DIY Electric Car Forums banner

Battery Heater through internal resistance

9K views 29 replies 8 participants last post by  Tony Bogs 
#1 ·
As it's currently winter here in Europe I'm coming to think about battery heating again. I'd already picked some heating film with a total output of 240W.

But now I'm thnking: what if you just run a constant current through the battery? That should heat them from inside out. The voltage drop is pretty significant at low temperatures so even low currents should cause heating.

I'm thinking about an inductor and an IGBT in series that is being clocked at a low duty cycle. Or shifting charge in and out of a capacitor. Whatever it is it shouldn't cause too much heat by itself. So a simple resistor is not a option.

Has anyone ever tried that? Couldn't find anything on the forum
 
#2 ·
There has been a discussion on this subject in "charger control lines".

DCB posted a very interesting link about heating up the battery internal resistance
with an AC current (think 50 to 120A peak). I am going to try HF AC heating.

http://www.diyelectriccar.com/forums/showpost.php?p=714290&postcount=5

An inductor and IGBTs is possible with no external power source.
Store energy in an inductor and deliver the energy back through freewheel diodes in a H bridge once the desired amp level has been reached.
One leg hi side IGBT, lo side diode, the other leg lo side IGBT, hi side diode.
 
#5 ·
Did a simulation that dumps a 250Hz sine wave onto the motor using two of the 3 half bridges. Assuming 500V DC voltage, 1 Ohm total internal resistance and 10mH motor impedance I get an RMS current of 14A with an average value of -2.6A. So 2.6A are actually drained from the battery while 14A are flowing through it back and forth. The current rises as a lower frequency is used.
So basically it's the AC heating from the quoted paper but using onboard power. It can even be done while the charger is plugged in.

Will test this in the car hopefully next week. Have to think of a way to put the inverter into battery heating mode.

Cool thing is that this is a "software heater".
 

Attachments

#8 ·
JHuebner said:
Did a simulation that dumps a 250Hz sine wave onto the motor using two of the 3 half bridges. Assuming 500V DC voltage, 1 Ohm total internal resistance and 10mH motor impedance I get an RMS current of 14A with an average value of -2.6A. So 2.6A are actually drained from the battery while 14A are flowing through it back and forth. The current rises as a lower frequency is used.
The math is easy. Assuming 1 Ohm internal resistance (battery, high value):
Input power: 2.6 * 500 = 1300W. Output = SQR(14) * 1 = 196 W.

So basically it is mainly (~ 85%) a DC load and ~ 15% AC heating.
 
#9 ·
Hmm, seems you're right. Today I found that two mosfets in my simulation were still generic NMOS. After changing that the result looked quite different. I think my simulation is crap.

Will run a test in the actual car. How do I determine heating power and power being pissed away by motor and inverter? I have a clamp meter for DC and AC.
 
#11 ·
Ditto, we know that flat out charging frozen lithium at anything other than a trickle can cause "separation " issues .

Someone should test a single cell for capacity after performing this procedure. Hopefully ac has a different effect than straight charging a frozen battery

Given the main reason to heat a battery is to charge it, it would only make sense to use this technique when the frozen battery needs charging. Coupled with insulation this might be a cheap way to thermally manage batteries.
 
#12 ·
I am aware of a method used for his master thesis by a friend. I'll post and leave any and all use rights up to you (whomever you are) as this information was not passed to me in confidence. The subject was a 2 way DC/DC converter placed mid pack which would charge one half of the battery from the other half, and then vice verse. Shuttling charge back and forth but without any additional storage components (capacitors). I think it worked well and was a small compact device. It did not need the primary charger on.

Just f.i.y.

major
 
#13 ·
Experts in battery chemistry in the mentioned paper indicate that DC and low frequency AC charging is bad. 50Hz is already high enough.

JHuebner said:
How do I determine heating power and power being pissed away by motor and inverter? I have a clamp meter for DC and AC.
I'd measure the AC and DC (average) current in the line from battery to inverter.
And the voltage ripple on the battery. Maximum output heat is Vripple,rms * Iac,rms for a fully resistive battery impedance.
Idc,average * Vbattery,average ~= input power (close enough).
The DC link capacitor in the inverter is a bypass to ground for the generated AC current.
Must have a low value. This issue is addressed in the paper (link) posted by dcb.

My take on AC preheating with power from battery:
A boost converter supplies a voltage inverter with twice the battery voltage (up to 1000V).
The inverter generates the AC current (10kHz) with an inductor between output and battery.

I am aiming for 90 to 95% conversion efficiency. So 5 to 10% loss in booster and inverter. Up to 100Arms.

Lower amp values close to lowest battery operating temperature (discharge), Usually -30 to -20C. Below that no AC preheating.
 
#14 ·
JHuebner said:
The paper dcb posted is no longer available
This link to the paper dcb posted is still active

http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.216.8902&rep=rep1&type=pdf

In the paper AC preheating is supported by a charger to supply the necessary power and keep the battery DC current 0.

In the schematics in the paper a circuit with an external IGBT inverter + inductor is used.
It takes only a few components (IGBT/mosfet/diodes) to convert a buck output stage of a charger into a combined booster / single leg inverter. The inductor is already there.
More on this can be found in my thread on the SiC/LLC modular charger design
http://www.diyelectriccar.com/forums/showthread.php/sic-llc-modular-charger-design-162082.html
 
#15 · (Edited)
This paper http://epg.eng.ox.ac.uk/sites/default/files/Howey/evs27_Howey_full.pdf shows that it is possible to inject ac current
in the battery internal impedance with the aid of an EV drivetrain.
Figure 9 shows the presence of DC-LINK cap.

The goal of the authors is to measure the battery impedance in the 1 to 2000 Hz range.
An acceptable accuracy can be achieved with a low cost measurement system.
 
#16 ·
Yes, cool paper (the link has an excess http at the end), thanks for sharing :)

I couldn't find any info on how they excited the DC bus.
I did some experiments of my own yesterday which show "room for improvement". I excited the motor with a 1000Hz sine. Basically sin(w) on Phase one and sin(w+180°) on phase two. That should produce that same sine wave on the DC terminals.

Results:

  • DC current 12A, AC current 16A
  • Prohibitively loud (could be used to honk at people)
Will do more experiments with different frequencies and amplitudes
 
#18 ·
Just did more testing.
First of all my algorithm was crap it was actually dumping a large DC component on the motor. Now thats fixed.

I could only achieve about 2A AC current at 100Hz before the noise gets too loud. At higher frequencies like 1kHz it looks like that precious AC component is being sucked up by the DC capacitor.

This was with my 24V (lead acid) test rig, will see how things go in the car at 500V.
 
#19 ·
Not much better results in the car. Could reach 8A AC at 4A DC. This produced 80A in the stator winding. So it's more like a motor heater...

It looks like high frequencies don't make it out of the inverter because of the DC link cap while low frequencies drive the stator in to saturation and cause heat in there.

Next idea would be to use the kinetic energy stored in the spinning rotor. Basically short accelerate/regen cycles. Would also be fairly noisy and you shouldn't forget to put your transmission in neutral...

Any other ideas?
 
#20 · (Edited)
Sure, assuming that a little extra hardware (you have already picked up heating film) is acceptable.

AC preheating with the aid of a series resonant tank.

Unfortunately uploading timed out. I'll post the schematic of the circuit later on. It has an inductor, a capacitor (bank) and two IGBT/DIODE combinations.

About 70A peak (~50Arms).

 

Attachments

#21 ·
Just realized that I missed a couple of posts further up the page. Sorry about that.

The mid-pack tap looks quite promising. It's just a half-bridge with a little micro generating the simple pulse pattern.

I'm just working on a compact half bridge module anyway.

No, I haven't picked up heating film.
 

Attachments

#22 ·
I really like the resonant idea. Very easy to retrofit, compatible with all BMS variants, easy control (ATTiny85 SOIC8) and a very high efficiency. The parts are easy to get.

I'm going to spin off another thread for the design.
Or maybe use it as an option in my modular charger design.
 
#23 ·
Just occurred to me, why not just pull the same trick as with the boost mode charger.

Use a relay to connect center pack to one of the motor phases while driving another phase with the suggested pulse pattern.

Will check on my bench to make sure the large bus cap doesn't interfere with these plans. In the paper they suggest 20µF, I have almost 1000µF.
 

Attachments

#24 ·
I have done the math for the 2uF input cap (with 22R series resistor) of the HV buck converter that powers the low voltage part of my AC resonant circuit design.
The input is the battery pack. Assuming a 1R internal resistance and 25Arms output of the resonant tank,
about 0,7 Arms will flow through the input cap (and its series resistor).
The math for the resonant pre-heater is really simple, because the current is purely sinusoidal.

So I am going to lower the input cap to 0.22 uF.
 
#25 ·
I would think that you would be "using up" battery cycles by doing this. Better to have an enclosed battery compartment with a small resistance heater. I used a 300w heater in both front and rear batt compartments on my ev Golf. I have found, though that if the outside temp (and batt temp) is above -10celsius, that I can drive the batteries warm in about 25km. My bms will not allow charging if any cell is below -5c. I'm running 96v, 200ah cells. Driving them warm works good good for me most of the time. Heating them with a heater for 24hrs costs 14.4kw, about the same as driving the car 100 km! With that equation, it costs more to heat than to drive.
 
#26 ·
Well with LFP cells I wouldn't be to worried about cycle life. I'd reach 360,000km if I were to use the full 3000 cycles and thats unlikely going to happen within the cells shelve life.

Anyway, the point of heating through internal resistance
a) Cells are heated from inside out which is much more efficient than outside in
b) Less/almost no additional hardware needed

I'd mostly use battery heating on trips that get close to the maximum range. In that case driving the batteries warm is not an option as you've already lost too much range once they are warm. Practically I struggle to get 90km in winter (with the heater off) but safely get 120km in summer.
 
#29 ·
I think the logic of using a trickle or constant current source to keep batteries warm is seriously flawed.

Assuming the batteries are fully charged up, if you force current into a fully charged Lithium battery is a recipe for destruction and/or fire. You would be seriously over charging the battery.
 
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