[bvip]

Hi,

I am trying to simulate a battery pack cooling system for an fsae vehicle. I have managed to use the built in battery models to simulate the cells.

Current is drawn from the cells, generating heat. This is simulated using a user-define 0D RCR equivalent circuit model. Cooler air flows over the cells, cooling the cells down.

A test module contains 18 cells. Simulating all of these cells individually uses up a lot of computer resources. The cells will all be under the exact same load, and will therefore produce the same amount of heat.

I created the mesh for a single cell, then used a mesh pattern for the other cells. To do this, the first cell needs to have a mesh created before using the pattern. However, I also need to create a conformal mesh between the cells and the air to get the best heat transfer results. To do this, all parts need to be within the same automated mesh. I don't know how to both create a conformal mesh and use the mesh pattern. The aim is to reduce the time taken to get results for the same mesh size and other settings.

I have attached an image just to show the current arrangement of the cells. Any help is appreciated.

v13 6x3 Module Simulation_Temperature - Surface upload.jpg
v13 6x3 Module Simulation_Temperature - Planes upload.jpg

[bvip]

I thought I would try to test whether the software is calculating the heat generated for each cell separately.

I ran a simulation with a single cell, and then again with 5 cells. All other factors were kept the same, and the simulation completed in nearly the same amount of time (within 5% difference).

So the software isn't simulating the repeated cells separately. This thread can be closed, although I am not sure how to do that, sorry.

[cramr5]

Hello,

I have never used the pattern function but a couple of things come to mind they can help you.

1) do you need to simulate the 18 cells exactly? I would be very tempted to split the model in 2 and use a symmetry plane to simulate just half the model.

2) Maybe you can start using a model of the cells only and adding a convection heat transfer on the surface and get a 1st estimate.

Also, which hardware are you using? how many cells this model has? when you talk about "Simulating all of these cells individually uses up a lot of computer resources" what do you mean? takes 1 day to simulate? 1 week? do you have any acces to a University cluster?

Edit: Oh, I see now that you are using the battery model of Star-CCM+. Is that maybe including extra calculations? maybe you can extract the heat generated for the cells in W first and the add that to the CFD model as a heat source and you avoid having all this models interconnected and working at the same time (which is nice but I am sure it complicates the model a lot and maybe not needed right now)

[cwl]

Originally Posted by bvip

Hi,

However, I also need to create a conformal mesh between the cells and the air to get the best heat transfer results.

You won't get conformal mesh with patterned mesh.

What's your issue with non-conformal mesh?