Learning Forum

[phgems00]

I am trying to simulate a two-phase heat pipe in ANSYS Fluent and I am currently facing issues defining the porous wick region and achieving a stable convergence. 

The heat pipe is modeled as an axisymmetric 2D geometry consisting of three regions:

• Copper wall

• Porous wick structure

• Internal vapor/liquid volume

 

Boundary Conditions

Thermal boundary conditions are applied on the outer wall:

• Evaporator: constant temperature (heat source)

• Adiabatic section: insulated wall

• Condenser: constant temperature (heat sink)

 

Multiphase Model

I selected the VOF multiphase model with:

• Phase 1: liquid water

• Phase 2: water vapor

The wick region should behave as porous media so that the liquid return flow is driven by capillary forces and pressure gradients.

Therefore I want to assign:

• Porous media properties

• Viscous resistance / permeability

• possibly Darcy–Forchheimer model

to the wick cell zone.

 

Problem

I am unable to correctly define the porous wick region:

1. When the cell zone is defined as mixture, Fluent does not allow me to set the viscous resistance parameters for the porous region.

2. When I change the zone phase (e.g., to water vapor or liquid), I can access the porous parameters, but this obviously breaks the multiphase setup.

3. Changing the phase definition in one cell zone also changes it for the other zones.

This makes it impossible to correctly model the wick as porous media within the multiphase system.

 

Convergence Issue

Whenever I try to run the simulation, It give a non-physical result.

 

Main Questions

1. How should the porous wick region be defined in a VOF multiphase heat pipe simulation?

2. Should the wick be a separate porous fluid zone, or should it be modeled differently?

3. Is the VOF model appropriate for this type of heat pipe simulation, or would another multiphase model be more suitable?

4. Are there recommended initialization strategies for liquid/vapor distribution in heat pipe simulations?

Any advice or references to similar Fluent heat pipe models would be greatly appreciated.

[Rob]

Porous media is set in the cell zone, and if I understand you're seeing mixture, phase-1 & phase-2 options on that cell zone. That's correct, the whole fluid domain(s) is multiphase but some of the settings are at the mixture, or phase threads. So you're not changing the phase, merely the phase you're setting things on. 

Porous media is a lumped model for pressure drop. So, there's nothing to drive a wicking flow: you'll need a UDF for that. 

VOF can work, but you then need to model each and every droplet and film on the walls and wick surfaces (which you may need to add). Eulerian might be a better option. However, all methods require additional user coding and aren't simple. As Ansys (Synopsys) staff are restricted in the level of detail we can provide you'll need to see what the wider community come up with.