BiduleOhm's project design WIP (30 kWh battery, 10 kW inverter, 16 kW PV)

[BiduleOhm]

Hi everyone,

I bought a house a year ago (first floor of 90 m² (970 ft²), second floor not yet done but should be about 60 m² (650 ft²), so about 150 m² (1620 ft²) total) not very far from Paris, France and I want to be 100 % off-grid.

My main problem will be to power a 4-4.5 kW (electrical, 16 kW heat) DIY heat pump during winter (so far less sun available than in summer...). Including everything else I'll need a total of of 16.4 kWh/day (5.5 kWh/day just for the heat pump) on average during the coldest months.

I've been doing some research since a few weeks about the details of what I'll need to meet these goals:


Battery

I plan to use 8 Tesla modules in a 48 V configuration (i.e. 4 pairs in // of 2 modules in series) and between 10 and 90 % SoC to extend the cycles life for a total of 33.6 kWh, 30.2 kWh usable with a 90 % efficient inverter, so a bit less than 2 days of autonomy worst case.

I'll make a DIY BMS per module pair based on an Arduino (or maybe a Raspberry Pi, I'll see the pros/cons) to monitor, balance and protect the cells + I'll have full hardware redundancy for the critical protections (over/under voltage, over/under temp, over current) as I'll definitely not rely 100 % on software for that. As it's Li-ion and not LiFePO4 safety will be number one priority everywhere.


Inverter

As I need to power a big inductive load (albeit with a VFD so soft-start is possible) I was thinking about a LF inverter, a 10 kW one (I should not use more than 8 kW 95 % of the time so 10 kW seems a good size to aim), this one to be exact: https://www.ebay.fr/itm/10000W-LF-P...C-48V-to-AC-230V-Charger-UPS-LCD/113493808631


PV panels

As I'm in a region with a big solar difference between summer and winter I'll need to over-panel like crazy to meet the demand in winter. I'm thinking of minimum 52 panels of 315 W (that's 89 m² and i'm not even sure the roof is big enough... I need to do the calculations about that) for a total of 16.4 kW.

I have a 4-sided roof with the smaller side facing SW and a bigger one facing SE and about 45° of inclination for both (so at least better for winter than summer, good news about that).

Edit: thanks to @nosys70 who catched a big mistake I change the number to double that; so 33 kW of panels.


SCC

See question 3.


Questions

1. Is the battery size enough for my needs? I'd like to put more but my budget is tight so...

2. Does anyone know if the inverter is ok quality-wise?

3. I'm debating using the MPP PIP 5048 series as it includes both the inverter and SCC but I'll need at least 3 of them to handle the solar power and I'm not sure about the AC quality and surge capability, especially when paralleled; do you have any input on that?

4. Do I need more panels? less panels?

I'm open to any other advice you have to offer (I can provide more data and calculations details if needed), and thanks for reading this long post

 

[nosys70]

you need more solar power than battery, else how will you recharge them (and not only recharge, but also run device when sun shines) ?
usually you are running battery longer than solar, since useable daylight is about 5 hours, and a active day frorm 8h to 22h = 14H.
So in pure theory you would need 3 times more solar power than battery power (just to charge battery)
il faut plus de details

 

[BiduleOhm]

I see, I calculated for 1 day usage: (16.4 kWh / 5 h) / 0.2 = 16.4 kW (the 0.2 is for the winter correction) so it's basically a 1:1 relation here. So it's okay in normal conditions but I neglected the battery recharge need after a day of no sun for example, luckily you did catch that, thanks

I don't need to recharge the whole battery in one day, but I do need more solar. So if I want to be able to compensate in one day for one day without sun for example (seems reasonable) I need twice what I calculated. Hum... I really need to see what space I have on the roof...

Also, I forgot to mention it in the OP but I'm ok with using a small (2-3 kW) gas generator for the really bad days, but the goal is still 100 % solar so it should be really exceptional to use it, like a few times per year.

 

[nosys70]

if you want to skip bad days, then you can plan for a huge battery (like 1200Ah) that you will hardly discharge from day to day, but in case you need to run several days on battery , there would be no problem. Then you will need also several days to recharge or use the generator.
it is all about why you need solar, to fill in excessive but occasional consumption, to get a baseline, to act as UPS in ermegency situation etc...
usually the best is to plan like if you run only on solar. then calulate battery capacity for the remaining time (that could be the same as the one you calculated for solar) , then add this capacity as a load (recharging).
so you can fill a full cycle. then if you want to skip a cycle , you just need to multiply battery capacity by 2, 3, or more.

 

[BiduleOhm]

I can't do that because of the limited money, however I can more easily add panels (far cheaper than battery) the problem then is roof space.

The goal is "just big enough to run off-grid with the very very occasional help of a small generator".

The whys are because I'm in a pretty rural zone with aerial power lines so I know that too much wind, thunderstorms, etc. will sooner or later cause power outages, also because I plan to live here for a while so I should be able to have some ROI.

Now what I can do is test with not too much panels for a year or two while still having grid power as a backup, and then change what I need to change to solve any problem I had and be really off-grid.

 

[nosys70]

i would also skip on elctrical heating that is common thing in france but an heresy everywhere else (turning eletricity into heat... beurk!)
ultra vacuum tube can produce hot water even under snow.

 

[BiduleOhm]

Well the heat pump is electrical but with a COP of 4 it's only 1/4 of what pure resistive heating (what the house had actually...) would use.

 

[BiduleOhm]

Ok, I just found this thread: https://diysolarforum.com/threads/j...1260-00-usd-delivered-grade-a-to-canada.3387/ and I must be dreaming because 31 kWh (usable) of these would cost me 6.1 k€ where 30.2 kWh (usable) of Tesla modules would cost me 10.4 k€ so that's almost half the price of what I thought was the cheapest battery I could find and with the advantages of LiFePO4 on top of that (mainly: the safety).

Did I miss something? because I'm really close to click "order"...

Edit: after more calculations I can have 51.6 kWh (usable) for 10.2 k€ (almost the same price as for the 30 kWh of Tesla modules) so a bit more that what's needed for 3 days of autonomy (49.2 kWh).

 

[nosys70]

yep just ordered 8 cells 2 weeks ago, as per UPS tracking they are currently in Corbeil Essonnes, France.
1200 swiss francs for a 290A 24V (6KWh) . Also need to add BMS (i took 200A dely common port)
but frankly , if you need huge battery capacity, OPzV cell will be cheaper up to 2000A and up to 10C discharge
for 220V,it is advised to go at least 48V, so it makes 32 cells to get 500A at 48V ( 24Kw).
start to be a nightmarish cell management and these cells are not made for high current.
(terminals too small, discharge rate 1C)
Make sure also that if you have 24KWh, it is 5KWh of panel to juste charge the battery.
For such intensity, i would go with OPzV2 at 600A or more. (lead should not be discharged more than 40%)

 

[BiduleOhm]

I should have been more precise: cheaper per kWh/kcycles, not cheaper to buy at the start, so everything lead based is dead and buried (also, anything lead based would be too heavy and big).

Yep, BMS isn't a problem, I can buy a cheap one (Daly as Will loves them apparently) or make my own. I will probably do that as it'll be far more flexible and overbuilt than a Chinese BMS (I don't like to mix Chinese electronics and hundred of Amps...).

 

[nosys70]

for such high amount of current, you will just use the bms for measurement, and put a relay (500A) instead using the mosfet of the BMS.
i plan to use the battery a 100A maxi so i oversized BMS and battery by factor 2 to get more capacity on the duration, but the inverter will be
2400W anyway. i think going with these battery for a 30KWh capacity is a serious risk to go wrong.

 

[BiduleOhm]

I plan to make a solid state relay (to avoid the power hungry coil of a classic relay) driven by the arduino/RPi based BMS and the hardware safeties.

And as I'll have a few strings sharing the current it'll not be super high anyway (10 kW is roughly 200 A @ 48 V so with 4 strings in // that's only 50 A per string) but I plan on being able to run on one string if needed (redundancy...) so I'll size each string to be able to supply 200 A.

Go wrong as in fire/smoke/whatever? or as in made in china quality?

 

[nosys70]

if you move 500ish amps, the relay coil will be the last of your problem.
i mean if you take the cheapest cell on the market, there is a string attached for sure.
does not mean it is bad product, just means you get what you pay.

for the "each string to be able to supply 200 A" you are aware then the battery will be empty in just over one hour ?

 

[BiduleOhm]

Why 500 A ? and having something using 10-20 W all the time for no good reason is not really great, wether you switch 5 A or 500 A.

Ok I see, as with everything else from China. The tests performed thus far are very encouraging though. I'm tempted to wait for more tests but at the same time I don't want the supplier to run out and have no cells left when I want to buy them.

Yes, I know, but that would be at full load (should pretty much never happen, especially if I know I'm on one string only and I'm a bit careful) and it's only in case of a severe problem with rest of the battery (again, should never happen either but you never know).

It's just that it's easy and very cheap to add some redundancy now even if I'll probably never need it, because if I ever need it I will be very thankful as I'll probably be in a crappy situation to start with.

 

[Solarfun4jim]

Just thinking out load...and just noted the date of the previous last post.....so your project might already be completed...lol

Pity you couldn't put in a ground source heat pump instead of an air source heat pump, as i think the COP doubles, plus it's much more stable all year round. If you had PV panels on the best two roof spaces and over capacity evacuated tubes on the third roof space, then when your solar thermal store was topped up, the extra capacity can be dumped to floor space heating in winter or else the ground collector during summer.

 

[michael d]

I see, I calculated for 1 day usage: (16.4 kWh / 5 h) / 0.2 = 16.4 kW (the 0.2 is for the winter correction) so it's basically a 1:1 relation here. So it's okay in normal conditions but I neglected the battery recharge need after a day of no sun for example, luckily you did catch that, thanks

I don't need to recharge the whole battery in one day, but I do need more solar. So if I want to be able to compensate in one day for one day without sun for example (seems reasonable) I need twice what I calculated. Hum... I really need to see what space I have on the roof...

Also, I forgot to mention it in the OP but I'm ok with using a small (2-3 kW) gas generator for the really bad days, but the goal is still 100 % solar so it should be really exceptional to use it, like a few times per year.

Generator noise is rather obnoxious, but you may have it (or want it) for an emergency just in case( i wore out 3 or 4 generators in off-grid use in the mountains of Bolivia where no electricity was available. I would suggest more solar panels and buy a bigger battery bank. I bought 64 272Ah Lishen cells for about 90 dollars each including shipping. the second 32 cells should be here this week as FedEx finally has them here in the USA.
use dc lights where possible. I bought 8 48-volt led dc powered lights the work great directly connected to the battery without any inverter.
solar panels will put out about 10 percent on a cloudy day. mine do.
my current project is in South Dakota.
cheers!
France sounds like an interesting place!

 

[Solarfun4jim]

I know you are good at designing electronics, but how are you with heat pumps? How about designing a ground source heat pump type system, but one which utilises an evacuated tube array instead of a ground collector, thus giving a much 'higher grade' heat source to start with and potentially a better COP???

 

[BiduleOhm]

NB: this thread is outdated as for example I changed the battery type I'll use from Tesla modules to EVE 280 Ah cells.

Just thinking out load...and just noted the date of the previous last post.....so your project might already be completed...lol

Pity you couldn't put in a ground source heat pump instead of an air source heat pump, as i think the COP doubles, plus it's much more stable all year round. If you had PV panels on the best two roof spaces and over capacity evacuated tubes on the third roof space, then when your solar thermal store was topped up, the extra capacity can be dumped to floor space heating in winter or else the ground collector during summer.

Nope, sadly the practical side didn't advance much, only the design and theorical side (badly injured my arm and then the covid lockdowns...), so there's time to make changes before I start doing the real thing

Actually it'll be a ground source HP more infos on this thread (in FR): https://tech-masters.fr/showthread.php?tid=25

Generator noise is rather obnoxious, but you may have it (or want it) for an emergency just in case( i wore out 3 or 4 generators in off-grid use in the mountains of Bolivia where no electricity was available.

Yes, that's exactly what I'm planning. I shouldn't have to use a generator at all, only for emergencies

I would suggest more solar panels and buy a bigger battery bank.

Problem with more panels is that I need more space and my roof isn't infinite. But I need to do the exact calculations on that now I'm more familiar with the subject. This thread was a first try at the project, I need to make some changes I have in mind and redo the calculations precisely.

I bought 64 272Ah Lishen cells for about 90 dollars each including shipping. the second 32 cells should be here this week as FedEx finally has them here in the USA.

Yep, I plan on using a 4x 16s 280 Ah battery bank.

use dc lights where possible. I bought 8 48-volt led dc powered lights the work great directly connected to the battery without any inverter.

Yes, I even plan to have DC outlets besides the AC outlets because lots of devices can be run on DC easily...

solar panels will put out about 10 percent on a cloudy day. mine do.
my current project is in South Dakota.
cheers!
France sounds like an interesting place!

Thanks for the infos, and yes, France is pretty nice

I know you are good at designing electronics, but how are you with heat pumps?

HPs are actually simpler to design than any serious electronic project. The only serious problem I have is to find a 48 V DC compressor (or more realistically 2 lower power ones). Otherwise I'll need a big inverter, although I could also have a big DC/DC PSU and power the VFD directly on its DC bus, but that's still one more conversion than with 48 V compressors (so more losses) and more money needed.

How about designing a ground source heat pump type system, but one which utilises an evacuated tube array instead of a ground collector, thus giving a much 'higher grade' heat source to start with and potentially a better COP???

I thought about thermal panels but they're expensive and not really adapted to a HP (unless you have a big water tank as a buffer, ground is far better) so you might as well heat your water directly with them instead of using a HP. Also, I'm lucky to have pretty ideal soil (very wet, shouldn't have rocks, etc...) so there's not really a reason to switch to thermal panels.

 

[michael d]

one other thing to consider. I chose to do a ground mount as I did not want to climb up 20 plus feet to the lower edge of the roof on my 1840 to 1880 vintage house with a ladder in the winter or summer for that matter. I do enough ladder work without asking for more. climb a ladder in the wind and you will realize the roof is not always the best choice. tie that ladder off at all times for safety. I do now.
it also needs the roof replaced. I built a separate insulated power shed for all the batteries and solar charge controllers with a locking door to keep the young ones away and any other nosy people. I went with 24-volt inverters but I also have a 48volt build going. I went with Electrodacus SBMS0 with Electrodacus DSSR20's for the charge controller. The SBMS0 is the brain or controller and protector for the expensive battery bank LiFePO4 wired 2P8S for a 13,926.4 Wh potential capacity using 16 272Ah Lishen cells. I am going to build 4 of these battery banks so will have over 13926.4 x 4 = 55,795.6 Wh capacity potential in these 4 24-volt LiFePO4 batteries.
easy to clean the snow off the ground mount. just go over and broom it off.
I have 2 arrays I am working on now. of course, the south-facing array works great now, the west-facing array is to gain a little solar later in the day.

 

[michael d]

one other thing to consider. I chose to do a ground mount as I did not want to climb up 20 plus feet to the lower edge of the roof on my 1840 to 1880 vintage house with a ladder in the winter or summer for that matter. I do enough ladder work without asking for more. climb a ladder in the wind and you will realize the roof is not always the best choice. tie that ladder off at all times for safety. I do now.
it also needs the roof replaced. I built a separate insulated power shed for all the batteries and solar charge controllers with a locking door to keep the young ones away and any other nosy people. I went with 24-volt inverters but I also have a 48volt build going. I went with Electrodacus SBMS0 with Electrodacus DSSR20's for the charge controller. The SBMS0 is the brain or controller and protector for the expensive battery bank LiFePO4 wired 2P8S for a 13,926.4 Wh potential capacity using 16 272Ah Lishen cells. I am going to build 4 of these battery banks so will have over 13926.4 x 4 = 55,795.6 Wh capacity potential in these 4 24-volt LiFePO4 batteries.
easy to clean the snow off the ground mount. just go over and broom it off.
I have 2 arrays I am working on now. of course, the south-facing array works great now, the west-facing array is to gain a little solar later in the day.

the picture today January 12th, 2021 is the Electrodacus with only 4 250-watt 60 cell polycrystalline panels wired in parallel in sets of 2. 2 panels per Electrodacus DSSR20 which is controlled by the Electrodacus SBMS0. the SBMS0 is the BMS. the DSSR20 are solid-state charge controllers. the SBMS0 can control up to 18,000 watts of PV array. this system I am putting together will have 10,000 watts running the 24-volt battery banks, etc.
each DSSR20 can control 2 60 or 2 72 cell panels and has a 20 amp rating but can go up to 24 amps each. the 4 panels (2 pairs) are putting out 30.194 Amps.

 

[BiduleOhm]

one other thing to consider. I chose to do a ground mount as I did not want to climb up 20 plus feet to the lower edge of the roof on my 1840 to 1880 vintage house with a ladder in the winter or summer for that matter. I do enough ladder work without asking for more. climb a ladder in the wind and you will realize the roof is not always the best choice. tie that ladder off at all times for safety. I do now.
it also needs the roof replaced. I built a separate insulated power shed for all the batteries and solar charge controllers with a locking door to keep the young ones away and any other nosy people. I went with 24-volt inverters but I also have a 48volt build going. I went with Electrodacus SBMS0 with Electrodacus DSSR20's for the charge controller. The SBMS0 is the brain or controller and protector for the expensive battery bank LiFePO4 wired 2P8S for a 13,926.4 Wh potential capacity using 16 272Ah Lishen cells. I am going to build 4 of these battery banks so will have over 13926.4 x 4 = 55,795.6 Wh capacity potential in these 4 24-volt LiFePO4 batteries.
easy to clean the snow off the ground mount. just go over and broom it off.
I have 2 arrays I am working on now. of course, the south-facing array works great now, the west-facing array is to gain a little solar later in the day.

Ground mounted array isn't really ideal because I have a lot of trees around my garden, and for aesthetic reasons. If I really need more panels I could in theory add a pergola on the terrace and put panels on it, but I'd rather not have to do that.

 

[michael d]

Ground mounted array isn't really ideal because I have a lot of trees around my garden, and for aesthetic reasons. If I really need more panels I could in theory add a pergola on the terrace and put panels on it, but I'd rather not have to do that.

I totally understand the aesthetics issues.
I am out in the country without any real close neighbors on a long private driveway, so free to play with panel location a bit more.
did you look at the Electrodacus dssr20 with diversion as a heating option? the excess electricity from the solar panels can be diverted into hot water heating or into resistive wire heating. I bought 4 of the dssr20's to try out and bought 3 600 watt 24-volt water heating elements.
Dacian also talks about using 2 panels to generate heat to one panel by shorting 1/3 of the one panel.. you could also increase the panel mounting area by putting arrays on the other sides of the roof if possible. others talk about this idea.
I am using a big array small array setup idea and will have the big array facing south and the small array facing west.
time will tell how it all works but currently working good.
I like the Electrodacus approach to keep it simple and low cost - more bang for your buck.
the solid-state relays (dssr20's) replace the mppt's.
just some ideas for you to think about!

 

[BiduleOhm]

did you look at the Electrodacus dssr20 with diversion as a heating option? the excess electricity from the solar panels can be diverted into hot water heating or into resistive wire heating. I bought 4 of the dssr20's to try out and bought 3 600 watt 24-volt water heating elements.

Yes, long story short I think the design isn't adapted at all to my case; not only a 24 V system would make wires bigger (and I already need pretty big ones...) but the direct connection PV/battery would need to have hundred of € in wires and it would still have bigger losses than with a high voltage system (I plan to be at around 300-400 V on the PV side). Also, given clouds and rain are common here I would basically crush the already low power available in winter without a MPPT.

I'll have diversion to resistive elements to heat water

Dacian also talks about using 2 panels to generate heat to one panel by shorting 1/3 of the one panel..

What would be the goal of that besides loosing precious surface area?

you could also increase the panel mounting area by putting arrays on the other sides of the roof if possible. others talk about this idea.
I am using a big array small array setup idea and will have the big array facing south and the small array facing west.
time will tell how it all works but currently working good.

I have a 4 sided roof and the sides are facing almost perfectly N-E, S-E, S-W and N-W. What I thought was a disavantage at first is actually an advantage as I can overpanel more and not have a big spike in summer as if the panels were directly S oriented.

I like the Electrodacus approach to keep it simple and low cost - more bang for your buck.
the solid-state relays (dssr20's) replace the mppt's.

Not until you factor in the cost of the wire. And I can't just cheap out and use thinner wires as then losses would be enormous and efficiency is a priority here. And, again, no MPPT in winter...

 

[michael d]

Yes, long story short I think the design isn't adapted at all to my case; not only a 24 V system would make wires bigger (and I already need pretty big ones...) but the direct connection PV/battery would need to have hundred of € in wires and it would still have bigger losses than with a high voltage system (I plan to be at around 300-400 V on the PV side). Also, given clouds and rain are common here I would basically crush the already low power available in winter without a MPPT.

I'll have diversion to resistive elements to heat water




What would be the goal of that besides loosing precious surface area?




I have a 4 sided roof and the sides are facing almost perfectly N-E, S-E, S-W and N-W. What I thought was a disavantage at first is actually an advantage as I can overpanel more and not have a big spike in summer as if the panels were directly S oriented.




Not until you factor in the cost of the wire. And I can't just cheap out and use thinner wires as then losses would be enormous and efficiency is a priority here. And, again, no MPPT in winter...

the panel with 1/3 shorted would be located in your living quarters where you want the heat. the 2 other panels would be outside. this requires no other components.
I am not sure where everyone(some people report incorrectly) keeps saying it takes larger gauge wire they are obviously wrong and mistaken. i have all panels wired with standard 10 gauge wire. i use PV red and PV black 10 gauge for easy color-coding. 10 gauge is the same wire use on almost all commercial PV solar panels. no big wire sizing required. I have read their incorrect quotations about the wire size and they are definitely wrong.
my polycrystalline 250-watt 60 cell panels produce 10 percent on cloudy days.
if mounting real estate is short the monocrystalline put out slightly better with a slightly smaller footprint. the advantage of dssr20 is no combiner boxes, no mppt's with short lives due to the capacitors. no high voltage to have to deal with special permits.
all good for a DIY person.
today I am hooking up the 3rd set of 2 panels.
if your distance to the panels to the dssr20 is less than 60 feet 10 gauge is fine for 60 cell panels up to 75 feet then you may want to use 72 cell panels. mine work great at about 60 feet with 10 gauge PV wire. I am in winter here now and the solar panels prefer the cold. no mppt in use in this off-grid build. just the solid-state relay dssr20. below freezing every night and almost all day for the next month or 2.
24volt LF PSW inverters are readily available on the internet and the distance from the inverter to the battery is always short 6 feet or less. I use 4/0 cable from battery to busbar which I connect the inverter to after the class t fuse. i have a 6000-watt LF PSW split phase inverter with 400 amp class t fuse.
just explaining how it (Electrodacus system)works. you decide what you build. there are 50 ways to accomplish the same thing.
good luck keep up the good work.

 

[BiduleOhm]

the panel with 1/3 shorted would be located in your living quarters where you want the heat. the 2 other panels would be outside. this requires no other components

So you use a PV panel as a resistor? I can't see any advantages to do that and it would be an expensive resistor...

Plus it would have a COP of 1 were I'll have a COP of around 4 thanks to the heatpump, so you would need a lot more panels and battery. Resistive heating is so inefficient it doesn't make sense to use it (excepted for power diversion for example).

NB: I'll have hydronic floors.

I am not sure where everyone(some people report incorrectly) keeps saying it takes larger gauge wire they are obviously wrong and mistaken. i have all panels wired with standard 10 gauge wire. i use PV red and PV black 10 gauge for easy color-coding. 10 gauge is the same wire use on almost all commercial PV solar panels. no big wire sizing required. I have read their incorrect quotations about the wire size and they are definitely wrong.

Let's take an example: 10 kW array with 350 V vs 10 kW array with 24 V.

At 350 V you need to transport 28.6 A, only one wire is enough to do that with say 3 % losses.

Now at 24 V you'll need to transport 416 A... so you'll need two 4/0 gauge wires at best or a bunch of smaller wires (essentially one per one or two panel(s)) to have the same 3 % losses. I'll let you compare the cost for a battery which is 50 feet from the panels for example... plus 4/0 wire (or a bunch of 10 AWG ones if you prefer) will be a PITA to route.

the advantage of dssr20 is no combiner boxes, no mppt's with short lives due to the capacitors. no high voltage to have to deal with special permits.

I'll not need combiner boxes. Capacitors are easily replaced. I'm off grid, I don't need permits.

just explaining how it (Electrodacus system)works. you decide what you build. there are 50 ways to accomplish the same thing.

I know how it works, and that's why I know it's not adapted for my case.

good luck keep up the good work.

Thanks, you too