NEVER Build a sealed building to house batteries

[Korishan]

View: https://www.youtube.com/watch?v=6LbBryib8yY


The owner built everything to higher standards and tried to do everything right. Heavy duty plywood, double insulation, climate condtioned, BMS's, built away from the main home.
But yet, because he built it too good and sealed, he created a potential explosive force, which ended up happening. The details as to why the cell(s) failed is unknown.

The biggest take away here is, Do NOT build a SEALED container for your battery storages. Always provide a ventilation path to the outside atmosphere. It's better to waste some extra energy trying to maintain temps, cooling/heating, than it is to potentially have to replace the whole structure and anything else damaged in the process.

Be careful folks!

 

[Fukuokian]

Korishan, thank you for posting this! I definitely was heading towards a similarly over-sealed setup! I'm currently looking at adding a H2 sensor that will trigger my ventillation fans.​

 

[Korishan]

I wouldn't rely on a sensor that could possibly fail. Cold air sinks and H2 rises. Have top vents on both sides of the room and an exhaust fan always blowing out. Doesn't have to be much flow. Being at the top allows heat to escape as well. Any heating needed to could be localized to the cells directly.
This would allow for good ventilation and almost zero risk of pressurization

 

[OffGridInTheCity]

Very interesting.

Air flow - The powerwall under my house which is 'open to outside air' via typical crawl space vents around the perimeter. The heat/cool of the house above keeps this area within 55F/13C in winter and 75F/24C summer... which I believe is reasonably 'gentle' on the 18650 cells temperature wise.

BMS can fail - Absolutely! I'm in my office most every day - so I rely on the Batrium Toolkit Blue-Bar app and a custom web page (with bold, red alerts) on a 3rd monitor 24/7 for peace/confidence as to what's happening. Would not be as comfortable relying on BMS with no monitoring of the BMS itself.

I have a stand-alone, purchased, LifePo4 48v battery powering my APC 3000 - and I occasionally check the attached monitor, but it makes me more nervous that then powerall because I don't see it daily.

My camp trailer battery (27kwh) I monitor on my desktop as well.... it get's charged up to 4.15v/cell for trips, so I really depend the Batrium Toolkit to track what's happening as that's only 50mv margin to overcharge.

Operating voltage range - Powerwall charging to 4.0v max leaves a 200mv buffer on the upper end for some messiness / to avoid overcharge. Discharge to 3.5v leaves plenty of room on the bottom end for messiness.
**I wonder if things overcharged in the youtube?

Operating current range - 50ma/cell (overall average) to a max of 250ma/cell - pretty low stress.
**I wonder the 'stress level' of the powerwall in the youtube?

Linked Fire Alarm - right over the powerwall.

Cell level fusing - each cell has 28awg wire to buss... hopefully, a big short will tend to tamp itself down.
Note: LifePo4 cells of larger size don't lend themselves as easily to cell level fusing - and perhaps there's a larger potential here than with smaller a 18650 cells? I'm not sure - no one does cell level fusing that I've read, only battery level fusing for LifePo4.

Conclusions for my situation:
With these operating circumstances, and 8 years under my belt, I'm not thinking the powerall will fail catastrophically within 'hours' for no reason - I'm thinking weeks and a trend (sagging pack or voltage irregularities) will be manifest in plenty of time (weeks) for an intervention.

However, fire could erupt per cell? or electronic equipment! But that's true of the oven in the kitchen as well. I did go ETL equipment.

Nothing is 100%.

 

[Fukuokian]

I wouldn't rely on a sensor that could possibly fail.

Yep, I agree! My original plan was to have the vents sealable for the winter to keep the heat. (I've quickly rethunk that idea ).

I'm pretty sure I'll have enough natural convection from my vent placements (two 15cm openings high & low at opposite corners of the box) during the winter, but I still like the idea of venting out very quickly in the case of a thermal runaway. In reality, this is not very likely to happen; but, it still is a possibility. There's a way to use an arduino and cheap sensors to trigger the fans. The sensors have a typical 3-5 year lifespan, so I was already thinking to have several in operation at once, and replace them at regular intervals.

 

[Korishan]

What I would recommend for controller the fans is to use solid state controls, not software. However, software can override and/or monitor what's going on.
Transistors, thermistors, and diodes works really well at controlling switching without the need of programming. Then use the arduino/mcu to monitor what's going on, or use another transistor switching to bypass the solid state devices to force the fan early