Battery selection for large off grid system

[karrak]

If you were to use an LiFePO4 (LFP) battery the minimum size battery for your 15kW of solar panels with 3 Midnites and your 12kW of inverter would be around 500Ah (27kWh). Cost would be somewhere around $10,400 plus $500 for a BMS.
If you were only going to be running lighting, fridge and freezer, computers etc at night I can't see that you would need much more than that.

The two LFP batteries I have installed are over four years old and still as good as new. There are a large number of other people on other forums with batteries around the same age that say the same thing. steveg has an LFP battery that is over 7 years old and reports the same thing. It is worth looking at his posts. I can't see any reason that an LFP battery that is properly looked after won't last well in excess of ten years.

One plus with LFP batteries is that unlike Lead Acid batteries it is OK to add capacity to them at a later date.

There is no regular maintenance. Only issue is that they have to balanced correctly as part of the commissioning process, correctly set up and have some sort of Battery Management System to make sure they are not damaged by being taken outside their safe operating zone.

Simon

Off grid 24V system, 6x190W Solar Panels, 32x90ah Winston LiFeYPO4 batteries installed April 2013
BMS - Homemade Battery logger github.com/simat/BatteryMonitor
Latronics 4kW Inverter, homemade MPPT controller

[AzRoute66]

ButchDeal,

It is actually one of the very first articles that I found when starting to look at solar. The one that got me thinking, "Hmm. This might be more interesting than it sounds." and started a good two month pondering on the merits and mechanics of homemade solar trackers. He provides his rationale, and even some of the formula used in his simulation. His 'summer' is mid April until mid October. His website is solarpaneltailt.com Great read for the geek in you. (I am in Sierra Vista, the same latitude as Nogales, and with much higher elevation (4,600')).

[AzRoute66]

[...] Once of twice a month you will spend a day cleaning and watering your batterries. Once a week measuring specific gravity every 15 minutes on every cell while you charge them up on your very large thirsty generator.

I think that is a bit extreme, in the real world not sure anyone would, or should, do watering more than one a month, and probably not checking SG at all unless a problem is suspected. Forgive me for not thinking this through first, but couldn't a clever enough guy use Zones 5 and 6 to equalize Zone 3?

[ButchDeal]

ButchDeal,

It is actually one of the very first articles that I found when starting to look at solar. The one that got me thinking, "Hmm. This might be more interesting than it sounds." and started a good two month pondering on the merits and mechanics of homemade solar trackers. He provides his rationale, and even some of the formula used in his simulation. His 'summer' is mid April until mid October. His website is www.solarpaneltilt.com Great read for the geek in you. (I am in Sierra Vista, the same latitude as Nogales, and with much higher elevation (4,600')).

Sorry not impressed with that. I feel that as long as PV modules are as cheap as they are there is little need for tilt systems.
Further, clearly he is mistaken on any assumption of north being optimal for anything other than a very short time frame in June the continental US. Maybe you two have a different definition of optimal than most though.

Unless there are some shadows in the mix that could easily changes things.

if anything the south is even better in Sierra Vista:

Month	Sierra Vista Energy Production [kWh] North	Sierra Vista Energy Production [kWh] South
Jan	26.08	31.45
Feb	27.17	30.77
Mar	39.9	42.68
Apr	51.45	53.09
May	55.74	56.1
Jun	56.15	55.86
Jul	44.93	45.07
Aug	43.94	44.95
Sep	38.08	40.3
Oct	34.11	37.9
Nov	25.18	29.81
Dec	23.22	28.52
Total	465.95	496.5
Summer	183.1	186.18
April-Oct	324.4	333.27

[AzRoute66]

[...] Maybe you two have a different definition of optimal than most though. [...]

Jeez, Don't shoot the messenger. It's not my simulation. If you read it he agrees with you to add a panel and leave it alone - he also provides two different 'fixed array' philosophies and the groundwork for a novice to create their own version (to deal with shade, etc.). If you read his article thoroughly, and disagree with the result that is fine. I'll continue to report his news as 'my infotainment gospel' until you or somebody else shows me the error of his ways (or the 'more better' of yours).

[ButchDeal]

Jeez, Don't shoot the messenger.

You made the claim then amended that it came from someone else. In either case there is no backing for the claim.

I'll continue to report his news as 'my infotainment gospel' until you or somebody else shows me the error of his ways (or the 'more better' of yours).

You claimed that the optimal tilt angle for summer production in southern AZ was 3 degrees north.
I have twice demonstrated that that is false unless "summer" is solely defined as July only. However you then stated that Summer meant from mid April to Mid October (about half the year) which is also is not the optimal tilt angel for those months.
The only way it could be true is if you have a very specific and local shadow to the south, making the north tilt option better.

Did you pay that site to generate you an optimal tilt angle?

References: I am listed on a patent for solar production modeling

[bcroe]

ButchDeal,

It is actually one of the very first articles that I found when starting to look at solar. The one that got me thinking, "Hmm. This might be more interesting than it sounds." and started a good two month pondering on the merits and mechanics of homemade solar trackers. He provides his rationale, and even some of the formula used in his simulation. His 'summer' is mid April until mid October. His website is www.solarpaneltilt.com Great read for the geek in you. (I am in Sierra Vista, the same latitude as Nogales, and with much higher elevation (4,600')).

With cheap panels and PV systems being installed where clouds and snow are common, the problem is a
lot more complex. Bruce Roe

[AzRoute66]

[...] Did you pay that site to generate you an optimal tilt angle?

References: I am listed on a patent for solar production modeling

I won't get sucked into hijacking this very cool thread, for which 'we' have only managed to do the obvious Ah calculation, select FLA batteries, and point out that the charge controllers were insufficient. 'We' have not suggested a better charging situation, inverter recommendation, or generator discussion yet. 'We' can do better.

I'll fan your flame by mentioning that has many patents, he covered 95-ish percent of the world's populated area with three simple algebraic equations and offers to do a site specific recommendation (which I am sure would agree closely with yours), he gives an overview of his technique to include his air mass consideration formula used, has an MSEE from MIT, and provided a page on 'how to repair a toilet valve' in his bio page. Now, go find a panel tilt thread and revive it (or start a new one) and I will gladly defend this man's honor at high noon as if it was my dear mothers, assuming it stays fun. Until then, can't SOMEONE help Mr. Silver with his 15 KW problem?

[ButchDeal]

that has many patents, he covered 95-ish percent of the world's populated area with three simple algebraic equations and offers to do a site specific recommendation (which I am sure would agree closely with yours), he gives an overview of his technique to include his air mass consideration formula used, has an MSEE from MIT, and provided a page on 'how to repair a toilet valve' in his bio page.

Can he walk and chew bubble gum at the same time too. Has nothing to do with your claim. Can't even find where he makes a similar claim either, his formula comes up with 4.7 degrees tilt and south.

You even just stated that you feel his assessment will agree with mine.

You are making the claim but tossing out as evidence that someone else (who doesn't make the same claim as you) is an MSEE and can fix a toilet??
just face it north tilt in AZ or any of the continental US is not optimal unless you have very specific shadow issues.

There are so many silly claims about optimal direction etc that novices are confused by the sheer volume of FUD.

As for the OP, no one can help him till he evaluates his needs.
Off grid system design is simple, you start with load, then get a battery and inverter that can handle the load, then an array that can charge the load.
He started at the other end.

[AzRoute66]

Butch, you have me giving up on this thread. You won't move it elsewhere, fine, I quit.

Mr. Silver, I suggest you contact Morningstar, and do whatever they suggest. They have a global presence and I have no doubt their solution will be of the highest quality. They might bring in Sunwize, a distributor of theirs that specializes in light industrial/commercial applications to help out with all of the periphery of site survey, permits, etc.

[Sunking]

I think that is a bit extreme, in the real world not sure anyone would, or should, do watering more than one a month, and probably not checking SG at all unless a problem is suspected. Forgive me for not thinking this through first, but couldn't a clever enough guy use Zones 5 and 6 to equalize Zone 3?

Sure fire way to destroy a battery. You are correct, not many would do that. Thus why they should use commercial power. They get a wake up call and education when they have to replace expensive batteries in a year or two.

[Sunking]

If you were to use an LiFePO4 (LFP) battery the minimum size battery for your 15kW of solar panels with 3 Midnites and your 12kW of inverter would be around 500Ah (27kWh). Cost would be somewhere around $10,400 plus $500 for a BMS.
If you were only going to be running lighting, fridge and freezer, computers etc at night I can't see that you would need much more than that.

LOL. Do not listen to this fool. He just told you how to start a fire. 300 plus amps on a 500 AH Chi-Com POS LFP battery. That should get you banned real quick again. Lets see what the Mods do? Gotcha again. C'mon mods you know that is a recipe for a FIRE

[sensij]

Butch, you have me giving up on this thread. You won't move it elsewhere, fine, I quit.

Mr. Silver, I suggest you contact Morningstar, and do whatever they suggest. They have a global presence and I have no doubt their solution will be of the highest quality. They might bring in Sunwize, a distributor of theirs that specializes in light industrial/commercial applications to help out with all of the periphery of site survey, permits, etc.

Morningstar will probably suggest that the OP attempt to characterize the loads before they make any recommendations. Even if you want to approach this from the supply side, at some point you still need to set a supply budget, and figure out what can be reasonably powered from that budget.

The idea of running daytime loads only when there is more generation power available than what the battery needs to stay healthy is great in concept, but more difficult to implement. Some simple timers or relays actuated by voltage sensors will allow some excess power to be skimmed off, but to try and push a significant portion of the load onto the "run only when conditions permit" subpanel will be challenging to control. Not impossible, but requiring a level of automation or user engagement that are not present in most of the discussions here.

My real goal is to find the maximum amount of amp hours that the battery bank should be sized at based on the existing panels and equipment. I will adjust the loads accordingly.

If we take this statement as-is, what do we get.

karrak says the minimum battery size with LFP chemistry would be 500 Ah (27 kWh).

Wy_White_Wolf says the battery size range for FLA would be 2520-3780 AH based on panel rating, although as SunEagle points out, the controllers won't deliver that much on a 48 V battery. With output per controller at 76-80 A depending on the PV string voltage, that probably drops the battery size range down to 1824-2400 Ah, or 87500-11500 Wh, which gets you roughly the same usable energy daily as the LFP bank karrak suggested without worrying too much about timing individual loads. (as an aside, having more peak panel power than what the controllers can convert to deal with real world variation isn't a bad thing.)

Is this enough for the OP to feel like his primary question has been answered? Is it necessary to dive into the derivation of any of those recommendations?

[Further aside, solarpaneltilt.com very clearly states their formulas maximize insolation on the array, not energy generation. What they show is probably correct, but I'm not sure why anyone would want to maximize insolation for its own sake. Maximizing energy production as ButchDeal has done takes more than 3 simple algebraic equations, and is covered nicely in the technical documentation for PVWatts, SAM, or Sandia's PV site, or a number of other first sources, with Duffie and Beckman's work a good reference to check out at some point]

[Sunking]

Sensij panel wattage and battery type dictate what the minimum and maximum size battery can be supported. For FLA the range is C/8 to C/12. Yes I know there are some FLA that can go to C/6. AGM can go as high as C/2, but AGM cost twice as much and last half as long. So FLA wins in both longevity and birth to cradle cost.

So with that fact if you have 15,120 wat panel system @ 48 volt battery, Max Charge current is going to be at least 15,000 / 50 volts = 300 Amps. Using C/8 max and C/12 minimum for FLA the smallest battery is 8 hours x 300 amps = 2400 AH, and the largest is 12 hours x 300 amps = 3600 AH. Based on 5 day minimum anatomy yields 23 Kwh/daily with a 2400 AH battery up to 34 Kwh daily for a 3600 AH battery. I have lived in large luxury homes with heavy Air Conditioner loads in NV, OK, and TX. Never have I ever used that much power in a day.

As for Karrak is smoking crack again with a 500 AH Chi-Com POS LFP battery 300 amps exceeds C/2 charging rate which I hope the Mods Pick Up and ban that fool again for making dangerous suggestions. Make it permanent this time Mike and SE. Cheese and Rice trying to take 23 to 34 Kwh/day from a 27 Kwh battery is as stupid as stupid gets. For a lithium battery, they are sized for 4 day antinomy which yields the same usable 3 days of a Pb battery sized for 5 days. Clouds and rain re going to happen. That is why you need 5 days of autonomy and a generator. Based on just 20 Kwh/day would require a 1600 AH LFP battery at 48 volts. That sucker would cost you 3 times as much as a PB battery and not last as long. Stupid is what stupid does.

As has been stated many times, the very first step before you do anything is determine your worse case daily usage. The OP has not done that and failing to do so means a 90% chance of failure. Three things happen if you guess, and two are really bad.

[sensij]

Sensij panel wattage and battery type dictate what the minimum and maximum size battery can be supported. For FLA the range is C/8 to C/12.

...

As for Karrak is smoking crack again with a 500 AH Chi-Com POS battery 300 amps exceeds maximum C/2 charging rate which I hope the Mods Pick Up and ban that fool again for making dangerous suggestions. Make it permanent this time Mike and SE.

The OP has 3 Midnight Solar Classic 150's. Those will put out a maximum of around 80 A on a 48 V battery. It doesn't matter how much PV is installed, the charge controllers will never produce more than 240 A. I think that leaves karrak 's recommendation intact.

Whether 15 kW of PV on those charge controllers is optimum is sort of a secondary concern, since they are already bought. Having the extra PV power creates some design flexibility in array orientation, allowing East-West (or SE-SW) layouts to be considered for increasing the length of the charging day, or making more of cloudy days.

Yeah, my statement on LFP autonomy was garbage. I was thinking about it in terms of 100% discharge for load shifting under TOU plans, which is fine because there is no consequence for not fully replacing it the next day. Off-grid, worst-case production potential needs to be looked at more closely.