Can LiFePO4 Batteries Combust in Thermal Runaway Event?
- Thread starter Will Prowse
- Start date
Any idea what happened? It has been said the prismatics are not suitable for mobile use.. also, the cells needing a physical separator between each cell, compression…
Yep, missed that. I only checked the last couple of comments that were posted today.
The electrolyte is flammable. If a cell is physically damaged, you can catch it on fire. But for an exothermic reaction to occur to cause combustion, from the literature I've read, is impossible with lifepo4. Well, unless you had a pile of kindling next to a spark generator of sorts ? or a dead short next to a fuel source. But to have combustion caused internally, is impossible.
Yessir. It's the reason that an LFP starter battery can be 3 inches under your butt when using them in motorcycle applications, like Shorai, Antigravity, Earth-X and a whole host of others.
Extenuating circumstances, like wrong charging voltages and poor wiring infrastructure, bms failure etc can cause them to cook off the electrolyte and melt poor wiring becoming the primary source of trouble - not the cells themselves.
But at the end of the day, the greedy little iron-phosphate molecules in the active material simply won't release oxygen. So yes, you are absolutely right. It's the other things that if handled casually, are the instigators.
Probably the reason that these companies have been in operation for many years, and not sue-balled out of existence.
Extenuating circumstances, like wrong charging voltages and poor wiring infrastructure, bms failure etc can cause them to cook off the electrolyte and melt poor wiring becoming the primary source of trouble - not the cells themselves.
But at the end of the day, the greedy little iron-phosphate molecules in the active material simply won't release oxygen. So yes, you are absolutely right. It's the other things that if handled casually, are the instigators.
Probably the reason that these companies have been in operation for many years, and not sue-balled out of existence.
Last edited:
Could you elaborate on how you dealt with this? Did you pull it out while it was going crazy? Is the yellow stuff from the cell?
First of all I am glad you and your van are ok and there wasn't more damage.
The cell is severely bloated and I have seen worse bloating and the vent did not pop. But the vent on your cell did. It appears the cell got hot because the PVC wrap is deformed on that cell. So the BMS reported 3.4 volts for that cell?
These cells can be used in a mobile environment if properly compressed to reduce terminal stress. Tape alone isn't going to work.
Cells should be insulated from each other and anything conductive that can come in contact with the cells. The PVC wrap can't be relied on especially in mobile environments.
The bolt/screw torque of the cells terminals should be checked on a regular basis to make sure nothing is coming loose and all connections are good. Do you know how much torque you used on the bolts when you assembled the pack?
If possible I would like to know what happened. Is there any more information you can provide? I would also like to know the AH rating of the cells and the manufacturer. Thanks.
So… I honestly have no idea what happened. This is a 120ah build of cheap aliexpress cells and a overkill BMS. I use a red arc 1225 charge controller to manage the 100w solar panel and alternator/starting battery To keep the cells charged. They have worked flawlessly for a year with delta between cells typically 0.02-0.05 and I think 3 recorded overvoltages, all of which I witnessed via the BMS Bluetooth app. My van has been sitting for 2 weeks with hardly any use and a 12v fridge consuming power so the batteries were discharged around 40% of the 120ah rating The day before. When I got in the van and checked the state of charge it read overvoltage disconnect and a fully charged system. This was odd as there is no way my little solar panel charged up those cells to 100% in a day and overcast skies. Ambient temp outside around 92F so it was a hot day bit the overkill temp sensors were showing 37-38C in the battery box. I drove to the store and watched the BMS to see if it would kick on charging again and it never did. Driving back home I heard a hiss and smoke immediately released into the vehicle. I pulled over and looked under the seat quick where the battery is and the smoke stopped. I checked the BMS and saw the voltage in cell 4 rapidly dropping while the other 3 remained stable. I used the BMS to disconnect the cells and drive home (3min drive). I got home and looked under the seat and saw the cell was significantly deformed and swelling and closee the box and immediately a loud bang and a ridiculous amount of “smoke” (fumes really)…. I ran for a fire extinguisher and was overcome by the fumes and the heat n the battery box and just called 911. Fire department came, put on respirators, and used a yellow powder fire extinguisher to stop the reaction or “put it out” and pulled the cells from the vehicle and that is essentially it. I have a number of fuses setup in the system and none were blown. Zero damage to the other cells and any of the components of the system. Just this one cell failed, cell #4.
In another forum, people are saying the casing on these cells is charged and say the cell was shorted out on the metal battery box Which I find ridiculous. Even if the cell was negatively charged, it’s already grounded to the car so contact with the metal box should not do anything in my estimation. If they are positively charged, I would assume this would be happening a LOT and all 4 cells, being connected by the bus bars, would have been damaged.
The cells were separated with very sticky foam tape and taped together… I believe I saw this done in one of Will’s videos?
I am all ears as to what you guys think happened or I need to do differently to prevent this again. I hope to replace the system as it was working awesome and way better than the lead acid i used to use.
Do you know what the voltages were on all the cells then? It sounds like from what you describe cell 4 was seriously over voltage which would explain why the pack read fully charged even though you thought it shouldn't be. Can you pull cell 4 out carefully and fully examine it?
My best guess at this time: stress on the terminals caused by the movement of the individual cells and transferred through the bus bars damaged the cell internally.
The issue isn’t that negative shorting to case or vehicle ground… remember you have FOUR differing polarities, if ground on one shorts, unless it is the main negative cell, there will be a short circuit. So, yes… a cell casing shorting to ground would be an issue, because in a series build of cells, negative on one cell is connected to the positive of the other cell… thus a series.
And can anyone confirm the casing on these cells is in fact this dangerous and any contact with metal can cause a short circuit?
I can only say for sure that the EVE 280 cells (version E I think, the ones that were available originally last year) can only supply a small leakage current through the case (case is positive). I noted that here:
Heltec BMS' (up to 350A) with Active Balancing & Independent Active Balancers
Maybe check with a light bulb and see if it is just a tiny leakage current, or if it can actually pull current to the case. Some meters are very sensitive and a tiny bit of moisture can carry enough current to show full voltage. But if it won't light a small bulb, it is likely not a concern in...
diysolarforum.com
Other cells may very well be different.
DerpsyDoodler
Solar Addict
- Joined
- Jan 10, 2021
- Messages
- 2,247
wow! where were you when it happened?
I have some 2019 Eve cells, and some Shenzhen TopBand cells that I can test and report back. It might be a good resource to have the case "polarity" of different cells noted.
I was in the van in my driveway!
DerpsyDoodler
Solar Addict
- Joined
- Jan 10, 2021
- Messages
- 2,247
glad you're OK and glad you shared. Maybe you should start your own thread about this. I bet you'll have a lot of interested forensic (shadetree) analysts to help diagnose.
Similar threads
