Pylontech US3000 Battery Economics

This is my first post on this forum! Something I would like to get some input on, is whether there is an economic benefit of owning a large battery bank. That obviously depends on the cost of batteries at any given point in time, the cost of grid electricity and the specifics of the battery to buy. The reason for my post is because I currently have a Victron Multiplus 5kVA with 2xUS3000 batteries. I am contemplating whether to get a third, which seems to be more “in spec” than only two.

Some data I managed to gather:

1. The Pylontech US3000 warranty can be summarised as follows:
1.1 4500 Cycles @ 90% DoD (My understanding is that 6000 @ 80% is also possible, and because 4500 @ 0.9 < 6000 @ 0.8, I’ll use 6000 @ 80% in my calculations).
1.2 60% nominal capacity remaining after 10 years (or 70% after 7 years).
1.3 I’ll assume that capacity degradation is linear over 10 years.
2. I’ll assume a 90% DC to AC conversion loss, assuming that you use the battery with an inverter.
3. Current cost of electricity in Bellville, Cape Town is R2.12 for the first 600kWh and R2.92 thereafter.
4. It is impossible to estimate future electricity increases and future investment return on other investments (if you didn’t buy a battery) so I will assume that the electricity increases and investment return are the same (negating the need for an unreliable time value of money adjustment). This would hopefully also lead to a somewhat prudent calculation.

Over a US3000’s life cycle - ignoring the 10 year lifetime, you can therefore expect to cycle the following amount of electricity from it:
6000 (cycles) * 3.5 (nominal capacity) * 80% (DoD) * 80% (average nominal capacity) * 90% (to get the power on the AC side) = 12,096 kWh

However, more likely that you get 365.25 * 10 (days in 10 years) * 0.8 (1x80%DoD cycle per day) = 2,922 cycles over the 10 years, assuming you have enough PV every day to cycle it. I’ll make some guess work and say that roughly 20% of the days in a year you don’t do a cycle: 2,922 * 0.8 = approx. 2,340 cycles.

As such, over 10 years only about 4,700 kWh.

With the price of electricity at 2.92 (the most I can pay), this adds up to R13,724 of savings. Currently the cost of a battery seems to be roughly R21,000 (just quickly checked on an online store, sure the price might fluctuate and this might not be the cheapest).

So my questions:

1. How likely is it that you get more than 10 years of service from your battery if you treat it well? (I’ll make another post at some point asking advice on what it means to treat your battery well, let’s just assume for now that it is being treated well)
2. Are there any assumptions in my calculations that are wildly wrong?
3. Am I missing something because, unless you have a very large backup need, it would seem to me that it only makes economic sense to purchase enough batteries to match your inverter’s maximum demand, but no more than that?

(This edit was for science)

My calculation looks a bit different.

They state you can use 80% of the capacity 6000 times, therefore I do it as follows.

3.5kwh(Capacity) x 0.8(DOD) x .9(Conversion Losses) x 6000(Cycles) = Expected Saving of 15120kwh x R2.92 = R44150

But I might be missing something.

I do not own a pylontech so I am going based on the spec sheet.

Useable capacity is stated as 3.2kWh (if you have a system that allows the BMS to drive the bus I assume the battery will disconnect at 80% DOD?) so the 3.5kWh is a false starting point me thinks.

If we assume the people at pylontech have a very good reason to have a 20% capacity floor, and if wanting to play extra safe then I guess your system preset for DOD should be 60% of nominal? (80% DOD on nominal capacity will be 100% of what pylontech have as the BMS cut-off).

So for basic cautious calculations I will lean towards
3.5kWh x 0.6(DOD) x 0.9 (losses) x 6000 (cycles) ~ 11300kWh x 2.92 (electricity rate) = R33 112 as the theoretical maximum amount you can prevent leaving your wallet and going to council/eskom. If you do one full cycle per day this is however ~ 16 years’ worth of cycles and by that time the warranty is long gone.

But unless you stole the battery R33 000 is not your saving. R21 000 (Battery Price) - R33 000 (not to eskom amount) = R12 000 is the actual saving potential over 6000 cycles. (obviously not allowing for best guess eskom increasess over time)

So then comes the question, if you had rather invested your 21k, would it have yielded 12k in interest over 16 years? The answer to that is of course yes, you’ll make twice that even at 5% interest.

But I suppose the debate is the same for owning a house (to live in) vs owning a house (for leasing). The first house has the additional advantage of also giving you a place to stay, while the second one must absolutely provide an above-average ROI or otherwise you’re going to get rid of it.

On that topic… OT and all… I predict a downturn in the property market, followed by an absolute shortage in about a decade.

To my mind the Return on Investment concept on solar panels / batteries etc. is a bit of smoke and mirrors at times - especially for most residential installs. Why? … simple (keep in mind my forum name / handle / nick or whatever the thing is called by the cool kids…) . I am pretty sure many installers/finance providers etc. can easily show that PV (? + battery) can yield a lower cost per kWh than eskom / municipality rates in a relatively short time (hell, over the past 10 years the average eskom increase is something like 13.6%… the most recent 5 years at ~ 6.3% being overlooked). Be that as it may, even if we assume you can produce electricity for own consumption cheaper than eskom from day 1, what do many (most?) people do with that saving? Probably nothing that will constitute “investment”… “Hey I have a R1000 extra this month because I beat eskom… so I only need to borrow R9000 to put new mags on the Uno”…

If you run a business and lower your overhead by getting cheaper electricity the ROI thing is more relevant.

1 Like

1. cover inverter and/or battery warranty requirement
2. cover essential loads when sun / eskom is not around

If not, then you want one and then the issue is most of the time “I think I want XYZ… see if you can talk me out of it”

1 Like

I based my degradation allowance off the warranty that says 60% nominal capacity, but it would change the equation if I don’t allow for it.

However, I think the biggest issue is still how to actually manage 6k cycles before 10 years are up, assuming something else breaks then. I’d obviously be stoked if it doesn’t!

Only if you assume 0% electricity price hikes, because then investing the R21k really doesn’t need to earn much to outperform the battery.

I’ve assumed that the hikes are equal to investment return, so that I don’t need to do a more complicated calculation. In my experience, if you can make reasonable assumptions and do a simplistic back-of-the-envolope calc, you wouldn’t be able to flip the outcome by doing a much more rigorous calculation… which might also just be me trying to justify my laziness for not doing a full discounted cash flow valuation.

How relevant is the warranty/10 years really? I come from the perspective that it is not like owning a car where you can (maybe) get your insurance to give you rental if your car goes up in flames (or a courtesy car if it goes for a service).

If the system needs the battery to function (i.e. no battery, no inverter coming on) then if the battery goes caput you are somewhat up the creek with a spaghetti strainer regardless of the warranty. Realistically how long will it take before the battery is replaced/repaired?

So you might as well work on 6000 cycles irrespective of time (will be interesting to find out whether time independent of the efffects of charge/discharge related thermal changes is correlated with some form of cell chemistry degradation).

The other questions we need to ask is, how do the bms work out the cycles. Say like mine is 140ah size the lifepo4 bank, and my first bank did 12000 CycleAH so far, how do you work out the cycles on that. Do I take 12000cycleAH and divide that with 140ah to get 85.71 cycles so far for the 246days it’s running.

My BMS seems to calculate a cycle as the full usage of Ah capacity.
So my battery is set as 80Ah. If I discharge half the battery twice (40Ah and another 40Ah the next day) then that would be 1 cycle.

So @Gman I would then use that and say device your capacity into CycleAH to get your cycles

@Louisvdw great stuff. Ok. So basically CycleAH divide with AH and you get your = Cycles.