-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Plan: Off grid solar with a Model S battery pack at the heart
- Thread starter wk057
- Start date
-
- Tags
- Energy Environment Policy
While my setup is certainly on the larger end of the spectrum, no offense, but yours is practically on the useless end. 2.5kWh of storage? Even before efficiency losses that would power the average home (according to the EIA at just over 900kWh/mo average) for almost 2 hours. Unless I missed something and Earth now has a 3 hour rotation I'm pretty sure this is useless for the vast majority of people as far as an off-grid solar setup goes and pretty useless even for standby power.
Assuming storage sized for 72 hours of normal usage (which seems pretty reasonable for a true off-grid setup) you've sized your setup for an average load of less than 35W. That's like a couple of LED bulbs. Sorry, no one in the 21st century can practically live with that type of energy constraint. You can barely charge a phone and run a light with that. So, if you will, please stop referring to your setup for comparison as if it were actually practical for anyone.
Now for something realistic, taking the average home at 900 kWh/mo, that's 30 kWh/day or 90 kWh for 72 hours of power. We'll start with that.
Storage: That's just over one full Model S 85 kWh pack. Would be 17 modules, or just about half of my rack. We'll round down to 16 for simplicity and multiple of 4 (nice even discharges/charges). Half of my rack is still 44" high and would take up the same floor space. If I dropped it to one column it'd take up half of the floor space but be the same height. And this is using Tesla's cells, which are more energy dense than anything else readily available.
Power: Then, for 30kWh/day we're looking at an average load of 1.25kW. Per NEC, for a battery based setup, the minimum inverter continuous output size is the size of the largest single load possible. For most people this will be either a hot water heater, an electric range, or a central AC/heat pump unit... all of which clock in somewhere in the 5kW area. Add a 40A EV charger and we're at 10kW. Add common sense (someone will take a shower or cook some spaghetti while the AC is running one day, I promise) and the minimum certainly isn't enough for normal use and should be at least doubled. Using the 5kW number, that's 10kW. I'll be conservative and say that one of my 8kW inverter units would meet a minimum requirement for this setup due to it's exceptional surge capacity.
Space: So, even using the half-height rack, and building the balance of the setup on top of it (inverter, a couple of charge controllers, transfer switch, etc) assuming it would all fit the battery bank would still be the largest component. And this is using Tesla's modules. Jump into LiFePO4 and we're *at least* doubled the volume (and weight, which isn't as much of an issue) of the battery portion. We're still looking at a minimum of about a 4'x4' area of floor space needed. To make it legal we need NEC working space requirements, so, this isn't going to get stuffed in a closet somewhere. We're talking an additional 4' wide area at least another 3' out from all sides of the equipment that needs to be able to be accessed. Again, I'll be generous/conservative and say just the front side. So we're now at 4'x7'.
That's about 32 square feet absolute minimum space needed to make a minimal off-grid setup for the average home using current tech and rules and the most energy dense battery tech available, not even counting solar panels or other renewable input.
Let's switch to LiFEPO4. We'll say 2x as large for the same energy (your number). Well, at 2x as large we'd need the full height rack setup again, so no putting the equipment on top. So we're up to an 8x7 area now, 56 square feet, minimum, after working space requirements.
Now for example, my mother's house is pretty average size. Her electricity consumption is a hair above average, but we'll ignore that for now. I'd be hard pressed to find even 32 square feet somewhere at her place where it would be practical to set this up, but it would be doable probably by taking a chunk out of an existing room. Bump that to the LiFePO4 size of 56 square feet and we're talking about most of a small bedroom for a usable off-grid storage setup with this tech.
Oh, and we haven't even considered ventilation needs for this space either.
Suffice it to say, anyone saying that energy density vs volume/floor space for this application isn't an issue is just being ridiculous. Granted there are likely ways to save floor space, technically, with any setup (a raised false floor with the batteries below was an idea I had for my setup at one point), but they don't change the actual volume needed.
----- *sigh* -----
OK, now the cycles life nonsense.
First, you're contradicting yourself. You say "show me the spec for NCA" then in the same post claim that the NCA has a 300 cycle life. Make up your mind. Either you have the specifications for the cell, or you don't. Since Tesla has not released them, and it is unlikely you work in Tesla's engineering department, I'm going to put my money on you not having the cells specs.
Now, even comparing to crappy NCR variants with the 300 cycle count, these are full DOD cycles. No one actually does this in practice, and neither will I. It is well know that the cycle count is extended dramatically on NCR chemistry by limiting the DOD on the extreme ends by even a few %.
While I don't have the exact data yet, either, using data available from Model S owners it is already very clear that this same 300 cycle count does not apply to the Tesla cells even in the harsh EV charge/discharge environment. Combine very low C charging and discharging with slightly less max DOD in an off-grid setup like mine. Even ignoring that realistic average DOD with a large pack like mine is very low, something like 20% overnight, I'd expect far more than 300 cycles.
But anyway, in a couple of months time I should have both some real world and controlled experiment data on the cells from my off-grid setup and experiments with single cells, respectively. I plan on using the cells from the module I'm breaking down to do cycle life testing at various DOD in various SoC windows at various average current draws. I expect that the longest of these planned tests, using 1000 cycles and 1A average discharge/charge (1/3C) rate will take about 6 months. 1000 cycles at 10A discharge and 0.7C charge should take about a month assuming no thermal failures. I'll be running multiple tests with multiple cells simultaneously (two of each type of test, probably 6 different tests) once I get time to get it all setup.
Last edited:
Solar portion still needs to be permitted and inspected, just don't need certain things required for grid tie setup like an interconnect application and other such things.
I expect to be calling for inspection on phase two (basically everything except the solar panels) next week.
I have the balance of the modules hooked up and charged to about 90% SoC with the inverters in UPS mode. I essentially have a pretty bad ass whole house battery backup for now until I get the solar installed.
Just some finishing touches (casing around batteries) and cleaning up before I call for inspection and move on to the solar portion.
I suspect that your inspector will be winging it anyway, given the uniqueness of this type of setup. You've already shown your attention to NEC detail, so it's likely you'll be doing more education.
electrodacus
Member
2.5kWh storage is not useless to me. My house is more energy efficient than an average house by quite some margins about 10x was my target. My house is about 3x smaller than an average house and about 3x better thermally isolated so 3x3 close to 10xSo my average 80kWh/month is also around 10x lower than an average house but can easily scale to that you can get a 25kWh LifePO4 pack and larger solar PV array.
I'm also in a cold climate and the heating requirement is also about 10x lower with no cooling required. Actuality for comparison I need about 1000kWh in the coldest month for heating that means that I can heat my house in this cold climate (Saskatchewan Canada) for the same amount of energy people use in an average North American house (average house in Canada uses about the same amount of energy as those in US).
Average house hold in Germany uses 284kWh and Italy about 228kWh that is average household there are many that use close to what I use for my smaller house.
All my electronics are used directly from 24Vdc that way I save 20kWh standby current of my 2400W inverter that is only used at most 1h/day when is needed for AC devices usually electric cooking (baking a bread 0.4kWh baking a 30cm (12") pizza 0.35kWh this is done normally during the sunny days so battery is mostly a buffer)
And I use my core i7-3770 with GT730 to replay to this comment and this computer works at least 12h/day no problem. I may use my Dell laptop just 10W idle in cloudy winter days to save some energy vs the 30W idle on this computer including the 18" full HD IPS LCD.
The Heating will also be done with PV panels but that will be a separate PV array about 12x larger than this one and will require no chemical battery energy will be stored in a 14 cubic meter concrete slab that acts as a thermal mass or thermal battery.
Is quite understandable why most houses need cooling even in spring since all that energy 900kWh average house ends up as heat inside the house (remember that is what I need to keep my house warm in a cold winter month).
You have not read probably my earlier comments on the subject. My position is that NCA is not better in cycle life than NCR or at leas not by any significant margin. If you have the spec to contradict me please do so. In a car battery is not cycled 100% DOD as in that test so it will probably last much more cycles in this conditions about 1200 from my rough estimation with cycles between SOC 20% and SOC 80% or about 300km range that gets me to 240km x 1200cycles = 288000km say new NCA is 25 % better at cycle life that is 360000km before 20% of original capacity is lost and that is probably good enough for an EV with limited range. Is anyone here with a Tesla and over 300000km about 200000miles of driving ?
I have an old gasoline car and I'm not even at half that mileage.
It makes no sense to repeat the calculations here but even 1500 cycles at 60% DOD equivalent will not compare with 6000 of 100% DOD on the Sony battery or even 7000 cycles 70% DOD on Winston. And reducing the DOD on Sony or Winston or any other LiFePO4 of course will increase significantly the number of cycles.
LiFePO4 is just superior when it comes to cycle life in the detriment of energy density so much better suited for stationary energy storage. The safety aspect of LiFePO4 is also important (no thermal runaway) then maybe even the fact that is more environmental friendly no Cobalt and uses less lithium for the same capacity may be considered important by some.
With cars you care less about this since you want the best range so energy density is the most important all other things are secondary at best.
Thanks for doing this that will be great but it will take quite some time as you correctly estimated. I will also test the capacity loss on my GBS battery in about one moth from now when they get 2 years off full offgrid use and one extra year of storage before. I have a video with the initial capacity test and I will probably be able to replicate that accurately since I still have all that equipment.
Dacian.
I suspect that your inspector will be winging it anyway, given the uniqueness of this type of setup. You've already shown your attention to NEC detail, so it's likely you'll be doing more education.
This is pretty likely.
You again make an assertion, then ask for the specs. Again, you can't have it both ways. Either you have the data/spec and can make such an assertion, or you do not.
@wk057 - Your project looks amazing. I was wondering if you would give me your thoughts on the dimensions needed to take my house off the grid.
Now I don't know what US citizens do to use that much electricity but that's another discussion. Our house uses about 350 kWh per month with a peak daily usage of 16 kWh. This is actually pretty bad for 700 square feet home but we have some electric heating and that really sucks up the juice.
Thoughts on battery dimensions, solar panel size and so forth would be much appreciated.
Now I don't know what US citizens do to use that much electricity but that's another discussion. Our house uses about 350 kWh per month with a peak daily usage of 16 kWh. This is actually pretty bad for 700 square feet home but we have some electric heating and that really sucks up the juice.
Thoughts on battery dimensions, solar panel size and so forth would be much appreciated.
JRP3
Hyperactive Member
As I posted up above, and you apparently ignored the data, I'll try it again. http://www.teslamotorsclub.com/show...k-at-the-heart?p=954009&viewfull=1#post954009
Simply put your position is just wrong, no matter how often you repeat it.
mitch672
Active Member
This has not been said, but I am going to say it:
electrodacus, no one cares about your puny system, and we don't care to debate cell chemistry on wk's thread about HIS project, we read this thread for how HE is using salvaged Model S packs, Please stop "threadjacking" and create your own thread for YOUR system
electrodacus, no one cares about your puny system, and we don't care to debate cell chemistry on wk's thread about HIS project, we read this thread for how HE is using salvaged Model S packs, Please stop "threadjacking" and create your own thread for YOUR system
Last edited:
Not quite true: I am interested in seeing both extremes--both wk057's setup as well as electrodacus's setup. I agree, however, that electrodacus should stop claiming that NCA is not better in cycle life than NCR without proof, and that he should create his own thread to discuss his system.
I think electrodacus is doing well in reminding us how we have gotten used to an extravagant energy-consumptive lifestyle, and we should reduce our energy consumption. I am amazed that he can live in Saskatchewan, Canada, and not use more electricity. I would like to see more detail about his setup (in his own thread!)--electrodacus, how do you heat your home? Surely not electric! Oil? Wood? Is your electric consumption so low because your home is vacant most of the day? Or, do you use alternative lighting sources, such as oil lamps? How many people are supported by your setup? Just yourself, or do you have a family?
wk057, I am envious! You are inspiring me to consider blocking off a large portion of my basement & roof to reduce my dependance on the electric grid!
jvonbokel
John VonBokel
Sounds like Tesla is not far behind you, wk057.
Tesla Motors, Inc. Will Announce a "Major New Product Line" in April (SCTY, TSLA)
Tesla Motors, Inc. Will Announce a "Major New Product Line" in April (SCTY, TSLA)
electrodacus
Member
I do repeat my position because is the right one base on the current evidence including that document.
In that document there is a 3000 cycles test done at a very narrow DOD from 4.05V down to 3.6V and back up (that based on the okashira discharge curves will represent less than 30% DOD) I'm sure any conventional LiCoO2 will be able to perform the same or extremely close to this in the same test situations.
So yes having 25 or 30% DOD with 3000 cycles will still get you to about the same numbers I posted earlier with around 300000km before 20% capacity is lost it just depends how you present the numbers.
With this sort of narrow depth of discharge a LiFePO4 can do 20k to 50k cycles based on test on many batteries made by different manufacturers so is more about the chemistry than about the manufacturer.
Note also that those tests where done at high charge discharge rates and degradation is mostly linked with the amount of time the battery is spending at high voltage and in solar applications charging is done much slower thus battery spends much more time at those high voltages.
- - - Updated - - -
I will like to talk more in details about my system but this is a Tesla forum and I do not think posting about my solar Off Grid system that has noting to do with Tesla car or battery is a good idea.
Look the replay I made to JRP3 should be above this one regarding NCA and that available document on the 400mAh cells from Panasonic.
For now as temporary solution I heat with propane I use about 45 small 20lb barbecue propane tanks per heating season about 10 to 12 in the peak heating month usually January. Each of those propane tanks contain almost exactly 100kWh of energy and my propane heater is about 80% efficient.
I will use electric as final heating solution and for that I will need a PV array about 12x the size of the current one about 9kW array that will be directly connected to resistive heating loops embedded on concrete floor under ceramic tiles (I will need to design and build what I call a Digital MPPT)
No chemical battery will be needed for this part and energy will be store as heat in the 14 cubic meter concrete slab that acts as a thermal storage able to store about 10kWh/degree celsius.
I do live here full time so the house is not vacant (I'm now there writing this message from my computer) Light is done with LED I have a 5W one in my lab working about 16h a day and another one in bedroom working less about 8h a day same 5W I have some other Lights but those are used much less not significant. Just me and my wife so two people are living here full time.
JRP3
Hyperactive Member
Maybe you didn't notice that the test stopped at 3k cycles but the flat capacity curve in no way suggests 3k cycles is the maximum.
For my purposes, if the cells have 50% capacity in ~15 years (pretty good estimate for ~3000 cycles worth of charge/discharge in my setup taking day and night usage into account in my simulation) I'll be happy. By then there will likely be something better worth replacing them with at a reasonable cost.... perhaps even more of the same salvage modules
Keep in mind that from sun down to sun up in my setup at most I may use, on a bad day, 50% of the pack capacity overnight. Solar will, even on a crappy day, offset the majority of my day time usage plus be able to recharge the pack a bit. On a cloudless spring/summer day the pack could technically be recharged from 0% to 100% by about noon or 1PM while also maintaining home loads. On a crappy day (5-10% capacity, reasonable for a cloudy/rainy day) I could still recover something like up to 1/4 of the pack capacity during the day with no direct sunlight since my array size is so huge. It will be pretty rare for there to be a day where there would be no usable solar energy.
My setup is about as easy on the cell chemistry as things can possibly get for a real world use case. Super low charge and discharge rates. My average discharge rate will be less than 1/50C! My max peak discharge is just about 1/3C.... and that's running two HPWC @ 20kW plus max winter HVAC (something unlikely to happen often, but the setup is sized for it). For charging the peak rate will be under 1/4C, with averages a bit lower. For comparison, a supercharger hits about 1.4C with an average about 0.9-1C. Realistically, if I have less than 80% of day 1 capacity in 10 years I'll be very surprised.
Also keep in mind that I have no real reason to ever deep discharge the pack. At ~8-10% SoC my setup will just top off the cells from the grid or backup generator. That's why I'm not fully cutting my utility service. Sure my bill will be < $20/mo year round, but I'll have that connection for those times when I just need that extra power. Say a caravan of Model S comes through and needs a few hundred kWh of charging one day... I'll be able to oblige.
Keep in mind that from sun down to sun up in my setup at most I may use, on a bad day, 50% of the pack capacity overnight. Solar will, even on a crappy day, offset the majority of my day time usage plus be able to recharge the pack a bit. On a cloudless spring/summer day the pack could technically be recharged from 0% to 100% by about noon or 1PM while also maintaining home loads. On a crappy day (5-10% capacity, reasonable for a cloudy/rainy day) I could still recover something like up to 1/4 of the pack capacity during the day with no direct sunlight since my array size is so huge. It will be pretty rare for there to be a day where there would be no usable solar energy.
My setup is about as easy on the cell chemistry as things can possibly get for a real world use case. Super low charge and discharge rates. My average discharge rate will be less than 1/50C! My max peak discharge is just about 1/3C.... and that's running two HPWC @ 20kW plus max winter HVAC (something unlikely to happen often, but the setup is sized for it). For charging the peak rate will be under 1/4C, with averages a bit lower. For comparison, a supercharger hits about 1.4C with an average about 0.9-1C. Realistically, if I have less than 80% of day 1 capacity in 10 years I'll be very surprised.
Also keep in mind that I have no real reason to ever deep discharge the pack. At ~8-10% SoC my setup will just top off the cells from the grid or backup generator. That's why I'm not fully cutting my utility service. Sure my bill will be < $20/mo year round, but I'll have that connection for those times when I just need that extra power. Say a caravan of Model S comes through and needs a few hundred kWh of charging one day... I'll be able to oblige.
lolachampcar
Well-Known Member
In summary, can you list-
Battery capacity
Inverter capacity (AC)
PV capacity (DC)
Budgetary cost not including your labor (if you do not mind)
Thanks
Bill
Battery capacity
Inverter capacity (AC)
PV capacity (DC)
Budgetary cost not including your labor (if you do not mind)
Thanks
Bill
- Battery capacity
- 36 modules from 2.25x Tesla 85kWh packs
- 191.25 kWh (DC side)
- ~4,200 Ah
- 43.2V nominal @ 3.6V per cell
- 15,984 cells (!)
- Inverter capacity (8x Outback Radian GS8048A)
- 240VAC @ 60Hz w/neutral
- 64kW continuous AC output
- 30 minute surge: 72kW
- 5 second surge: 96kW
- 100ms surge: 135.76kW
- Grid->Battery Charging Capacity: 57kW
- Expected AC output from pack after safe SoC window and efficiency considerations: ~160 kWh usable AC
- PV Capacity (In Progress)
- 102 Panels @ 435W (20% efficiency)
- 44,370 Watts DC
- Split into 17 sets of 6 panels (3 parallel of 2 in series)
- 17 individual MPPT charge controllers (Midnite Solar Classic 200)
As for cost, I'm going to actually keep that total to myself for the moment. It, admittedly, has gotten a bit out of hand, but it has been too awesome of a project not to see through to the end... budget be damned.
Last edited:
