Off grid fridge project in west africa

babish

New member
Joined
Feb 10, 2018
Messages
26
As there are frequent power outages where I live in west africa, I would like to try setting my fridge off grid.
(I am also very keen to learn about solar and batteries)

The first guidance I am looking for is about the type of cells to use in the battery pack: ambiance temperature is fairly high there, around 30 degreesC (from 26 to 35), so are LiFePO4 cells much better than regular 18650 Li-ion (egPanasonic NCR18650PF)in this case?

I am not too price sensitive, I am looking for the safest, best long term solution.

One other issue is that I cannot ship anything there (no service and customs issues), so I plan toship everything to the uk where I regularly go, and bring it back by plane, where there is usually a 160wh limit per battery size.So I can only go for the smaller LiFePO4 (40Ah).

Thanks for the help!
 
If you dont have a price limit and safety in mind then its LiFepo 10x before you go LiIon.
LiFePoe4 have alot higher cycle count and you cant kind of not get them to start a fire :)

LiFePo4 bricks would be nice for your task and those also are better suited for 12V system if thats your plan. LiIon is either 24V or 48V+. Not 12v...
 
Thanks.

So I am now looking at buying Winston 3.2V/40Ah cells.

I was initially thinking of going with a 24v battery, as I understand it is more efficient and will be easier to scale up if I later need more power. Are there any advantages of 12v apart from the fact that you can tune more easily the number of cells and solar panels?

Given that locally I have only found 24v/250W panels, I am now hesitating between 3 panels at 24V and a 12V battery, or 4 panels at 48V and a 24V battery.
My fridges uses about 1.4kwh/day, and from the solar map I am supposed to get between 3.16 and 4.26 kwh/day per kw depending on the season, so 3 panels sounds enough, which would plead in favor of a 12v battery, with 4s6p for a total of 3kwh.

Any advice for a MPPT charge controller? The outback flexmax sounds quite popular, does the efficiency and features justify the relative high price?

And finally any suggestion for a bms?

proabably too many questions!
 
It depends on price again. In some places 12V systems are cheaper but if you need the output go 24 or even 48V.

Outback is nice stuff. If its worth the extra money im not sure. The efficiency is not that much better but the guarantee is in many times
 
I finally found time to make some progress on my system.
It would be much appreciated if someone can take a quick look in case I have missed anything, in particular safety wise.
I will use:
- 3 275W pv panels (rated 31V and 8.9A)
- MPPT PCM60X charge controller
- 16A && 32A DC breaker (https://www.ebay.co.uk/itm/382128784099)
- 16s 40ah LiFePO4 battery, with batrium watchmon4 bms
- 80A fuse on the battery (https://au.mouser.com/ProductDetail...=sGAEpiMZZMsh2y49K8ANrQADpSXrq5qSqSHwvyEKiyg=)
- 2kw inverter (https://www.amazon.co.uk/gp/product/B00DFNE820/ref=oh_aui_detailpage_o06_s00?ie=UTF8&psc=1)
- 4mm2 cable from pv to charge controller
- 16mm2 cable between battery/charge controller/inverter
- 10mm2 cable for grounding pv panels and charge controler

see attached schema


image_idmhch.jpg


Is a shunt trip advised for my system? (I have not ordered the batrium expansionBoard...)
I am not sure I fully understand in what situations it would help vs the fuses and breakers.

Thanks!
 
I have used the breakers from here as they combine postage so a few breakers turn out quite cheap, they do work well :
https://www.aliexpress.com/item/2P-32A-DC-440V-Circuit-breaker-MCB-for-PV/623939138.html

The polarity and current direction of the DC breakers is important so the feed / surege source is wired into the top, i.e. battery pack in the top, inverter in the bottom.

Suggestion / Option
With the 275W panels, buy an extra panel and use a PWM controller, change to 15s and charge to 4V max.

Overall you should get more energy at a lower cost with the extra panel and a cheap PWM controller and the PWM controller should be more reliable over time. Less costly to replace as well..

The Vmp from 2 panels in series is 62V (drop 1V for inline diodes) so 61V. Drop another V for cable losses and you end up at 60V. 15s at 4V per cell tops out at 60V...

Sometimes MPPT route can be more costly with smalleer systems than a good panel voltage and a cheap PWM controller.
 
Your wiring diagram looks good except near PV1, PV2, PV3 you only need 1x fuse there.
In the blue battery box, take care that the body of cellmons do not touch anything else - the black paint is not full insulation & they could short out, eg to another cellmon, post or copper cell link, etc.

Try to have some tilt with your solar panels so dust gets washed off - otherwise dust will build up. 5 degrees is enough.

I hope the round table is strong enough - batteries are very heavy!
 
My batteries are dead!!! :mad: :huh:
I had to leave for 6 weeks so I disconnected the battery (I switched off the DC breakers between the panels and the charger, and between the charger and the battery, and I switchedoff the dc cut off between the battery and the inverter).

I only left the batrium bms on. The battery was about 3/4 full I think, soaround 1.5kwh.

When I came back the cells had ballooned and the total voltage was8V, with cell voltage from 0.07 to 1.6V, so I guess there are all dead.
I am not sure what could have happened?
Is it possible it got depleted by the bms? what is its consumption?
Is it possible there was a short circuit or a lightning strike?

I plugged the usb to the watchmon in the hope to get the historical data which would give more info, but it does not connect. Any idea if the historical data logs are still available and how to access them?
 
Unless you have implemented some form of LVD, leaving the BMS connected will have depleted your battery.

6 weeks is a long time, the consumption of the BMS isn't negligible.
 
yes 1W consumption is about 1kwh in 6 weeks, I guess I should have thought about that... though quite ironic that the bms killed the battery...
 
babish said:
though quite ironic that the bms killed the battery...

If there's no LVD that's just expected behaviour - the WM4 has a wide operating voltage so will continue to function down to around 8 volts.

I would not have expected over discharge to have caused the physical damage you have described, however over charge would.
 
could anyone tell me the exact consumption of the WM4?
Also is it wise to try to apply some charge to the battery to see what happens?
 
babish said:
could anyone tell me the exact consumption of the WM4?
Also is it wise to try to apply some charge to the battery to see what happens?

It rather depends on how its configured and what its connected to or driving - but likely less than 20mA with no outputs being driven, so minimal but not negligible.

Dont try charging visibly damaged cells.

Are you sure they weren't left connected to a charger (that wasnt configured correctly ?)
 
Dala said:
Not if they have balooned :(

Pics of the damage?


image_orqetp.jpg

image_cqobvc.jpg

not that easy to see the slight curvature on the pictures, but they have expanded


Sean said:
babish said:
could anyone tell me the exact consumption of the WM4?
Also is it wise to try to apply some charge to the battery to see what happens?

It rather depends on how its configured and what its connected to or driving - but likely less than 20mA with no outputs being driven, so minimal but not negligible.

Dont try charging visibly damaged cells.

Are you sure they weren't left connected to a charger (that wasnt configured correctly ?)

no outputs driven and yes I am sure there was no other loads connected;
my mistake was also to not have rewired the bms power connection after the shunt, then I would have seen itsconsumption when everything else was disconnected.
here is the voltage per cell:
1 0.4
2 0.28
3 0.43
4 0.46
5 0.4
6 0.6
7 0.64
8 0.09
9 0.09
10 1
11 1.1
12 1.03
13 1.14
14 0.7
15 1.31
16 1.63

is there any point tryingto remove the lowest voltage cells (also the most visually damaged) and charge them under low voltage and current to see what happen?
or there is no hope anyway as those cells are not supposed to go below 2.5V? (https://files.ev-power.eu/inc/_doc/attach/StoItem/1123/GWL-Winston-LFP040AHA.pdf)
 
Just charge them and try it. Most likely they're still fine with some degradation of power in some cells more than others. That means it won't keep a good balance if you try to charge it to the max. So reduce your usable voltage range and you'll be ok.
 
You could try recharging them gently but have a read of the "over discharge" part here:
http://nordkyndesign.com/lithium-battery-banks-fundamentals/#high-SOC-degradation
(the article seems conservatively/strong on the safety side written)
You don't have any reversed voltage cells but some issues from dissolved copper might come out later.
Maybe try charging them gently, maybe 1A at first until you get to about 3V per cell.
Probably should plan towards replacement later.
 
at what voltage should I charge them? should I try a lower voltage first, eg 12 or 24V?
 
Voltage across the packs probably won't make that big of a deal. So you could charge them either 12V or 24V setup. Or even individually in parallel.

The key here is to charge slowly, with low amps per cell. Redpacket says 1A, I might even would go a little lower to 500mA if possible (per cell). If the charger will output 5A, then have 10 in parallel. The slower the charge current, the more gentle the recovery and less likely any damage will occur.
 
Back
Top