I am just having some trouble modelling a radiator.
For some reason the simulation indicates that the system can cool from inlet to outlet about 30 degrees; however, in real-life testing the actual system achieved a factor of ten less!
I have set the materials aluminium 6061 casing and water coolant, assigned inlet conditions as 15l/min at 75 degrees, outlet conditions as atmospheric pressure and then put a flow coefficient of 0.024 to model ambient air at a ref temperature of 20 degrees.
In fact, sometimes the simulation gives next to no thermal properties of the system.
What can I do?
Also, if I wanted to put an external airflow through the fins whilst keeping the internal coolant flow how would I go about doing it?
Solved! Go to Solution.
You have applied your film coefficient on both the internal and external walls. It should only be on the external walls.
The easiest way to do that on your model:
- When you are in boundary conditions, the default selection is surfaces and direct picking. Hit the select all icon to select all the surfaces in the model. Next, in the ribbon, change the select by from direct to volume (not volume select, but rather the icon directly above the select previous icon).
- LMB on one of your inlets...you are essentially selecting all of the fluid surfaces (attached to the volume you just clicked on), but since they were already in the selection set, you are unselecting them.
- If you plan on modifying such a large selection set or doing anything with it, it makes sense to RMB at this point somewhere off the model and create a group.
I would try this and see how your analysis results look before putting air flow through the radiator fins. One other thing to consider is running to full convergence (more than 100 iterations are typically required).
Hope this helps.
Product Manager - Simulation CFD
I know this may seem like a stupid question but what exactly is the film coefficient and how do I calculate it for my situation?
Also, for modelling the external air flow through the fins, is there an easy way to do it? My approach creates a giant box around the radiator that is part of the same part if that makes sense?!
Hence the material properties are the same for each...
For typical values, we would use something like 5W/m2/K at ambient temp with just naturally convected air externally. This value might increase to 20 for forced convection, it is an assumption though.
To model the external air, just extrude a cuboid on either side of the radiator, in-line with the fins. This will be your cool air with a flow rate and temp on one side and a P=0 on the other.
In the end I used 30Wm2/K, which may appear high but as I was using a rather simplified model that was missing a large finned surface area I thought it was justified. What do you think? At the very least it matched experimental data.
On the topic of the cuboid extrusion, how do you stop the two fluid, i.e. coolant and external air flow, coming into contact with each other?
If it matches test, I'd say you are OK. although your mesh did look pretty coarse. 4-5 elements through a gap is recommended.
Regarding the 2 fluids touching, just model it so that they do not. Maybe focus the inlet just on the tubes, or add a cap on the coolant which you can supress from the mesh (essentially leaving a hole in the model) where you can apply the flow rate.
Would you like to share an image of what you have?
Okay I will have a look into it!
I have attached what I did, it isn't pretty though!
Thanks for sharing. I know you are already making changes but there are a few things to mention here:
- There is no need for all the film coefficients. You cannot use them internally - remove those. The external ones are unnecesary, this is essentially a just wind tunnel
- The mesh does need to be finer - check out our SimTV eposides on meshing
- The inlet and outlet needs to be much longer, more so downstream. Something like this, assuming this is the downstream side (I used a meshing region to show the dimensions, you will need to adjust the CAD. Notice that is only encompasses the inner section of the rad