My Project Lithium Battery Caught Fire

Didn't mean to sound like I was exiting the thread, just explaining why the replies are bit all over the shop without a lot of direct references made.
I believe this thread is about an LFP battery that caught fire. It wasn't punctured, so the staking is not particularly relevant, yet the fact that additional holes to allow air to enter so oxygen was introduced into the cell was required to get one of the test cells to actually catch fire.

The suggestion hydrogen was a gas in the vented electrolyte vapour, yet the testing was only carried out on burning electrolyte vapour, leaves many unanswered questions. Using that flawed methodology to assume hydrogen is 50% plus of the ejected vapour just muddies the water, more valid testing would be required.
The possibility that the hydrogen was not actually part of the vented electrolyte vapour but rather part of the by product of something within the vapour burning in a restricted oxygen supplied combustion situation causing an incomplete combustion, needs to be answered one way or the other .... one paper started with "Theoretically" regarding the hydrogen content .... we all know the difference between theory and the real world, in theory they are the same, in reality, they are often not even close .....

Hopefully once I get home and the dust settles, I'll have more time to do research regarding back up testing by the likes of Carnegie Mellon (sorry about my mutilation of the name) NASA etc, they had a lot of research work they conducted 12yrs or more back when I first became involved with LFP cells and their various construction designs and why that design was chosen. Paint me sceptical, but I don't trust any of the specs or testing from Chinese "research" facilities, very few of them have panned out to match my test results and when questioned, a whole different set of parameters were used to obtain the results they advertised ..... Winston was one of the exceptions, as was A123, but they only sold to major manufacturing concerns.

The 18650 for example is two long sheets sprayed on both sides with the active on one sheet and the graphite on the other. Designed for high C rates, not longevity or the storage and release over longer periods at a lower C rating. Prismatic, the type I prefer because of the multiple layers spreading the variations out that are inherent in mass production, a spring loaded vent, insulated case and capable of being compressed to minimise case bulging ..... the pouch cells came later and to be honest, not a fan, too many failures and poor heat dissipation potential. I'm intrigued by the new cylindrical cell sizes as to why? I'm guessing to house the increased capacity and possibly integrating the multiple layers found in prismatic cells with the long contact facing available in a cylindrical cell.

I am interested in your ideas regarding other reasons a cell would heat up, if the cell is cold enough to freeze the electrolyte, the wrong compound electrolyte was chosen and probably the wrong combination of elements in the active material. This is what directed me to the Winston LYP chemistry, all the benefits of LFP with an added minimum threshold of -36*C and upper threshold of 70*C .... I like the sodium ion cells I'm part way through testing because they stretch this a bit further with better under voltage and over voltage tolerances before they cease to function as specified ... actually, I'm still in the process or discovering if they meet the manufacturers specifications or just what specs they really have in regards to their expected performance.

I'm not really interested in supporting or rejecting anyone's claims against how someone else went about their BMS design, either it worked or it didn't. Did it catch fire? No evidence has been forthcoming that there was actual combustion involved. If the electrolyte vapour had been vented outside the vehicle cabin, would this circus have even got the tent up?
I know for certain the NiMh modules catch fire if over voltage charged, the insurance investigator backed that up, the melted modules and blacked steel cases just confirm there was combustion ... the flames licking out of the vent by the back passenger door was enough to convince me it was on fire :lol:

T1 Terry

 

Sorry I didn't reply sooner... I got really sick last week... seems like a combination of covid and allergies, although I tested negative.

Please review the previous posts. One test (the one that yields 50%) was performed in a noble gas, hence there was no chemical reaction due to flame, unless that flame was oxidized by chemicals within the cell (unlikely for LiFePO4)).

I'm always down to read quality sources.

Yeah, A123 initially only sold to those that would allow A123 to review their designs. They didn't want some rando to tarnish their name by designing and selling dangerous products. Even the KillaCycle guy got a slap on the wrist after he crashed his A123-equipped bike into a parked truck. At the time A123 was the only game in town for high charge/discharge lithium, so you basically had to play by their rules. I'm still amazed they managed to go bankrupt.

Related: I still have an A123 cell on my long term tester that hasn't failed yet.

In general there are three broad categories:
-Overcharging (chemical breakdown)
-charging too fast (electrical resistance)
-failure to account for cell aging (NexPower packs lack any method to control this)

Interesting stance.
I would counter (again) that selling an LFP product without a BMS is a recipe for disaster.
I would counter that in at least five cases (that I know about), it "didn't work"... the pack experienced a thermal event.
I would also counter that the customer shouldn't expect their product is going to have a thermal event. Doesn't matter if that is fire, smoke, or just a thermite meltdown... the pack should have the requisite parts to prevent it from disassembling itself in a gooey form under the rear seats. No customer wants that, and well designed products have measures to prevent this from occurring.

I agree this happens, albeit infrequently.
Note that NiMH voltage dips slightly as each cell hits full, which the OEM computer can detect. This allows the OEM computer to determine when a cell is full. It's not perfect, but it's something. Note that this chemical property of NiMH cells is measurable even with multiple cells in series between BMS leads, whereas lithiums gradual (and then faster) cell rise is much harder to detect without per-cell BMS leads).