How to select best UPS and Solar to be off-grid 24 hour for 2kw

50 points by jongi_ct ↗ HN
Hi, I need to be off-grid for 24 hours, I will be powering a device that consume 2kw.

I live in Africa, so Sun is not a problem here. Thanks, jongi

45 comments

[ 3.1 ms ] story [ 85.9 ms ] thread
How often do you need to do this? (E.g. once, once a year, once a month, or once a week?)

And do you have access to / storage for liquid or gas fuels?

I would look at off road, camping, and overland forums for ideas.
That's 48kWh. You're going to be looking at a setup that's a few thousand dollars.
Few tens of thousands of dollars.

You're looking at 4 hours insolation in winter in subsaharan regions. Also assume cloud cover drops panel output to 50% for consecutive days. You're gonna need 25kW of panels and 30kWh+ of battery storage.

2kw draw for 24h is actually quite substantial. Don't forget regardless of how sunny it is in Africa, You will still need to have enough stored power to make it through the night.

I'm not saying it is impossible but suspect it will cost more in fuel to transport the batteries and panels than it would to just run a small inverter genset instead.

If your device doesn't need to be constantly powered then I would suggest bringing pre charged batteries and not bothering with the panels.

A generator will consume roughly a liter of diesel per hour at this load. For a stationary installation, a diesel generator would be a terrible idea.
As with many things - ‘it depends’. You’ll certainly crank up some terrible hour counts running purely off a generator, but for only 2kw continuous you can size a generator that will sip fuel. Depending on local fuel prices, it could take many many years before solar would be even close cost wise.

If only 2kw peak, a common strategy is generator to charge a battery bank, also charge when possible the bank with solar, and run inverters off the batteries (or switchover to the generator when charging the bank). Some people just skip the solar panels. Solar has many problems with unpredictability, and if you need to be sure you’ll have power year round, it isn’t a good idea to rely on it 100% without some other way to charge or supplement it. Dedicated off grid folks always have a generator in my experience, even if it rarely has to get fired up.

Sizing a battery bank to give you enough excess to not worry about a freak 2 weeks of bad weather in winter is extremely expensive and has other issues (like not being able to properly cycle or maintain all the batteries due to lack of peak power)

The claim of 'you'll spend more fuel transporting panels' is a total fantasay

"generator that will sip fuel."

For this discussion to be usefull, you have to provide some numbers. From generator efficiency I was able to find, you need to consume roughly 1 liter of fuel to create 2 KWh of electricity. Playing around with batteries won't change that.

If you look at my calculation above, a single solar panel produces the equivalent of half a ton of diezel. There is no possibility anywhere outside the arctic that a solar installation won't pay for itself.

Using battery banks helps quite a bit, as you can use a larger, more efficient generator to charge the banks quickly and can size it perfectly to do so, then shut it off. So you might only be running the generator for 4-5 hrs a day or less. This helps with wear and tear, lets you be more efficient, etc. How efficient you can get depends a lot on the specific battery chemistry (columb efficiency), bank voltage, max charge rate (C rate, which is very high for lithium, less high for lead/acid), etc. so hard to calculate here.

I personally use a small inverter generator to power ~ 2kw (1.8 specifically) loads for long periods of time off grid. I don't bother with a battery pack in this case, though eventually I will. I also use propane (since I can store/refill a 1000 gallon tank with no worries about it going bad - it will be fine for a decade or more, it's easy to transfer to portable bbq size containers, etc), and propane is $2.4/gal where I'm at in bulk.

It supports that 1.8~kw/h load continuously, and burns ~ a 5 gallon propane tanks worth in 10 hours at that load. Lower loads, it will go as long as 18. This is at altitude (4500 ft), no supercharging or anything, so you'd get ~ 30% more efficiency at sea level. That is roughly half a gallon of propane for 1.8kw/h, or about $1.50/kwh - pretty much the most expensive and inefficient way to generate power using a generator I imagine. Diesel generators are far more efficient for the power they produce, but generally also output much more power and idle burns more fuel, but fuel is less stable - important for me in this case, not if you're running it a lot. If I was using diesel, I'd be charging a battery bank. In general, larger generators are more efficient (often MUCH more efficient) per unit power produced, as long as you can sustain load within their sweet spot. An over provisioned generator will waste a lot of fuel idling, and an overloaded generator will burn fuel inefficiently. This 'small' commercial generator [https://www.ckpower.com/what-you-need-to-know-about-fuel-eff...] (CK300VM-T4 Final) outputs 225kW, and puts out 14.25 kwH per gallon of fuel (rated capacity). If diesel was $3 a gallon like it is here, thats $.21 per kWh - damn cheap!

So, assuming our friend here is running 24/7 at 2kw continuous using my terribly inefficient setup, he'd be going through 72 dollars a day in fuel. Quite a bit! But let's compare to solar!

A 48kwh system is non-trivial in size and expense. A system rated to produce ~ 40kWh/day in winter (wholesale, DIY installation, no racking or batteries) is ~ 36k, and has 54 large panels. [https://unboundsolar.com/1891324/unbound-solar/solar-kits/17...]

Battery option price varies. Lead acid is cheap, but needs to be over provisioned to avoid stupid short cell life (> 50% discharge will severely shorten cell life), and even overprovisioned if you're actually drawing down the cells much every day, you'll be lucky to get 3-5 years out of them. To have a day buffer (essential, even if you get perfect sunny weather 99% of the time) and not kill your cells quickly, you're looking at ~ $30k worth of batteries on top of that, at nnearly wholesale prices [https://unboundsolar.com/1891324/unbound-solar/solar-kits/17...]

Lithium allows faster charging, deeper discharging, and longer cell life, but is expensive. A ~ 45-50kwh usable battery bank is ...

Thank you for the breakdown. I think this is very fair - my assumption from the get-go was that it's a longterm stationary installation.

One thing worth mentioning, in some cases, fuel delivery is a expensive/unreliable, so that can neccesitate eenewables in its own right.

I was looking at installing a system for supplimental power of refrigerators (no batteries), but the project never materialised, so I don't have hands-on experience.

No worries, hope it was helpful. It gets more complex for a real world case, as there are a number of factors that need careful balancing - too big a battery bank and it's hard to do the periodic rebalancing you need to keep the bank healthy, as it requires a significant amount of power, but infrequently. Too small a bank and you'll cycle your generator too often (lots of cold starts cause wear). If you never run your generator for very long, it'll probably be dead in a year or two from corrosion or similar issues - some have decent self exercise routines, but for many it's inadequate. Also, it's almost never a consistent predictable load, especially for households. Grid power benefits immensely from law of large numbers - any individual spiky demand someone does blends in when averaged over 100,000+ homes or a city of a million people, so you can predict and size things better. If you're the only user, your usage can swing from 50mw (standby power for something), or even zero mw - all the way to 11kw in the space of half second (welder). Way more orders of magnitude increase, and way faster than the grid will ever have to deal with. You'd need to decide if you'd want to be able to handle that, and if not, what you'd need to handle - and pay up front for the equipment.

It's almost never a problem to have 'too large' an array from an electrical perspective, but from a practical use of land/space and cost (especially when racking, up to code electrical lines, installation) it's really hard to not cut corners there. Most people end up with arrays that are too small in winter, or don't take seasonal insolation variation, shade, or similar problems into account enough.

Part of what has been helping with solar is dirt cheap financing. Even with nearly free panels, if you factor all the other associated costs in (if you're in the US, permitting can be a nightmare for instance), break even can be a decade or more out even with free money. After that it's not so bad though.

Fuel prices and logistics are like you point out a big factor. In my case propane isn't overly expensive and I can get it delivered. Diesel isn't super cheap, but not bad either. If you're paying $10/gal for diesel delivered, or have to go months or more between deliveries of substandard fuel, it changes the equation a lot.

OP never did reply with details did they?

The answer depends on exactly what you want.

Here is a guy doing a solar shed/office with a similar power envelope as you are: https://www.sevarg.net/2016/05/15/solar-shed-part-1-overview...

Not really. I can run 1500-2000W during the day if I care, but there's no way I can run a 2kW load, 24/7, without massive generator use. My battery bank is way, way too small for that. I typically idle around 75W overnight in the winter, can run a bit higher in the summer, but 2kW would drain my battery bank stone, stone dead about 5h after sunset.
As some other comments noted, there is a lot of information missing to give you a good answer.

Do you need to be off grid 24 hrs out of every 24 hrs? (No grid connection, permanent off grid)

Do you need to be off grid UP TO 24 hours, with a grid connection most of the time?

Is your grid power cheap? If so, how cheap? It might be worth it only as a weather related backup if it is cheap, or might be worth going full off grid if not.

Do you plan on sustaining that 2kw the full 24 hr, or only peaking up to 2kw? If only peaking, for how long?

There are lots of potential trade offs that could help a lot in making it cheaper, or make it much more expensive if you can’t do them.

and maybe they will be off-grid just say, this thursday.
How much advance warning will you get? You need 48 kWh + charge/discharge losses. So if you get 8 hours of sunshine on 100 W per square metre panels you need 48/(8 * 0.1) square metres to charge in one day, that is 60 square metre. Or you can use two months to charge and one square metre of solar panels. More panels charge in less time. Then you need a battery about the size of that in a medium sized electric car if you use lithium ion or lithium iron phosphate. If you use lead acid batteries which are probably easier to obtain in Africa, and cheaper, you will need something like 50 car batteries or equivalent because each on has a capacity of about 1 kWh depending on the size of the battery. Don't actually use car batteries, there are batteries specifically designed for use with UPSes

How portable does this installation need to be? If it needs to be at all mobile it really would be simpler and a lot cheaper to use a 2 kW petrol or diesel generator

So to summarize you need to decide how long you can wait for the battery to charge, how often you want to go offline, how big a solar array and battery you can afford.

The OP needs to hire you.

And maybe you'll accept raw diamonds as a payment?

This is in extremely bad taste. What precisely about Africa makes you think anyone has easy access to Diamonds?
Sorry but am missing your point. Where is Africa mentioned?
The original questioner mentioned that they live in Africa.
Firstly - maybe someone knowledgeable will come along, but but I wouldn't hold my breath - most people here have never touched an off-grid setup. Some might have some form of feed-in tarriff with solar on the roof, but that's totally not the same.

Secondly - there is not enough information here to provide real advice - do you mean that you are normally connected to the grid, but need to survive an interruption of 24 hours? In that case the generator suggestion might make sence. Or do you mean that it is a total off-grid location that needs to be self-sufficient, in which case the generator idea is idiotic.

Just saying 'Africa' is not enough, look for solar irradiance map and and figure out your annual solar incidence. That will inform how you size your solar panels - their power rating are given under standard incidence of 1000W / m2, so you can basically multiply annual irradiance to figure out how much power they will collect in a year.

For example a single 300W panel in the middle of Sudan will give you 720 kwh per year, or ~ 2KW/h per day. Same panel in Gabon will givey ou half as much. Then you have to account for seasons, in UK we get almost 8 times less sun in winter than in summer - you can't compensate for seasons with batteries, so you need more panels.

Batteries compensate for periods of bad weather and nights - they need to be sized according to weather and importance of the load: what happens if there is some unusual bad weather and power runs out, how important is the load? Are we going to be stuck without a TV, or is someone's life on the line? There is also a balancing act - you can compensate for bad weather by getting more panels or more batteries. Example of a battery bank: https://www.photonicuniverse.com/en/catalog/full/359-24kWh-4...

Then there are furhter spesifics -> is space at a premium? Is the load DC or AC? With DC you don;t need an inverter and you avoid losses.

Here is a UK centric calculator, just to sanity check some numbers: https://energysavingtrust.org.uk/tool/solar-energy-calculato...

Generally you will have local firms that specialise in installing solar and they will do a better job of providing advice than NK.

AFAIK most prepackeged UPS devices are a battery and inverter. My hot take is that if you need to run this device continuously it will be a substantial (& custom?) build. The biggest factor is how portable it needs to be.

Overview: Most solar outputs 12 or 24v, so what often happens is solar -> controller -> battery (or bus) -> inverter -> your device.

Batteries: Many batteries are rated for 100ah at 12 volts (so 1.2 kw hours). You would need 2+ of them to run the system purely off them for about an hour. You can buy much beefier batteries designed for residential homes (e.g. 3' x 2' x 1' in size, 10kwh, $5,000+ for lithium, etc..).

Solar: The panels you might see on top of a camper often generate 100 watts each, so you would need 20+ of them to continuously sustain your device. Can be 250 watts or a bit more, so 8+ panels.

I've only worked on electrical for a camper, so could be missing other options! AFAIK your requirements aren't crazy for residential solar setups.

If you mean 24 hours a day forever at 2kw then the solution really needs to include a 2kw petrol generator for when it rains for two weeks and there is no sun. Otherwise supplying the last outlying 20% is going to get insanely expensive.
(comment deleted)
The single biggest expense will be your battery. If I understand you correctly, you want to power a 2kw device for 24 hrs.

thats 48kwhr. which is about 4 tesla power walls.

you then need enough panels to provide > 48 kwhr of power to charge it.

I've had really good results with SMA Sunny Island.

It acts as "the grid". It's connected to batteries, and gives you a 230/120V network. Then, you had a normal solar inverter. It'll recharge the batteries when the inverter gives power, and power your equipment when it does not.

Big plus is that if someday the grid arrives at your home, you can simply take it away and you won't have to touch your solar panels.

Many UPS cannot cold start (battery connected but no grid). 2 kW is substantial, a small ICE generator is going to be the cheapest solution by far.
A single solar panel will produce the equivalent of half-a ton of diesel/year for his location. There is no way in hell a generator is a better option for a permanent installation.

There is a reason all remote meteostations, etc. have long switched to renewable.

I would recommend watching Will Prowse YT channel. For batteries, I'd say BattleBorn. They're expensive but have a 10 year warranty and are just damn good batteries. Probably among the best you can get. They should easily last 10+ years. Will goes over everything. He tests MANY batteries, panels, inverters, etc., and goes over how to size your system based on your needs and location.

https://www.youtube.com/channel/UCoj6RxIAQq8kmJme-5dnN0Q

ClumsyPilot mentioned this, but it's worth repeating - if your load is expecting AC power but is converting it to DC (eg computer power supplies) you can probably save quite a bit by looking at fully-DC options so you skip the DC-AC-DC conversion stages.
This is a job for a generator such as:

https://www.harborfreight.com/3500-watt-super-quiet-inverter...

$850 US, supplies 3000 watts, runs 11 hours at 25% capacity on 2.6 gallons of gasoline. You'd be at 66% capacity, so fuel consumption will be higher. Bring 20 gallons of fuel?

If you go solar, figure you have 6 hours of usable daylight, you'd need seven rooftop sized panels (330 watts each) to supply 2kw for those 6 hours, so you need battery capacity for the remaining 18 hours, or 36 kWh. That's a sizable battery. You'd need 36 of these:

https://www.jackery.com/products/explorer-1000-portable-powe...

(That's obviously impractical so you'd have to custom build a solution.)

I don't know... I'd go with a generator and gasoline. Much cheaper, more compact, and available off the shelf.

Maybe you can get Rivian to lend you a truck for PR purposes. It has four on-board 120v outlets and a 105-180 kWh battery.

I love harbor freight for some types of things, but their quality varies widely: I got a cheap drill press and it was crap, I returned it and the replacement is fine.

So my rule of thumb with them is 1) never buy their products for anything safety related and 2) never rely on their products for something that will be truly critical if it breaks.

Similar experience here. I would only buy hand tools from them, or buying other stuff with the expectation that it's only going to last for one job. Anything electronic or with a motor doesn't seem to last.
when I first responded nobody had suggested a Honda genny

this surprised me considerably since Honda is synonymous with this kind of utility equipment for the best I know all around the world

then I suddenly thought that maybe economic politics is in effect here to avoid sending good hard earned dollars to Japan.

I so strongly agree with favouring American and English manufacturing that I feel compelled to point out that the Japanese economy is very much more circular with America at least than people probably think.

if Harbour Freight is only a name on Chinese made hardware then definitely a Japanese product is better for us in every way

not least because sub standard hardware costs our own economy in cold hard dollars straight away

definitely Honda sell dual fuel generators in the close price range to this about the thousand buck mark and never think you can't haggle even with online prices - the first hit I saw at British 1200 pounds id expect to take home for about 900 for sure.

the caveats ineedausernane minds you to be alert to suggest to me that this is a rebadge unit so all the above definitely does seem to apply

I've cut and pasted my impassioned original reply into my hn profile just because I felt so strongly to write a treatise about the relationship with Japan we have from history and how we shouldn't cut off Japan by reflex trying to protect ourselves because that just won't do us any good whereas quitting the habitual and repetitive purchase of substandard hardware from China is for absolutely certain at the least I reckon the poorest economic habit and one that thru future maintenance and goodness knows what else is still going to be biting us for years to come. Heck wasn't one of the biggest ever DDoSs suffered by Brian Krebs the one only Google could staunch, nothing but countless webcams taken over because 90 percent of all webcams sold to us never could even change their root passwords? I can almost understand the webcams bought cheap (but not the apparent inaction socially after proof emerged of mass blackmail of our young daughters the results of which being unfettered fraudsters at large) but absolutely I can't believe placing trust in emergencies in a unknown origin generator is any kind of a good idea to do.

I’m sorry what are you talking about.... are you still talking generators?
(comment deleted)
Please buy a Honda generator and don't waste time with a Harbor Freight (Hazard Fraught) one. You need this to work in a time of peril - not a use case for HF tools.
Everyone with outdoor access where they live should consider a dual fuel generator that can run some basic things for at least few hours. Dual fuel give you flexibility because gasoline doesn't actually have a great shelf life but a tank of propane will last longer in storage.

I know not everyone can swing the cost for a somewhat discretionary purchase in the $500 to $1000 range, but checking craigslist on a regular basis can turn them up: contractors often use them on job sites.

This thread is an interesting micro allegory for how the move off of fossil fuels will be extremely difficult if all we use are solar and wind power.

Kind of drives the point home about nuclear being the best option. And I know, "ohhh it's so expensive." Ok, so stop letting the markets dictate prices and tax all fossil fuels until gas is $25/gallon. Suddenly nuclear is cheap.

If you have access to a large moving water source, you could maybe do hydro. Is that 2kw continuous, so for 48kwh? If so, that's a pretty large amount. A diesel generator is probably your best bet.
(Solar Shed guy here, I currently have 5kW of panel hung on my office, 10kWh of lead acid, 2kW of inverter, etc)

Unclear as to what you mean "off-grid for 24 hours" - do you need to be off-grid 24/7, or simply run a 2kW device for 24h after the power grid goes down?

Both are quite expensive, the second somewhat less so than the first (though still very expensive). You can probably do the second for under US$10k, probably not the first.

And depending on the use case, the suggestion of "Use a generator..." is very, very much the cheapest option for short term use.

Taking both options here, and going with hardware I would trust to do the job:

If you want to run a 2kW load for 24h after the grid goes down, are OK with charging from the grid, and simply need a really long term, high capacity UPS for infrequent use, your best bet will be something like an Aims Power inverter/charger or similar. I would go with minimum a 24V system for this, and for 2kW, you really should be up at 48V. Yes, I know there are 5kW 12V inverters, and, no, you really don't want to be dorking around with hundreds of amps if you can avoid it.

A 4kW Aims Power inverter/charger will comfortably run a 2kW load, and will charge the battery bank afterwards. At 48V, you'll need a minimum of 1000Ah at a 20 hour rate, and will probably want 1250-1500Ah. If you can at all avoid parallel strings with lead acid, do so. A bunch of battery strings in parallel just guarantees early death, because they won't charge equally over time.

You could use a Trojan SPRE 02 1255 or SIND 06 1225 to build a single string that would run the load, as long as they're warm. This will almost certainly be cheaper than lithium unless you've got a good source of scrap cells and want to build your own. However, this is really hard on the cells (running them down to 80+% depth of discharge), and they won't last more than a few years if they're getting cycled like that daily. If it's infrequent use, have a ball, it'll be fine.

Now, if you want to do it with solar, it becomes a harder question. How long does it really need to run, and do you care about a generator? "Sun is not a problem here" is a common claim, rarely backed by data. You never get clouds or storms? With well aimed panels, in winter, you may only have 6h of usable sun (unless you're using trackers) to generate your 48kWh daily use (plus some overhead for charging), so you're looking at about 10-12kWh, /minimum/ - and realistically, if you want to ride through cloudy days, you'll need far more. Or, you have to have an automatic generator start. Another option would be to have far more battery capacity so you can ride through a few more days, but then you run into the problem of having to have enough solar to recharge it. Practically, for fixed panels, to ride through winter, you're looking at 20-30kW of panel hung.

Something like a string of SIND 02 2450 would get you around 115kWh nameplate of battery, which might be overkill, but you won't drain it too deeply with normal overnight use, and you have capacity for two days of "no solar input" with the pack capacity. The pack will be 3.25 tons, give or take, but still cheaper than a similar capacity in lithium.

I would estimate a system cost in the US$50k-US$75k range in materials if you do it yourself, easily $100k if you're having someone else do it.

The right answer here is to either connect to the power grid, or figure out a way to reduce your loads when there's no sun. Running a 2kW on solar, 24/7, is doable, but it's a very, very expensive project.