DIY Electric Car Forums banner

Cooling capacity required for Tesla battery pack while fastcharging

17K views 22 replies 10 participants last post by  GoElectric 
#1 ·
In my EV conversion I probably will be using Tesla battery modules and I am still in the process of investigating all ins and outs for doing so in a safe way.

One of the things is battery temperature management. Tesla has quite an advanced thermal management system which has a passive cooling target of 30,0 C, and active cooling target of 55,0 C and an active heating target of 8,0 degrees.
For active cooling they even use the airconditioning I believe, but that is only during fast charging (at 1.4C).

I won't push the limits that much and will only fastcharge at 0,85C max (and not even all the way, but until about 60% SOC).

My question is: How much heat does this create?
In other words, what cooling capacity do I need?

Have been searching a lot on the web, but have not been able to find any info. Does not necessarily need to be Tesla specific. Similar 18650 Lithium Ion info is also OK. I just need to get an idea about the number of radiators and radiator size.
Thanks in advance for thinking along.
 
See less See more
#2 ·
How many packs will you use in series?

plus what is the max ambient you are expecting? Since you live in the Netherlands (judging form your username and posts) I would say just a radiator setup should be fine.

Will you be monitoring the thermisters in the packs? I would suggest monitoring the temperature of the outflow from the packs and have it control an extra cooling fan for a nice two stage temperature control.
 
#3 ·
My plan is to use 8 modules in series (3 front, 5 back).

If the system can handle an ambient temperature of 25 degrees is sufficient.
Above that does not happen very often (I am based in The Netherlands indeed) and in those cases I can reduce the charging current.
Each module has two temperature sensors, so 16 in total which will all be monitored.
I agree a radiator + switchable fan will do , but my question still is:
- What size radiator do I need (BTU)?
- How many? Is one sufficient or is one in the front and one in the back in series needed.
For dimensioning the system I was hoping to find some info on heat development. Especially during fast charging since that will be the most critical point.
My peak discharge will be 3C (60 sec. max) which then will also be no problem I guess if I have enough capacity during fast charging.
 
#4 ·
No one, atleast that will share the data, has done this testing. I believe that a radiator from lets say a offroad motorbike should be plenty.

Due to your pack arrangement finding a pump that will have enough pressure might be a little bigger challenge. However the easiest way to test it is measure. Charge one pack at the given current and just have a continuous flow of water through it. Measure the cell temp and the coolant temp plus measure the flow once, this should give you enough data.

For the rest I would just build it and then test it, if it runs too hot get a bigger radiator. You should have plenty of space in such a large Volvo.

If all else fails, add two valves to your coolant loop for the inverter/motor, and cool your batteries during charging through your main rad.
 
#5 ·
Thanks for your reply.
I was hoping for some generic 18650 Li-Ion curve for charge/discharge rate (in C) versus heat development (in W).
In the meantime I learned that Joule effect is a useful term to search further.
On this page I found a promising comparison between a 60 kWh EV and a 6 kWh HEV pack. http://www.mpoweruk.com/thermal.htm#ev
Larger means low temperature rise at least at 1C.

However the easiest way to test it is measure. Charge one pack at the given current and just have a continuous flow of water through it. Measure the cell temp and the coolant temp plus measure the flow once, this should give you enough data.

If all else fails, add two valves to your coolant loop for the inverter/motor, and cool your batteries during charging through your main rad.
Good idea! Those valves will be there already to heat the battery when needed but can indeed also be used while standing still and fast charging since the motor and inverter will not produce any heat. On the other hand, the coolant temp in that loop might be quite hot from driving.

Good idea to just try to measure. I do not have the equipment for doing so (yet), but perhaps my conversion partner New Electric has.
And indeed as you say, just a relative small radiator will probably do and I can always install another with more cooling capacity.
 
#6 ·
Hi
Another way to look at this is to look at cooling capacity v possible power requirements

The radiator on a typical car has to be able to dump a bit more than the peak power - 150Kw? that that car can produce - in Arizona at 40C
BUT it can operate at 100C - delta T of 60C

You want a T max of say 40C (better keeping it cool) with a 25C environment -
delta T 15C

Supercharger - 100Kw?? - say 90%?? (I suspect its a lot better than that) so 10Kw of heat

So you want to be able to dump 10Kw with a delta T of 15C - equivalent to a car/bike with a 40Kw engine
 
#7 · (Edited)
Some typical discharge/temp curves for a Sanyo/panasonic BF cell (simiilar, but not a Tesla cell !)..
http://lygte-info.dk/review/batteries2012/Sanyo NCR18650BF 3350mAh (Red) UK.html
Suggests a 35C temp increase under constant 7 amp discharge (~30 mins).
I dont think charging at 0.85C (~3amps ?) is going to be a significant heat issue.
The uses feedback i have read suggest the main use of Teslas thermal management system, is to heat the pack for cold starts, and to cool the pack after hard discharges in hot ambient conditions ..whilst supercharging !
But as others said,.. It would be easy to do your own tests on a module.
Note.... For $50 you could buy a small RC charger with temp probes and data logging that would let you test a cell under various charge/discharge rates to confirm exaclly what the situation is.
 
#9 ·
Jack at evtv.me has done testing of cells at fast charge rates to see how the temperature rises, http://evtv.me/2014/11/announcing-evccon-europe-2015-march-29/

It seems that the rate didn't cause undue temperature concerns unless it was continued after the cell was reaching fullness. If you are not supercharging or overcharging the cells then you likely don't have to worry about overheating. At a fractional charge rate below 1C what is the mechanism to cause excessive heating?

He may have done some testing on Tesla modules/pack also.
 

Attachments

#11 ·
Unfortunately the IR of a Tesla cell is not known (as far as I know).
I've read on https://teslamotorsclub.com/tmc/thr...-with-up-to-77-kwh-usable.61896/#post-1352579
Further, these cells are almost certainly NOT Panasonic NCR18650B cells, like many have assumed. Close, but not identical. I've tested actual Panasonic NCR18650B cells on the same equipment as the Tesla cells, and the retail Panasonic cells always perform better and have slightly lower internal-resistance.
I've read that at a supercharger Tesla puts 1.4C in and then they use active cooling with the pumps and AC turned on. So at those levels there is something to take care of.My peak will be during fast charging at 100A (probably for about 15 tot 20 min. depending on the SOC and then drop).
Since this is 0,85C and the specs for NCR18650B is 0.5C - 0.7C I think in my case 1/3 of the size of regular radiator that used to be in the car is enough.
Another interesting video/read is this one:
https://batterybro.com/blogs/18650-...3090502-why-do-lithium-ion-batteries-die-long
showing that the Coulombic efficiency of Tesla modules is quite good.
https://cdn.shopify.com/s/files/1/0...chemistry-comparison.jpg?17426464464965184610
 
#15 · (Edited)
.....My peak will be during fast charging at 100A (probably for about 15 tot 20 min. depending on the SOC and then drop).
Since this is 0,85C .....
Curious what charger you plan to use that is able to supply 100A?

I am also using Tesla modules in my conversion. I will be using a salvaged Tesla fluid pump to circulate fluid in the modules and just plan to use a standard sized EV rad that is normally sold for cooling motor/controllers. I am using 2 parallel Brusa NLG513 chargers, which will put out a max of 25A total, which will give me a max charge rate of 0.1C, so I am not worried about cooling during charging, only during discharge. I will likely be using two of these heaters to heat the pack if necessary prior to charging during cold temperatures.
 
#12 · (Edited)
Of course the Tesla cells are not the NCR18650B panasonics.
Those cells are not EV capable realistically.
The
Tesla cells are "custom" in several aspects, but the closest commercial cells are the Pana "BE" or. "PF" models which have the same chemistry, capacity, charge/discharge characteristics, etc etc.
Exactly which cell is in your modules will depend on when it was built as Tesla have changed the cell spec several times..and have even used more than one manufacturer !
The only way you are going to satisfy yourself if you need cooling for charging , is to perform some testing on the actual modules you have

PS... What configuration/ capacity are your Tesla modules ?
..i thought they were usually 68p or 74p configured ?..IE ,.220 Ahr + capacity ?
..in which case your 100A charge would be less than 0.5C ?
 
#14 ·
Hi. For forum members, Lars and I are converting the 'same' car, both with Teslas.

Nothing like running a test, have Eric send you a few loose cells. I suspect you won't need any battery cooling at all, ever, but here is one way of getting an estimate:

I've measured those cells and cell packs. An individual cell is about 35mOhms, and I could not measure any appreciable change when I heated or cooled them within normal limits. Nor did it seem to vary much with SOC. So, just use 35mohms and calculate total amount of heat produced to run a current through them for 20 or 30 minutes. I know you are planning on the double-voltage packs which are 37P instead of 74P.

So if you are charging a 37P 'cell' at 100 amps, that is about 3 amps per cell? For a given charge time, you can use P=IsquaredR to calculate the heat-energy produced in Watts, and convert to Joules (4.12 Joules per Watt, or something like that). Then divide that number by the specific heat of a lithium cell - 0.83 J/gK....

Here is my source for that number - which I have found from several sources, but this one was done using 18650s:

http://www.inforlab-chimie.fr/doc/document_fichier_279.pdf

...When it is all said and done (assuming no heat-loss), you should be able to predict the temperature change of a pack by multiplying by 744 cells and dividing by the weight of a pack. Or use a single cell, if you know its weight.

You need to understand this yourself, so I won't do the calculations for you.

I can't do heat-transfer equations, but if you do need to loose a certain amount of heat, you can multiply by the number of packs, convert units again to BTUs and size a rad accordingly without any equations.

I'll be happy to help you out with these calculations, or accept any corrections from anybody out there - interested in the result.

(Lars: how are you going fitting those 5 modules in the back? Can you get them in there in two rows by expanding the cavity? I can send you some pics of the battery box I built for back there, it is pretty nifty.)
 
#17 ·
Thanks for the actual internal resistance value.
To add some information:
  • I intend to use modified modules so 12s37p instead of the original Tesla config 6s74p. The modules will be 118 Ah and 44,4V.
  • The 100A charger is the CHAdeMO fastcharger (DC). Currently up until 50 kW http://www.chademo.com/wp/technology/optimal/ but will even be 150 kW in the future. My socket is rated for 125A. My nominal voltage will be 355 so my max will be 35 kW.
  • I am now in the process of figuring out the optimal balance between pressure drop and flow in order to still be able to use one pump for the whole system (preferably the Pierburg CWA50). Then the question is, what is the best radiator size to take away the heat produced. I have someone helping me out and will share the results so we can discuss further here.
  • There is enough space to make a stack of two and one of 3 in the back assuming the modules become 9cm high including room for cables and casing. I am doing a wagon and will use both the space of the fuel tank and the spare wheel. In a sedan it is a whole different story since you only have the depth of the fuel tank since the spare wheel sits on the side.
  • Those Calix in-line water heaters look interesting. That will however require to rethink my setup even further since I only choose a 400W DC/DC converter. I was thinking of heating using 220V while plugged in and using heat from the motor circuit while driving (if heating is still needed since the batteries themselves will also generate heat during discharge).
To be continued! Thanks for all in input, haven't even processed and read everything in detail but will do.
 
#18 ·
Building on the calculations and work of Jim in his "Working with Tesla packs" topic I did some further calculations.
I used Joules law and the Isobaric mass heat capacity under a couple of conditions and found the following delta T per module.
- Fastcharge (100A, 20 min) = 7 degrees
- Peak power controller/motor (351A, 0,5 min) = 2 degrees
- Max. continuous power controller/motor (158A, 37 min (pack empty)) = 33 degrees
- Average (guess) (45A, 120 min) = 9 degrees
By the way my pack voltage is 355V nominal
So my impression is that there is nothing that I cannot cool down with one radiator/fan combo but will investigate that further.
My idea is to control the pump speed depending on the temperature to reduce power consumed my the pump.
 
#19 ·
Just did some charging on a Tesla single cell and can confirm low internal-res.
I got about 50°C on a full 5A discharge, thats 1.7C.
Below 1C barely any temperature rise in the cells.

You should correct your energy values, unless you plan on buying modules from a brand new S90. S85 cells are in the 3100mAh range when new.
You will end up with approxx 108Ah @ 44,4V, 1y used cells are at 2900mAh, cirka.

- Fir3Storm
 
#20 ·
Thanks! That is good news.
My motor/controller can only handle the peak (350A) for 30 seconds.
With 12s37p that is about 3C.
Under normal operation, even fastcharging I will be below 1C.

I will keep the radiator, but think in the end warming up during the winter is going to be more important than cooling.

Thanks again for the test and data!
 
#22 ·
I will be using those Pierburg pumps as well.
Currently I have two options in mind.

First one is the Siemens Logo.
I have this PLC on board already for some other tasks.
My idea is to also use it for the thermal management.
Switching on electric heater, open valve to add warmer coolant from inverter circuit, etcetera.
Have not investigated yet, but I am hoping that I can also use it to control the pump speed via PWM.

The alternative is to control the pump speed "stand alone" using tinyCWA
http://www.tecomotive.com/en/products/tinycwa.html
This device is developed to do that. The "intercooler firmware" version matches my expected temperature range, but they can flash custom firmware with custom ranges on request.
 
#23 ·
Hi Lars. Good for you. I'm using a different pump, but as usual there will be some overlap between us. Mine will be controlled via an Arduino. We have variable frequencies, but cannot make the pump respond, it is either on or off only.

Must be some kind of sequence.... Any info appreciated from any-body!

J
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top