Finally some progress, cooling layers mk3

More cooling plate woes, however some success finally 

I'm guessing this won't be the only post that I start with. "It's been far too long since my last post"

To be quite frank, I've been at a bit of a loss with the cooling plates for a long time now. There are a few key problems. 

I'm trying to keep them 8mm thick to reduce weight and to avoid having the batteries sit too close to the bonnet.

Safety is a big concern for me, making sure they won’t leak, both under normal vibrations over hundreds of miles and in the event of an accident.

Manufacturing—Building an EV is not cheap by any means, but I still have a theoretical budget to stick to. If money were no object, I’d have a solid CNC billet plate with the cooling track cut out, and a top plate friction stir-welded in place.

 Example of a friction stir welded cooling plate.

Fit and Function—The company I purchased the batteries sell universal aluminium cooling plates. If you’ve watched Electric Classic Cars on TV (Vintage Voltage), you've probably seen these in use. I could use these however they are a universal solution, meaning although they will fit, there are then certain compromises, I wouldn't be able to use these universal plates and still get the 28 modules in. So I do unfortunately need a bespoke solution.
https://shop.fellten.com/shop/coolantplate-cus-ev-battery-module-coolant-plate-double-20843?category=6

See the plates between the modules here in the Universal solution

I'm sure you have all seen the failure that was my bonded solution. It worked ok however it was far too restrictive. Yes I could have solved this by opening up the channels but then I also though about the thickness. It’s not just about the cross-sectional area of the cooling tracks. For example, if you have a thin, flat cooling track with a cross-sectional area of 40mm², it won’t perform the same as a round pipe with the same cross sectional area. The flat one has much more resistance. 


Also the safety aspect, I was very concerned that in the event of an accident what would happen? The bonding glue I used was high-temp and could theoretically do the job, but it wasn’t exactly rated for this application, which worried me. If I spring a leak inside of the pack, this would be a serious problem!

Anyway, all these things were enough to put me off bonding the plates. So I looked at TIG welding. 
I designed up some plates with HUGE cooling tracks in them, there would then be a little shelf cut out of the 8mm plate and then a 2mm filler plate put on top.

See the pictures below, I've probably done a terrible job of explaining all of that.

But this didnt go to plan either. Now I have always said I own a welder but I am not a welder/fabricator, after a load of practice on some tokens and scrap I was happy to give it a go. I started putting down my first few tacks, all was going well, then... crack! The distortion between the 2mm and 8mm plates became too much, and the tacks began breaking.

At that point, I wondered if I just needed to have this done by someone who knows what they’re doing. So I contacted a good friend who is a very accomplished fabricator, he basically told me that he wouldn't advise it. The heat difference between the 2mm and 8mm plates was too much, and the geometry of the capping plate caused it to flex. It’s not like a square tank plate when it can distort the same over a length. Why didn’t I ask for this advice before spending days designing it and £1,000 on laser-cut aluminium? I hear you ask. Well you live and learn I suppose. 

Routing out the step so the blanking plate has something to rest on for welding. 


Here you can see the blanking plate, there will be 2 of these, one on the top and one on the bottom, with then the cooling fluid traveling in the void. 

The blanking plate here sitting on the "shelf".

Here you can see the distortion on the plate. 

 Again, I'm far from a fabricator but you can see the tacks staring to crack already. 

 

So now onto Cooling layer design 3... 
After doing endless research into commercial cooling solutions. I found this technique of copper pipe being "squashed" into a rectangular cut out on a structural aluminium plate. 



After learning from my past mistakes I'd try this on one plate before buying all 4, and also start with the top plate as it is the smallest / cheapest. 

So some more hours in CAD and I finally had a design to send to be manufactured. 

Once it arrived I then had the wonderful task of bending the copper to the matrix shape, I did a few tests to begin with, and even with being extremely careful with the pipe bender some of the inside bends were crimping. I found filling the pipe with sand really helped stop this and made it much easier to work with. So I 3D printed a small funnel that would go directly onto the pipe and began filling. 




Now it was a case of bending this matrix near perfectly, I began by marking the start of the bend, then the end of the bend, that meant I could then take the linear distance to the start of the next bend. Anyway I hope you are following there. 

I 3D printed some guides that I could use to keep checking the matrix was correct, a few mm out and it wouldn't fit in the aluminium plate, so I had to keep things tight to the drawing. 

So with my test complete it was now time to squash it and force it into the plate. In the video above they use a lovely press and do it all in one go, as you may guess I dont have that luxury. For test no1 a block of wood and hammer it was, to my surprise it was actually working, I soon realised the technique of doing the straights first so it locks the cooper into place. When you do the bends without the straights being locked in it tries to straighten out again, but once the straights are locked in then it complies nicely and fills the gap in the aluminium plate. However hitting it with a hammer did make far to many dents, although this will never be seen, I want it to be done properly.


I used a offcut of copper that wasn't long enough but gave me a good test. You can see all the hammer marks. 


The beauty with this design of cooling plate is if I make a mistake in the copper it is only that bit that gets scrapped, rather than the whole plate. So I bent up another matrix and this time used some soft jaws in my vice to do the "squashing". This worked really nicely!

For once I'm actually happy with the cooling plates, we have some success and can begin to progress. 
It might not come across how delighted I am by this, however this has been on my mind and stopping progress for nearly a year now. 



The vice method won't work fully on the full plate as it wont have the reach so I've ordered a workshop press that I can use this to squash the copper matrix far more effectively an get the professional finish I am after. 


Now I finally have a design that I am happy with it's hopefully time to make some progress. All 4 plates are now design and to be ordered. 

£500 worth of AN fittings are now on order to link all 4 plates up. 

Really looking forward to making a huge dent in this over the winter months. It's not a hard target but I would really like to have the pack all wired up or at least mostly there by the start of next year. So that's busbars and cell tap wires.