RC LIPO Combustion Process

[Punx0r]

Having spent some time last night watching videos of people deliberately lighting off RC lipo pouch batteries by over charging (and in some cases mechanical puncturing after over charging), I'd like to discuss the probable combustion process and how it might be best contained.

This is all based on <5Ah, <6S single packs. High voltage packs may be different. I saw no evidence of electrical arcing or plasma.

Combustion

The combustion appears to be a mini-BLEVE (Boiling Liquid Expanding Vapour Explosion). Heat within the cell appears to boil the flammable solvent and create internal pressure. At some point the combination of heat and pressure causes mechanical failure of the pouch and the gaseous solvent escapes as a vapour cloud. If there is any remaining liquid solvent, some likely flashes to vapour due to the sudden pressure drop.

Having mixed with atmospheric oxygen the cloud ignites. I'm unsure whether this is because the solvent is already heated above its auto-ignition temperature, or if a short in the cell caused by mechanical damage during rupture provides an ignition source.

This creates the sudden fireball that lasts for a fraction of a second. The plastic remains of the pack then continue to burn, perhaps along with any remaining liquid solvent for several minutes.

Mitigation

Clearly, a non-flammable casing for the battery is a good idea. There's no point literally adding fuel to the fire. Preventing the second-stage of combustion, where the plastic burns should also be pretty simple - just exclude air.

What would be ideal is to prevent combustion of the gaseous solvent when it is released. Conventional wisdom seems to be to have a fireproof case/enclosure, but to add a vent to safely release pressure. This seems intuitively to be a good idea, but we must avoid the situation of this act simply becoming a flame-thrower. If a pack is the ignition source of the vapour cloud then a metal gauze over the vent will act like a Davy lamp and prevent combustion. If the solvent is above auto-ignition temperature then this won't work, although a sintered vent material might sufficiently cool the escaping vapour to prevent auto-ignition when it reaches the air outside.

A compromise needs to be struck regarding inside volume to allow some room for expansion (to reduce pressure) and to limit the amount of oxygen available for combustion inside the case.

Prevention

Mechanical failure of the pouch and subsequent combustion following extensive swelling and heating. Either of these could be detected by a temperature or pressure/strain sensor attached to each pack. It should be quite simple to create a circuit that will break the mains connection to the charger via a contactor based on input from such a sensor. I'd recommend the mains isolation option as an over-charging event could be caused by a faulty charger or BMS and we should not rely on them to stop charging based on sensor input.

Caveat

This post supposes the use of RC lipo because an application demands it, either because alternative battery chemistries or formats are inadequate on technical grounds or not locally available. This post is about trying to mitigate against the particular risks involved in using it. Not using is not considered a solution. I say this because the topic is not about how much safer lifepo4 or 2C 18650 cells are.

 

[dogman dan]

I think your evaluation of the process is pretty good. I have taken the approach of trying to be sure the "flame thrower" is directed in a safe direction. But I do this only during storage, while I'm gone or asleep.

Packs are stored in metal boxes, and the boxes have vent holes pointed into the firebox of my fireplace. The hope is that any fire will not reach the ceiling. The flames will shoot into the fireplace and the tile surrounding it. But smoke damage would be severe, since the damper is not kept open on the fireplace.

While charging, the metal boxes are open, but I'm around and there is a smoke alarm close by. Very close to the fireplace is a door to outside, and a flat shovel for heaving a flaming metal box out the door. Fire extiguishers near, for putting out the wall or ceiling.

My pack is 25 ah, but most of it is in 5 ah sections, each one in it's own metal box. Thus, the flames will not be so big I can't shovel it out the door, box and all. First step, use the shovel to flip the top closed.

 

[dnmun]

the best way to stop the thermal runaway is to drain the voltage off of the pack immediately. do not use a dead short but something that still rapidly uses up the charge on top of the full charge state so there is not the continuation of the chemical reaction of the electrolyte with the carbon of the cathode in the cell. CO2 fire extinguisher will cool it off too so that helps terminate the thermal runaway.

 

[Punx0r]

Sounds like a good idea. Using a double-pole contactor for the emergency circuit could simultaneously disconnect the batteries from the chargers and connect them to a resistive load (bulbs or heating coil). In this case it would be better to disconenct the battery from the charger than the charger from the mains.

Projector lamps can be had in 24V 250W packages and take up little space.

 

[parabellum]

There are small automatic/manual turbine extinguisher systems, they are about handball size capable to extinguish flaming 700hp kerosene turbine (walter m601) for example, in 1 burst.
http://www.turbinerotables.com/safety.html
They are strictly regulated for aviation use but worth nothing after they reach resource life (on documents) :wink:. Very scary when they blast as well.

 

[dnmun]

Sounds like a good idea. Using a double-pole contactor for the emergency circuit could simultaneously disconnect the batteries from the chargers and connect them to a resistive load (bulbs or heating coil). In this case it would be better to disconenct the battery from the charger than the charger from the mains.

Projector lamps can be had in 24V 250W packages and take up little space.

duh, i expected you would disconnect the charger but all you have to do is open the throttle wide open and put on the brakes enuff to make a load for the motor and the voltage will drop instantly.

 

[liveforphysics]

Typically, the SEI starts mechanically breaking down from drastic over-lithiation in extreme over charge, which causes the exposed anode material to react with solvent and start off-gassing (this is why they start with a gentle puff at first, then run-away).

Contrary to what various marketing BS would try to get you to believe, it doesn't matter if your cathode is LiFePO4 or LiCoO2 or whatever, because it's the cathode wasn't the weak-link that started run-a-way anyways (at least when looking at the failure mode of over-charging.)

The mesh cage idea may have some promise to it, though it would need to be tested before I would believe it. I know those safety lamps are awesome and contain the flame with mesh, but I also know if you simply exceed a certain gas flow rate through the mesh you defeat it. It may require an impractically large mesh bag of many layers.

Once, over the course of 1 by 1 overcharging LiPo cells from the same 4s pack I had cut apart, I had all different types of failure modes. Not all involved flame, some were just thermal events where it just had a rush of dense white solvent vapor that never ignited. That failure mode is a bummer when you're wanting a satisfying fireball, but it would sure be a lot nicer if it was happening on your bike than the fireball.

It's worth exploring if a mesh cage can make it less likely to ignite.

 

[liveforphysics]

Sounds like a good idea. Using a double-pole contactor for the emergency circuit could simultaneously disconnect the batteries from the chargers and connect them to a resistive load (bulbs or heating coil). In this case it would be better to disconenct the battery from the charger than the charger from the mains.

Projector lamps can be had in 24V 250W packages and take up little space.

duh, i expected you would disconnect the charger but all you have to do is open the throttle wide open and put on the brakes enuff to make a load for the motor and the voltage will drop instantly.

Once you've overcharged enough to mechanically crack-up the SEI growth layer over the anode, you can't save it by simply dipping the voltage down, it's going to have the exposed anode surface start reacting just like it's going through formation all over again, but this time it won't be a gentle controlled trickle of current, it will be whatever current the cell can dump into itself.

 

[Punx0r]

Good info. So there is a tipping point where runaway begins and the process is unstopable? Roughly at what point does this occur? I'm trying to establish if there is a useful window between "completely abnormal heating or swelling" and run-away where a safety device could cut in.

This is the sort of thing I have in mind for a flame-trap style exhaust:

Exhaust silencer for pneumatic actuators. Cheap, widely available, easy to fit, fire-proof, tiny orifice size and high flow rate.

 

[fechter]

There has been some speculation that the cells could produce a small amount of oxygen during breakdown and could allow for combustion even in the absence of atmospheric oxygen. The total amount of O2 released from the cell (if there is any) will be quite small, so if air is prevented from reaching the solvent, it should keep it from burning.

Once a cell fails, essentially all the stored electricity will get turned into heat as the plates short out. Even if you avoid combustion, there will be a significant amount of heat relased. This is why a plastic housing would probably fail to contain it.

My observation of exploding Lipo cells is after the initial fireball, the remainder just looks like burning solvent/plastic stuff and should be easy enough to stop if air is kept away.

I have an army surplus ammo can I'm going to use for one pack made from 10Ahr prismatic pouches. I'm still considering the various vent designs and I'm also concerned about solvent vapor igniting outside the can and turning into a flame thrower. If there is ignition inside the can, then any kind of screen or sintered mesh will probaby be useless.

Another approach I considered is to have a vent hole on the can that's covered with a thin stainless steel 'flapper' that looks like a reed valve on a 2-stroke. Under pressure the flap opens and vents, then spring force closes the vent. You might still get a breif flame thrower action, but once the vent closes, the remaining solvent should not burn.

I think we need a test (with video!).

 

[dogman dan]

I can see it coming, exhaust pipes on ebikes. Point that fire behind you. Seriously, not a joke. A motorcycle exhaust handles a lot of hot gasses safely.

 

[amberwolf]

I think we need a test (with video!).

I was gonna do a test of dead lipo pouches and other types of cells inside ammocans vented in at least one way, but the bad cells/packs I was to do the test with were stolen (probably by the housefire repair/cleanup crew) along with the BBQ grill /etc that they were stored in, some months back. If I ever get more bad cells to test with, I'll be doing that test with video, in different ways if there are enough cells to do it with. (but each test should be done with enough cells to make a pack of similar size to what would actually be used in the ammocan).

 

[Samd]

An ammo can and those sintered vents would make for an interesting trial. And the exhaust tube idea from motorbikes I guess provides a chamber out with limited oxygen exposure back to the heat source.

Imagine if developers at the motorbike had abandoned the idea due to being afraid of containing an onboard fire.

Is there a chance that without adequate ventilation that an ammo can could collect pent up gas inside and cause a more energetic explosion in the long run? I guess what I am trying to gauge is whether you want lots of free space inside the ammo can around the pouches or not?

 

[Punx0r]

I imagine the pressure generated by the cell vapouring its solvent is fairly low and readily containable by a strong case. Having plenty of room for expansion also means having plenty of oxygen inside the case, which if you strike it unlucky could cause an explosive mixture when the cell pops and releases the solvent vapour. An ammo can might be strong enough to contain that but if it let go it would be a big "bang".

However, that would be pretty unlucky and I suspect that even in something as large as a 50 cal ammo can there would be too much solvent, making the mixture too rich to combust. If the amount of solvent in a typical 5Ah cell was known we could make an estimate. Limits if combustibility would probably be around 3-5% like many solvents.