Cap cooling of 18650, thermal output graphs?

I am researching 18650 cell cooling, one source said that tab cooling prolongs cell life by reaching into all the layers of the battery, is it the same for cylindrical batteries?

How many watts of thermal energy does a 3000mah 18650 generate at 1,2,5 and 10C?

I found a couple of fun videos of X-ray and IR cameras of batteries:

IR: https://www.youtube.com/watch?v=hwXccpeN6Qc
x-ray: https://youtu.be/4u2ZgFRosoQ

Alta motorcycles draws the negative charge from the shoulder of the 18650 cells, which adds resistance to the circuit, but it also leaves the bottom and sides open for cooling efforts.

They also use a nominal 355V pack so the amps per cell can be lower, to mitigate heat generation and maximize range...

Oh that's new!!! What's the specification for the sides of the cells? It would be tempting to build batteries where the negative contacts were on the sides of the cells. I want to read some good articles about 18650 cooling, i.e. costs and weights and materials for the task.

zzoing said:

I am researching 18650 cell cooling, one source said that tab cooling prolongs cell life by reaching into all the layers of the battery, is it the same for cylindrical batteries?

How many watts of thermal energy does a 3000mah 18650 generate at 1,2,5 and 10C?

I found a couple of fun videos of X-ray and IR cameras of batteries:

IR: https://www.youtube.com/watch?v=hwXccpeN6Qc
x-ray: https://youtu.be/4u2ZgFRosoQ

Click to expand...

In theory, You can calculate the watts of power dissipated by looking at discharge curves such as this example with the old Sony VTC4 cell :

At a given state of charge value (ex: 50% or roughly 1000 mAh), look at the voltage at 2A (1C) versus the voltage a 30A (15C) for example... they would be 3.65V and 3.15 V respectively. That is a voltage sag (dV) of 0.50V for a amperage difference (dI) of 30A-2A=28A... With ohms law (V = R x I, or more usefull here dV = R x dI). So R = dV/dI = 0.50V / 28A = 0.018 Ohms, or 18mOhm.

Knowing that P = V x I or that dP = dV x dI, the power sag (power lost in heat) is dP = 0.50V x 28A = 14W.
Alternatively, P = R x I^2, or that (dP) = R x (dP)^2 = 0.018 Ohms x (28A)^2 = 14W.

That's my way to estimate, but it's reductionist approach, as the data is for one single cell isolated in the air... not a whole battery pack

Matador

Matador

zzoing said:

How many watts of thermal energy does a 3000mah 18650 generate at 1,2,5 and 10C?

Click to expand...

We need to know the DC internal resistance of the cell you are using to answer this question.
To make things even more complicated, DCIR can vary with state of charge... such as in this example for a good old Sony VTC4 cell:

Matador

I have some data you are chasing. Will look for it today.

OK I know where the plotted data are but I won't be able to get them for a few days. But in any case, I suggest you read a paper by Drake et al. (2014) http://www.uta.edu/faculty/jaina/MTL/pubs/Drake-JPS2014.pdf

The thermal conductivity of a cylindrical cell is two orders of magnitude higher in the axial dimension, than that of the radial. However there is a large point of resistance between the end of the jelly roll and the base of the can. It's even worse for the positive end of the cell. Cooling the side of a cell isn't so bad as while there are multiple layers of resistance between each electrode, they all circle around to close to the same point, and critically there is a shot path from the hot jellyroll to the outside of the can. The problem as always is getting a high thermal conductivity from the cell surface to the cooling medium without electrically shorting the cell. Tesla uses a thick rubbery material on its Model S and X modules, but they have used a thermally conductive epoxy resin and silicone encapsulant for the Model 3.

zzoing said:

I am researching 18650 cell cooling, one source said that tab cooling prolongs cell life by reaching into all the layers of the battery, is it the same for cylindrical batteries?

Click to expand...

These tests were done on a high energy storage cell (probably a Panasonic GA cell) and the results of cooling axially versus radially were done. The cooling block was a machined ally block with water flowing through a backing maze; so effectively an impossibly good cooling system. Impossibly because there is no practical way you could build a battery with this sort of cooling system. Axial cooling of only the cell base did work, but it was far less effective than simply cooling the sides. Even 30% coverage of the radial surface was good enough. Cooling both ends would be better, but you would need some pretty impressive materials to conduct the heat away (electrically insulating, thermally conductive = unobtainium).

That said, I believe electric pickup truck company Rivan are using axial cooling on their 20700 cell packs. They admit it's not as effective, but it allows for a much bigger battery pack (which by extension, isn't working as hard under normal conditions).

Even not very effective heat transfer at measuring fixture BF-2A from cell axial surfaces plus and especially minus pole is sufficient enough to cool the cell to some extent. Heat transfer from shell to minus pole is enough to cool down the cell as it is all part of one can. Temperature measured by thermocouple, wire diameter 0,127 mm.










View attachment 4


Compare temperature rise of 30Q cell surface at BF-2A and at modified plastic cell holder. At the same conditions of 2 A charging and discharging, BF-2A fixture is cooling down the cell more than plastic holder.








Of course, the best solution for ebike is not to draw high current from the cell. Overdesign of the battery capacity is always effective way to keep battery cool and to improve lifetime generally. Drawing more than 2 or 3 amps from the cell and expecting good lifetime is nonsense.
Thermally conductive silicon glue can help to conduct the heat from the core and decrease the temperature gradient through the battery.

Thanks, that's great information. Am studying the graphs. Some varieties of Boron Nitride is a good unobtanium thermally conductive electrical insulator. It's a bit difficult to find but it would be fun to have some at home: https://www.accuratus.com/boron.html

here is a video about cap cooling.
https://www.youtube.com/watch?v=_jd8REVB-c8

Aluminum oxide is a good heat conductor and a good insulator and not stupidly expensive. But it is brittle and hard to machine.