Modelling transient heat transfer when opening a fridge
 

Hi,

I have been working on a group project that involves a comparison between different types of doors on commercial refrigerators. I created a 3D CAD model of a small commercial fridge and tried to simulate the loss of energy that is experienced by the fridge when the door is opened for a certain duration. The meshing and setup are mostly fine, so is the geometry, but I have two problems:

I do not know how to animate the movement of a sliding door in 3D (if that is even possible in FLUENT).

I have created the exterior fluid domain - the room - but cannot seem to create the interior domain within the fridge before it is opened. I want cold air in the fridge before it is opened and ambient outside, so that when the door opens the warm air will displace the denser colder air that flows out the bottom.

Note: I have used Boolean to subtract the fridge from the room, and again to unite one of the two sliding doors not used for the analysis with the rest of the fridge. I also considered the interior back panel of the fridge to be a sink and the exterior back panel to be a source. Is this advisable?

I have attached some images for reference, and would really appreciate some insight. Thank you very much. :)

hmm, sounds like an interesting problem but maybe someone who knows more can give you an answer

Fascinating but I'm not quite sure tbh. Perhaps an engineer grad can help you out on this one?

I would recommend, not to simulate the movement of the sliding door. Run the simulation with two cell zones with a wall in between. One cell zone is inside the fridge the other one is the sourrounding. Run it as steady state. When the solution is converged change the type from the boundary between the two cell zones from wall to internal and start the transient solution. You can even split the wall in segments to model a step-by-step opening of the door. Last hint, in reality the fridge is not completly tight. Be sure that you have the same pressure in both of your cell zones, before you open the door.

Hi. Thanks a lot. I'm honestly new to this software, but how exactly do you make the two cell zones different? Because whenever I create the fluid domain it makes the whole thing one cell zone. Also, how can the door type be changed from wall to internal mid-simulation? I didn't think that was possible. Thanks again.

Diffrent cell zones has to be defined in the cad modeler (number of fluid zones/domains). But as I think about it, it works also with only one cell zone. Just define a wall for the door of the fridge with zero thickness (only a face). For door open, the boundary typ is interior. For door closed changed it to wall, in this case fluents create a shadow wall, so that you can distinguish between both sides of the door. See attached pictures.

You have also to add a wall thickness to that boundary to model the insulation of the door.

I am by no means an expert in Fluent, but I second MKuhn solution for opening the door.

However, getting the energy loss information is a bit more tricky.

1. I would separate the interior of the fridge to be just a box in void, with the back wall being the cooling element like in a real fridge. You need to calculate a heat transfer coefficiency for the back wall at the temperature you want to have the interior of the fridge. 5C or whatever is the standard general temperature.

2. Run the simulation in steady state until you converge.

3. Open the door for some amount of time with the method MKuhn suggests.

4. Close the door.

5. After closing the door, measure how much energy is transferred in total "through" the back wall in order to reach the steady state cold temperature again, while running it in transient.

Disclaimer, the insulative properties of the door and the fridge itself might in reality increase the amount of energy lost, as in real world heat is always leaking into the fridge volume.

Thank you for the answers. Would I need to use UDF for the door? Also, I'm still not quite sure how you can change the door type mid-simulation. How do you program it before hand to do that? Also, I don't know how to find the energy expended by a single surface (i.e. the back wall of the fridge). I have played around with it for a while but cannot seem to find the right method.

As MKuhn suggested, you could remove the detailed mesh of the door and replace the door with a 0-thickness plane, and in Fluent give the door a virtual thickness and material in the boundary conditions menu. You can create such plane even just by simply splitting the air volume in Space claim, and specifying this split as a "wall." This allows you to have a door with the correct insulative properties, without having to mesh it.

To "open" the door, you can change that door boundary condition from "wall" to "interior". This simply tells Fluent that the fluid can pass through the plane as if it does not exist.

You can change the state of the boundary condition at any point and continue the simulation from that iteration forward with the different boundary condition. This is also how you can "switch on" and "switch off" a power source for example if you are running an transient simulation.

I think it is also possible to open the door realistically by using frame or mesh motion.

MKuhn also correctly points out that these doors are usually not completely sealing, meaning that over time the pressure outside and inside the fridge equalizes. But for a short amount of time there might be some milli-bar pressure difference after the door is closed, if the air cools down very very quickly. You might have experienced this if you ever put something in the freezer, closed it and tried to open it again immediatedly after, the door will require considerable effort to open. This might have an impact on how the air reacts to when you "open" the door in the simulation, since if there is a lower pressure inside the fridge, it will pull higher pressure warm air in with considerable amount of force when "opened".

No you would not.