Speculate what happened to my cells

Sunder

10 MW
Joined
Sep 6, 2011
Messages
3,054
Location
Sydney, Australia
Alright, so I have no idea what happened to my pack. It's a massive LiFePo4 8S23P battery, charged by a solar charger, and discharged by a 3000W inverter. I have several theories, one of which is a prime candidate, but I don't know how to test for it. They are:

1. The pack was over-charged. The BMS was correctly cutting off the charge at 30v, (24v) battery, so this seems unlikely, but if it wasn't balanced, possible
2. The pack was over-discharged. Again, the BMS does correctly cut power - but again, badly unbalanced, it's a possibility.
3. The BMS was continually bleeding cell row 1, causing it to run flat, while the other 7 rows were fine. This is my favourite.
4. Single possible single cell short, causing the destruction of every other cell in that group?
5. Something else?

Here's a photo of the pack:
Pack-small.jpg

The cells look like this:
View attachment 2

The rest of the pack looks like this:
Nextrow-small.jpg

The facts from testing are:

1. The puffed cells are at 30% capacity, and terrible internal resistance.
2. The non-puffed cells have 100% original capacity. I can't load test such a large pack for internal resistance.
3. The gaffer's tape holding every other string together have pulled off, showing that at one point they too puffed, but deflated back. This doesn't fit with my favourite theory.

I'm fairly certain I'll never find out for sure what happened, but I'm keen to know what you think might have happened.
 
Some events produce gas in a pouch. Overdischarge means the SEI layer gets de-lithiated to mechanical contraction induced fractures, followed by eating the copper metal current collection foil next if it's out of Lithium.

Overcharge breaks the carbonates in the solvent into CO2 and a handful of other gasses in half reactions.

Mechanically damaging the pouch layers through mis-handling may also result in excess gas production.
 
Wonder if I can puncture one and see what gas comes out then ;) What's the gas from overdischarge?
 
Sunder said:
Wonder if I can puncture one and see what gas comes out then ;) What's the gas from overdischarge?

Anything that disturbs the SEI will require re-passivating the surfaces exposed by the cracks. As it undergoes new formation/passivation, it makes CO2 and H2 dominately, and I don't know precisely how many sides reaction gasses.
 
Sunder said:
Wonder if I can puncture one and see what gas comes out then ;) What's the gas from overdischarge?

If the cell suffered from overheating (probably not the case here), Hydrogen Fluoride may have been formed by thermal degradation of the LiPF6 electrolyte. So I would recommend you to NEVER smell battery gas, since HF is lethal at ppm levels.

Though in your case it's probably just CO2 and whatever H2 has not left the pack yet.

You have to diagnose this failure which is probably related to your electronics having overcharged/overdischarged this cell bank. Although cheap LiFePO4 pouch cells are crap by design, it is unlikely that they are the culprit here.
 
I would guess something on your list caused number 4 to eventually happen. Is the pack very old, It might just be dying of age and high resistance. Can't handle that big inverter any more.

FIWW, the positive cell in a lipo pack seems to always puff first, or puff more. No bms involved.

You might have had this happen to the first cell in the 23P positive end, then domino on down the pack as each puffy cell caused a higher c rate on the remaining ones.

When my ping battery died of old age, it was definitely puffed on the + end, while on the negative end of the 48v 3p pack, the cells were still hard and relatively good. They still had high resistance, but had not yet puffed.
 
It was indeed the positive cell. I wonder why that is. Is there an explanation for that? I've seen chargers in series where the most positive charger seems to do all the voltage saging. Similar problem?

As for overheating, what are we talking about? This shed can get to 60*C in summer, but as LiFePo4, I thought it wasn't all that sensitive to high temps.

Thanks for the replies guys. I've since rebuilt this pack and it's working beautifully, but I do want to prevent a repeat.

My last thought, and given the state of the shed, is that something metallic fell across the terminals of cell 1, and causes a long term short. Seems the most likely now, so I've put it in a fully enclosed box.
 
Wow you have 644 cells and just four puffed ? I usually see a parallel group as a cell but 23p I would qualify them first. So four bad cells all at the 8s or positive end. ?
Edit: 184 cells
 
I'm a little lost. 8s 23p = 184 cells, and if you look carefully, one entire 23p row puffed. It's just 4 in the foreground, but you can see more puffed in the back ground.

All cells in the first row puffed, none in the other rows puffed.
 
2 years, but only in service for 1.
 
It seems to me that early positive end cell puff is often related to discharge rate. But yours was only a bit over 1c. Eventually, you would still expect it, but after 4-5 years.

One internal shorted cell though, dendrites or a contaminated cell, would then puff the whole row. Overdischarging it.

Cloudy week, trickle short in one cell, the whole row went too close to 0v. Could be random that it was the positive end.

BMS not doing the discharging on that cell group, right? Could also be that. See if its stable off the charger but on the bms. Or does it keep trickling down that group slowly.

You mentioned heat, that is not ideal, full and hot. Was it also stored full and hot before it went into service? Still, shoulda took more like 3-4 years to die from that. I'm still betting on one cell shorted internally early, and then trickled the other cells to death.
 
Back
Top