Inverter powered battery heating

[johu]

I want to take up the discussion from the old forum concerning using the inverter as a battery heater. The technical background can be found in this paper.

Basically you connect the pack midpoint to an AC output terminal, say of phase 1. Then you apply the described pulse pattern to phase 2. This will take some charge from pack half one, store in in the inductor (a motor winding) and then feed it into pack half two. Then it takes some charge from pack half two and sends it to pack half one. So really short (0.2-1ms) charge/discharge cycles.

It turns out, the lower the frequency, the higher the current.

The internal resistance of each cell will convert this current into heat and heat up the cell from inside out.

This would relieve you from using heating mats, fans /w PTC element or water pipes.

I have written the code for it, basically it creates a variable frequency PWM with a large deadtime.

Hardware wise you would need a relay that connects the packs mid point to a phase terminal.

[PRB617]

Don't even think about heating with a leisure battery. They simply cannot store enough power.
for other info see inverter review

[arber333]

@ PRB617: i think you misunderstood, the point is to use motor windings + inverter transistors to excite current flow through cells while stationary.
That way one could sit in a car turn on seat heater and/or Webasto and wait some time for battery to raise temperature. It is ingenious and if it is successful you can sell the concept to new EV aviation industry. There every gram counts .

However i am concerned with some of the details @johu
1. In case of BLDC motor will it not try to move if there will be alternatively polarized force applied to 1/3 of stator winding?
2. Wont the DC link cap interfere with the pulses? I bet the actual pulse will be more like a slope of a hill?
3. Will cells really heat up from its own energy appliead in short jolts? Did you measure the actuall dT/dt on 2S wiring?
I have 4S cell 10Ah LiFe configuration here. Also i have 600uH inductor taken from 5KW UPS and a 4700uF cap. Would you send me a simple wiring for a halfh bridge IGBT and inductor. I could use your R1 main board for a brain board. Cells are so small i bet i could measure dT soon. I would use 6mm2 wires.

[johu]

Yes thanks for the clarification.

I made a schematic how you would connect the battery. Like in charge mode phase 2 is used for switching.
Software-wise you have to put the inverter in opmode 6. Ampnom controls the the strength of the heating pulses, so to say. Below 20% nothing happens, then it current starts to increase. No regulation yet of any kind.

Your questions:
1. I will have to find out. A sync motor might spin a fraction of a turn but I don't expect it to actually spool up. Induction motor will not move.
2. Yes the bus cap actually sucks away about 50% of the current in my simulation
3. I expect the internal resistance of a cell to actually react like a resistor. It heats up according to P=I²R. So if we generate 20A RMS and assume an internal resistance of 8mOhm (in my Polo the battery sags by 100V at 80A and -5°C - so 100V/80A/152cells=8mOhm) we would generate 3.3W per cell or 500W for the pack.

Would be awesome if you could make a small test setup and report back. For a 4S pack you would need something like 10µH, otherwise the current will be too low. And capacitor should be film similar to what is used in OEM inverters.

[thingstodo]

Hi all,

I'm a newbie here ... perhaps I don't understand the idea ... so forgive me if this has been covered elsewhere

My lithium battery packs are built from salvage leaf cells. It is my understanding that it is *REALLY BAD* to charge them when they are below 0C, one of the reasons for the battery heating in the first place.

So if you can't charge below 0C, how does splitting the pack and using one half the pack to charge the other half of the pack ... when you shouldn't charge the pack AT ALL unless it is warm, work?

I understand .. mostly .. taking half the voltage, pulling some current, using the inductor in the motor for a boost so that it is higher voltage than the other half of the pack. The part I have issues with, or would like some more information on, is how you can charge half the pack, even a bit, when it should not be charged until it is warm. You can't warm it up until you pull out some current, but you can't really put any of the energy back into the pack ... right?

Is there an exception to the rule that I missed?

[johu]

So the point is charging at low temperatures is not generally "REALLY BAD". I can try digging out a paper where they charged cells at different temperatures and currents and the eventually took them apart for inspection. Only high currents (C-rates) AND cold temperatures are detrimental because that causes lithium plating. So for example the Nissan Leaf with its non-temperature controlled battery pack doesn't inhibit charging at low temperatures, it just derates charging current.

Now, the battery heating actually happens at high C-rates (about 0.5-1C) but pulsed. That is also known not to cause lithium plating.

But still, no reason to be completely sure, maybe there is a reason why not a single OEM uses that battery heating scheme. But that could very well be lack of imagination on their side.

[thingstodo]

So the point is charging at low temperatures is not generally "REALLY BAD". I can try digging out a paper where they charged cells at different temperatures and currents and the eventually took them apart for inspection. Only high currents (C-rates) AND cold temperatures are detrimental because that causes lithium plating. So for example the Nissan Leaf with its non-temperature controlled battery pack doesn't inhibit charging at low temperatures, it just derates charging current.

So charging the packs below 0C does not destroy them. But I need a charger that derates the current. Are there guidelines for the de-rating, or a curve to follow?

Now, the battery heating actually happens at high C-rates (about 0.5-1C) but pulsed. That is also known not to cause lithium plating.

OK. But any energy coming out of half-pack A is going into half-pack B. So the average C rate is what counts?

But still, no reason to be completely sure, maybe there is a reason why not a single OEM uses that battery heating scheme. But that could very well be lack of imagination on their side.

My 2011 polaris Ranger EV, that uses a Sevcon controller, won't even do regenerative braking when the temperature goes below -20C. Hmm. I guess I should say that it does not do NOTICEABLE regenerative braking below -20C. I do notice regen below 0C, but it is not nearly as aggressive. The brakes squeak since I actually USE the brake pads.

I expect that the SEVCON controller has an internal temperature sensor.

[johu]

So charging the packs below 0C does not destroy them. But I need a charger that derates the current. Are there guidelines for the de-rating, or a curve to follow?

Yes. I can't find that paper. It might have been in German. I'll keep looking.
I roughly remember that 1/8C was considered unharmful even at -20°C. E.g. 2.5kW for a 20kWh pack.

OK. But any energy coming out of half-pack A is going into half-pack B. So the average C rate is what counts?

The average C-rate would be 0 as you switch between charging and discharging.

My 2011 polaris Ranger EV, that uses a Sevcon controller, won't even do regenerative braking when the temperature goes below -20C. Hmm. I guess I should say that it does not do NOTICEABLE regenerative braking below -20C. I do notice regen below 0C, but it is not nearly as aggressive. The brakes squeak since I actually USE the brake pads.

I expect that the SEVCON controller has an internal temperature sensor.

Yes that makes sense. Effective regen always ends up with high C-rates especially at higher speeds. I manually turn it down in cold weather (which has been like 7 days this season...)

[Teknomadix]

It turns out, the lower the frequency, the higher the current.

Fascinating!

Read the paper this evening. Good science there. Great find! Thank you for the introduction to this principal.

Industrial Webasto / Eberspächer 'Thermo-Top' hydronic heaters are somewhat common in commercial use through out North America. But for regular Joe and Jane consumer use, these device are mostly unknown. Much less common than in the EU, and so hard to obtain secondhand. Thus adding more necessity for these methods to become a viable means of traction pack cold weather temperature management. Indeed, very desirable and necessary for helping adoption of pure battery powertrains in many North American markets during the long and hard winters. With the traction pack hydronic loop, also maybe adding to the comfort of the riders upon starting out on a frozen day? I can imagine the implementations being quite useful.

Look forward to the experiment findings!

[joromy]

Thinking outside the box = Johannes H. I love it!

But you will not get any less energy consumption, than using any other resistance heating.
And the uncertainty of battery health at low temperatures.
That's maybe why EV manufacturer don't use this.
And if the car already has heatpump, it would be dumb to not use this "free" energy.

So this is only for cost/labor saving?

I would prefer AC heating when plugged in. And DC heating if necessary, on the road.

Sorry Johannes H. for the negative feedback

[Teknomadix]

So this is only for cost/labor saving?

As you may find in reading the material; the research paper presents a novel way –with little to no extraneous apparatus needed, to quickly and efficiently heat the traction battery from −30ºC to 15ºC in as little as 5 minutes!
This in comparison to traditionally applied external pack heating in regards to the amount of time and energy needed to accomplish a similarly suitable internal pack temperature, offers us an efficiency increase of many orders of magnitude.

[joromy]

It's very interesting, have read it, don't get me wrong!

But if you think "normal" use of an EV, in cold winter climate:
Charging, maybe in a garage, over night, with in car heatpump.
And structured work time departure.

If you park outside at work, without charging, then you must rely on good isolation and big thermal mass, and maybe heat the pack some extra before leaving home.....

If you think long term use at this scenario, it may be most cost efficient?

[johu]

It is mainly meant as a measure to reduce component count and yes, the labor to fit these components. In an EV conversion I'm happy about every component that I don't need to squeeze in somewhere where it was never meant to be.
Whether or not this is feasible in a production vehicle is a bit beyond the scope I was thinking in

[johnspark]

I think the concept is novel, but i would want to use a relay with a second contact to create a simple circuit to prevent any other mode of operation for safety reasons.