

I'm not sure why the interface temperature for the compressive method is much lower than saturation but I obtain the correct film thickness, not to mention having much lower heat transfer coefficients as the heat transfer coefficient is a function of the film thickness due to thermal resistance?

Kind regards.

• Rob
Ansys Employee

How well refined is the mesh, and how's the convergence looking?

• Navsing
Subscriber

Convergence and mesh is fine. I also looked at the ansys users guide and said that compressive is typically the best for evaporation/condensation problems due to it being able to capture the diffusive interface as well as the sharp interface. And the slope limiter is essentially the degree of compression at the interface.

• Rob
Ansys Employee

How many cells over the gas/liquid layer?

• Navsing
Subscriber

Just over ten cells on average across the tube. I take the interface as the a volume fraction of 0.5.

• Rob
Ansys Employee

Why the 0.5 limit to condensate?

• Navsing
Subscriber

its typically the value to determine the interface criteria without interface reconstruction (i.e. geo-recontruct) from reading into the literature.

• Rob
Ansys Employee

It is, but that's when finding the free surface postion. I take it the condensate volume fraction is 1 at the wall?

• Navsing
Subscriber

Yes from when post-processing, I take the interface as 0.5. But from the simulation persepective, I think it uses a volume fraction of 1 to determine when the condensate forms.

• Rob
Ansys Employee

Not sure. You mention the heat flux is out as well? Does the heat in/out correctly match the latent heat?

• Navsing
Subscriber

I'm not sure about this. I do include the latent heat in the energy source term from the alternative-energy-treatment tui command. Here is the total heat transfer rate on my boundary conditions:

• Navsing
Subscriber

latent heat of vapourisation for water is 2257 kj/kg

• Rob
Ansys Employee

Check gas mass in and gas & liquid mass out. You're trying to see if what goes in comes out again.

• Navsing
Subscriber

Here are the values of the mass in and out for each of the phases. Just to note, the mass transfer rate in the udf is measured in kg/m3s.

• Rob
Ansys Employee

Where's the inlet?

• Navsing
Subscriber

There is no inlet. I just have a pressure-outlet on the outer fluid domain. I initalize the temperature of the fluid domain to be equal to saturation and set the wall temeprature below saturation to create the mass transfer between the vapour and the wall.

• Rob
Ansys Employee

With condensation I'd expect flow into the domain as the vapour condenses and temperature drops.

• Navsing
Subscriber

So I should have a pressure inlet as well as an outlet?

• Rob
Ansys Employee

Pressure Outlet is fine provided the backflow condition is sensible.

• Navsing
Subscriber

For the pressure outlet, the backflow conditions for temperature are equal to saturation and has a water-vapor volume fraction of 1.

• Navsing
Subscriber

Plus when checking the velocity vector contours, I do have flow coming into the domain from the outlet towards the tube wall.

• Rob
Ansys Employee

OK. If the temperature in the fluid domain is below saturation temperature you’ve potentially got a problem: read up on the VOF model and how it works.

It’s too early, I need more tea. Positive mass flow is into the domain. But you should hit zero when the film is in equilibrium with the surroundings.

• Navsing
Subscriber