Understanding, planning for, and detecting failure modes

[braincore]

Having seen/read about li-ion packs combusting, I'm interested in hearing what can go wrong.

First off, the pack to be built is 20S-7P (72V, 35Ah) of 21700 5mAh cells (supposedly panasonic 2170A). The pack will use a BMS (JK BMS) to balance, protect over-charge/discharge, limit current, etc. It doesn't look like these cells have a PTC built-in, but being new to all of this, I could be mistaken.

I've read about lithium dendrites forming which can cause an internal short. The worst case will be an internal short causing a 7P pack to short and triggering thermal runaway. I assume this is extremely rare for high-quality batteries that aren't overcharged or fast-charged, which will be the case with the BMS configuration. Nevertheless, assuming I have a microcontroller fully recording the BMS's measurements of each 7P pack, what are the signs that dendrites are forming to dangerous levels? Can it be fully inferred by a drop in capacity?

Are there any other catastrophic failure modes I need to plan for? Are certain types of physical punctures more catastrophic than others?

 

[E-HP]

Having seen/read about lipo packs combusting, I'm interested in hearing what can go wrong.

First off, the pack to be built is 20S-7P (72V, 35Ah) of 21700 5mAh cells (supposedly panasonic 2170A).

Those aren’t lipo.

 

[braincore]

Those aren’t lipo.

Mistype. Thanks for the correction. I'm inquiring exclusively about li-ion in this post.

 

[gromike]

Probably your best chance of a catastrophic battery failure would be in the building of the battery, and in the subsequent installation. Playing with these things, i.e. high power battery banks, sure can get serious -- quickly.

 

[braincore]

Just to be safe from internal shorting, I think I might add a "fuse wire" between each positive cell terminal to the parallel cell group bus bar. Anyone have experience sizing it or even material selection (tinned copper?)?

 

[DaLanMan]

Yes, however that was your one...

 

[braincore]

Just to be safe from internal shorting, I think I might add a "fuse wire" between each positive cell terminal to the parallel cell group bus bar. Anyone have experience sizing it or even material selection (tinned copper?)?

Answering my own question, looks like there's a goldmine of information here: Fuse-wire for individual cells, DIY Tesla style

And a chart on fuse wire sizing to get started: Fusing Currents versus American Wire Gauge for copper, aluminum, iron, and tin (Cu, Al, Fe, Sn), wire melting currents

 

[braincore]

Hello LLM! Let me know when I can buy an AI robot to build me a battery pack.

 

[glennb]

Having seen/read about li-ion packs combusting, I'm interested in hearing what can go wrong.

First off, the pack to be built is 20S-7P (72V, 35Ah) of 21700 5mAh cells (supposedly panasonic 2170A). The pack will use a BMS (JK BMS) to balance, protect over-charge/discharge, limit current, etc. It doesn't look like these cells have a PTC built-in, but being new to all of this, I could be mistaken.

You’re not mistaken. You could add them yourself though, they’re inexpensive.

I've read about lithium dendrites forming which can cause an internal short. The worst case will be an internal short causing a 7P pack to short and triggering thermal runaway. I assume this is extremely rare for high-quality batteries that aren't overcharged or fast-charged, which will be the case with the BMS configuration. Nevertheless, assuming I have a microcontroller fully recording the BMS's measurements of each 7P pack, what are the signs that dendrites are forming to dangerous levels? Can it be fully inferred by a drop in capacity?

Your assumption about rarity is safe.

I’ve never given any thought to dendrites.

Are there any other catastrophic failure modes I need to plan for?

Outgassing doen’t meet my definition of catastrophic, and is better regarded as a success than a failure, so that leaves only one mode: Fireball.

I suspect it is causes (other than dendrites) that you’re interested in though. Exceeded thermal limits are the only one I know of.

Dozens of things could conceivably cause cells to reach critical thermal limits. Yes, you can plan for all of them. You could make a thorough list of them, group them into subheadings, rank them in descending order of probability, and brainstorm solutions.

I’m sure you’ve got better things to do though.

Most people seem satisfied with:
- confining cell usage to within manufacturer recommendations
- using best practice for construction
- incorporating vibration dampening; and
- housing the battery in a sturdy container

Are certain types of physical punctures more catastrophic than others?

Catastrophic is almost an absolute, but punctures are obviously by degrees, from pinprick though to incendiary bullet. I doubt you’ll find much in the way of controlled studies, but I’m sure there’d be plenty of pyromaniacs puncturing cells for the purpose of generating YouTube content.

The only safety conformity type data sheet I bothered reading was for some Lishen cells, and from memory the puncture test was conducted with a “nail”. I don’t recall the type or size being specified, or the result.

 

[fechter]

Another failure mode is a short in the wiring, mostly the balance wires going to the BMS. There is a lot of vibration on a bike and any wires that can rub against something can wear through the insulation. Similar thing can happen between cell groups if things move around. Avoid by carefully routing wires and gluing things in place.