Tweet
Well, I think I've had my mind changed on NiFe. Thank you inetdog for recommending Mike's posts on his experiences with the technology. Sunking, I respect your experience and your knowledge, but your presentation comes across as a self righteous troll and I have a really hard time taking it seriously (it's how the internet at large has trained me). Maybe that works on others, but I need more than that to be convinced.
What did convince me is the following (most important to me first):
So, I suppose the next best technology for long-ish life and safe use would be Lithium Iron Phosphate. Reasons being:
Their biggest flaw is the high price. Someone in another thread found a good metric for calculating a common price point: dollars per cycle. I think this can be further improved by equalizing the price per kWh for different DoDs. I think this represents the true cost of having a battery running the house at night, compared to what the grid can provide, and it makes much more sense than just throwing numbers without context around. I wish battery makers would give these numbers, but of course they don't want to showcase how horrible it actually is to run off a battery with a short lifespan. We can go even further and determine the cost to generate the power necessary to store the electricity in the battery. I will use a 8280 Watt system, which I priced out at $16,000 for reference, and base all the calculations on it generating 500 kWh per month year round (low, I know, but it's realistic for Connecticut). I now know the minimum daily generation is too low for NiFe, but this is as big as I can make the system.
Average current daily usage: 16.67 kWh per day
Current grid daily price (based on $70 for 500 kWh, $20 flat connection fee per month):$3.00
Current grid rate (without connection fee): $0.14 per kWh
Flat grid connection fee per day: $0.667
Effective grid rate: $0.18 per kWh
Sell back price: $0.14 per kWh (not sure on this, may be as low as $0.06 per kWh, this amount will be in parenthesis where applicable)
Max sell back rate with Outback: 3.6 kWh per hour
Expected system (inverter/panels/etc.) cost over 30 years: $16,000 + $6000 (replacement components) = $22,000
Average daily generation (over 5 hours): 16.67 kWh
Cost per kWh per day (over 30 years): $0.12
Average hourly usage: 0.6667 kWh
Average daily generation available for charging / selling (total generation - 5 x average hourly usage): 13.333 kWh
Using 400 Ah 48V bank as example:
This analysis shows huge loses for all battery technologies, even with huge increases in grid rates. I'm sure all of you solar experts have done or seen all of these same calculations, but I haven't seen any of this on the internet. I tried to find something like this, but I guess my Google-foo isn't strong enough.
Now let's see how a fully grid-tied system works out, using current rates. I'll be sacrificing battery backup capability with this option, so if the grid goes down, my power goes down too.
No Battery (fully grid tied):
So what I'm getting out of this is that even if I only get $0.06 per kWh I sell, and have to pay full price (plus some replacements), I still end up breaking even with what I would be paying at the current rates. The reality is that I'll be getting some solar tax credits to subsidize some of the initial cost, but price changes in the grid rates and loan interest will probably offset any initial discounts. Theoretically, if I produce a yearly average of 750 kWh per month, I could make back $5400 at the $0.06 per kWh buyback rate, which would further reduce the overall cost. Wouldn't pay for the system, of course, but it would help.
The smart move might be to install grid-tied solar, finish other updates to the house, do some landscaping, and then sell it (or rent it) for almost double my initial price. I'm guessing my total return would be about 40%, with conservative estimates. The next house would be selected to have good prospects for solar and a ground source heat pump, so that I could finally realize my dream of being mostly oil independent. Either that, or make sure there's natural gas in the area, which is generally far less expensive for heating tasks compared to electricity.
Thank you all for making me pursue this full investigation. I may create a spreadsheet that does all these calculations automatically, and I'll make it public (so people can copy it for themselves). I'll post that here if I do.
--David
What did convince me is the following (most important to me first):
- 48v NiFe banks need 1.65 v per cell to charge to their full capacity, but water starts electrolysis at around 1.52v (I think). This means that I either lose a lot of capacity to avoid losing water, or I have water them a lot (per Mike's experience)
- The chemistry of the electrolyte is a lot more temperamental than all the literature claims. CO2 causes major issues with the lithium, and frequent watering accelerates the issue.
- The literature on NiFe assures me that it's fairly cheap and easy to get new electrolyte, but all the owners of NiFe banks have a lot of issues sourcing lithium that isn't CO2 contaminated.
- Build quality issues of the cells themselves. Edison's original design was very robust and seemed to work really well, but the Chinese copy takes a lot of shortcuts and suffers longevity and performance issues as a result.
- Charging efficiency (probably due to power wasted on electrolysis) is horrible (much more than the literature would suggest).
So, I suppose the next best technology for long-ish life and safe use would be Lithium Iron Phosphate. Reasons being:
- Specs and experience seem to indicate around 10 years at a 60% DoD, and proportionately longer for lower DoD and significantly shorter for higher DoD.
- They produce tons of power without damaging lifespan much, so large surges aren't an issue.
- They have literally zero maintenance, and no accessories needed.
Their biggest flaw is the high price. Someone in another thread found a good metric for calculating a common price point: dollars per cycle. I think this can be further improved by equalizing the price per kWh for different DoDs. I think this represents the true cost of having a battery running the house at night, compared to what the grid can provide, and it makes much more sense than just throwing numbers without context around. I wish battery makers would give these numbers, but of course they don't want to showcase how horrible it actually is to run off a battery with a short lifespan. We can go even further and determine the cost to generate the power necessary to store the electricity in the battery. I will use a 8280 Watt system, which I priced out at $16,000 for reference, and base all the calculations on it generating 500 kWh per month year round (low, I know, but it's realistic for Connecticut). I now know the minimum daily generation is too low for NiFe, but this is as big as I can make the system.
Average current daily usage: 16.67 kWh per day
Current grid daily price (based on $70 for 500 kWh, $20 flat connection fee per month):$3.00
Current grid rate (without connection fee): $0.14 per kWh
Flat grid connection fee per day: $0.667
Effective grid rate: $0.18 per kWh
Sell back price: $0.14 per kWh (not sure on this, may be as low as $0.06 per kWh, this amount will be in parenthesis where applicable)
Max sell back rate with Outback: 3.6 kWh per hour
Expected system (inverter/panels/etc.) cost over 30 years: $16,000 + $6000 (replacement components) = $22,000
Average daily generation (over 5 hours): 16.67 kWh
Cost per kWh per day (over 30 years): $0.12
Average hourly usage: 0.6667 kWh
Average daily generation available for charging / selling (total generation - 5 x average hourly usage): 13.333 kWh
Using 400 Ah 48V bank as example:
- NiFe:
- Capacity @ 80% DoD: 15.36 kWh
- Price over expected life: $18,000 + 4 electrolyte changes at $900 = $21,600
- Cycles at 80% DoD: 11,000
- Price per cycle: $1.96
- Price per cycle per kWh (battery only): $0.128
- Charging efficiency: 60% (estimate, based on Mike's and other people's experience)
- kWh to charge: 25.6 kWh
- Deficit from available solar minimum: 12.27 kWh
- Effective average daily capacity (min system gen x efficiency): 8 kWh
- Cost to charge (kWh to charge x solar system kWh cost): $1.60
- True System Price per cycle: $3.56
- True System Price per cycle per kWh (price per cycle / (effective capacity + 5 x hourly usage)):$0.314 per kWh
- Grid power needed (daily usage - 5 x hourly usage - battery capacity): 5.33 kWh
- Grid power cost per cycle (at current rate without connection fee): $0.746
- Total Price per cycle (with connection fee and grid use per cycle): $4.973
- Total Price per cycle per kWh (price / daily usage): $0.298 per kWh
- Reliance on Grid for average use: 32%
- ROI: None, extra cost of $1.97 per day ($21,604 over 30 years beyond what I'm currently paying)
- Increase in Grid rates needed to make system break even: ?? I'll have to revisit this, since there are a lot of factors to determine this
- LiFePO:
- Capacity @ 60% DoD: 11.52 kWh
- Price over expected life: $16,500
- Cycles at 60% DoD: 4000
- Price per cycle: $4.125
- Price per cycle per kWh (battery only): $0.358
- Charging efficiency: 90% (conservative estimate)
- kWh to charge: 12.8 kWh
- Surplus available power: 0.533 kWh
- Cost to charge (kWh to charge x solar system kWh cost): $1.53
- True System Price per cycle: $5.661
- True System Price per cycle per kWh: $0.491 per kWh
- Surplus sell back amount: $0.075 (or $0.032)
- Grid power needed (daily usage - 5 x hourly usage - battery capacity): 1.813 kWh
- Grid power cost per cycle (at current rate without connection fee): $0.254
- Total price per cycle (+ connection fee + grid use - surplus): $6.503 (or $6.546)
- Total price per cycle per kWh (price / daily usage): $0.39 per kWh (or $0.393 per kWh)
- Reliance on Grid for average use: 10.88%
- ROI: None, extra cost of $3.50 per day ($38,325 extra if I were to buy new batteries for the same amount)
- Lead:
- Capacity @ 50% DoD: 9.6 kWh
- Price over expected life: $5000
- Cycles at 50% DoD: 2000
- Price per cycle: $2.50
- Price per cycle per kWh: $0.26
- Charging efficiency: 80% (guess, haven't researched it enough)
- kWh to charge: 12 kWh
- Surplus available power: 1.333 kWh
- Cost to charge (kWh to charge x solar system kWh cost): $2.16
- True System Price per cycle: $4.66
- True System Price per cycle per kWh: $0.360 per kWh
- Surplus sell back amount: $0.187 (or $0.080)
- Grid power needed: 3.733 kWh
- Grid power cost per cycle (at current rate without connection fee): $0.523
- Total price per cycle (true system price + connection fee + grid use - surplus): $5.663 (or $5.77)
- Total price per cycle per kWh (price / daily usage): $0.340 per kWh (or $0.346 per kWh)
- Reliance on Grid for average use: 22.4%
- ROI: None, extra cost of $2.66 per day ($29,127 extra if I were to buy new batteries for the same amount)
This analysis shows huge loses for all battery technologies, even with huge increases in grid rates. I'm sure all of you solar experts have done or seen all of these same calculations, but I haven't seen any of this on the internet. I tried to find something like this, but I guess my Google-foo isn't strong enough.
Now let's see how a fully grid-tied system works out, using current rates. I'll be sacrificing battery backup capability with this option, so if the grid goes down, my power goes down too.
No Battery (fully grid tied):
- System cost over 30 years: $12,000 (built out a new system) + $4000 (replacements) = $16,000
- Average minimum daily generation: 16.667 kWh
- Cost per kWh over system life: $0.0877 / kWh
- Average daily production time: 5 hours
- Average minimum hourly production: 3.333 kWh
- Average hourly load: 0.7 kWh
- Average hourly net production: 2.667 kWh
- Daily minimum average kWh sold: 13.333 kWh
- Daily production credit: $1.867 (or $0.80)
- Average night grid usage: 12.673 kWh
- Daily cost (night usage + connection fee): $2.441
- Total daily net: $0.574 (or $1.641)
- Total kWh cost (total daily net / usage + system cost per kWh): $0.122 (or $0.186) per kWh at current sell/buy ratio with minimum expected solar production
- ROI: Probably none, but I'll theoretically be paying less monthly
So what I'm getting out of this is that even if I only get $0.06 per kWh I sell, and have to pay full price (plus some replacements), I still end up breaking even with what I would be paying at the current rates. The reality is that I'll be getting some solar tax credits to subsidize some of the initial cost, but price changes in the grid rates and loan interest will probably offset any initial discounts. Theoretically, if I produce a yearly average of 750 kWh per month, I could make back $5400 at the $0.06 per kWh buyback rate, which would further reduce the overall cost. Wouldn't pay for the system, of course, but it would help.
The smart move might be to install grid-tied solar, finish other updates to the house, do some landscaping, and then sell it (or rent it) for almost double my initial price. I'm guessing my total return would be about 40%, with conservative estimates. The next house would be selected to have good prospects for solar and a ground source heat pump, so that I could finally realize my dream of being mostly oil independent. Either that, or make sure there's natural gas in the area, which is generally far less expensive for heating tasks compared to electricity.
Thank you all for making me pursue this full investigation. I may create a spreadsheet that does all these calculations automatically, and I'll make it public (so people can copy it for themselves). I'll post that here if I do.
--David
Comment