The Goal
This post is a first-cut attempt to determine the capacity of Li-ion batteries and the associated costs required to convert my zero-net-usage net-metered home to a home which has zero power flow to or from the grid. The cost will be calculated to achieve zero power flow from the house for the following lengths-of-time: 1 hour, 1 day, 1 week, 1 month, 3 months, and 9 months.
The House
- Latitude: 39 Degrees North
- 3000 sq.ft.
- Typical construction, 200-A electrical service.
- All-electric loads including heating and cooling using a 19-SEER, 8.3-HSPF heat pump (exception is propane cooktop).
- 12.75-kWp PV array
The Data
My power provider has recently made available hourly energy flow through the meter for the past few years. I have analyzed this data from March 1, 2016, until March 1, 2017. Net energy flow during this entire period was 112 kWh of production. The data has been analyzed so that the PEAK consumption and production have been determined for each of the periods listed above.
Here is one interesting bit of data which is not covered in the details I discuss below: Through the course of a year, my house draws (and replaces) about 10 MWh of its annual usage from the grid. That is compared with a total consumption of about 18 MWh. In other words, about 56% of its total consumption comes from the grid. The other 44% comes directly from the photovoltaics without being "stored."
A Few Assumptions
- Power flow during each hour is constant. (This is certainly false, but it allows me to proceed. This assumption also allows me to use the one-hour consumption number to approximate peak power flow.)
- No changes are made to the home except the addition of batteries. (Perhaps this is not ideal, but it is a starting point.)
- The weather during the analysis period is typical. (Perhaps, but I doubt it. In any case, what is *really* needed is a worst-case analysis, not typical.)
- I do not account for the additional PV which would be required to provide for the energy loss in the batteries. (There is plenty of room on the roof for the approximately 10% extra production which would be required. That much additional PV with inverters would cost about as much as two Enphase AC Batteries.)
- Self-consumption of the battery and their associated electronics are ignored. (This is a rather important assumption, since it is unlikely that current-technology Li-ion battery systems could store energy for up to 12 months as would be required to meet the longer-term requirements.)
Battery Requirements
For this section, I will compute the minimum number of Enphase AC Batteries which would be necessary to meet the energy and power requirements for a given time period. Here are the pertinent specifications I will use for these batteries:
- Energy which a full AC Battery can deliver: 1.08 kWh
- Maximum continuous AC power (input or output) of the AC Battery: 270 VA
- Cost for each AC Battery: US$1000
Code:
---------------------------------------------------------------------------
| Time | Peak | Peak | Limited | Number of | Cost |
| Period | Consumption | Production | By | AC Batteries | |
| | kWh | kWh | | Required | |
|--------|-------------|------------|---------|--------------|------------|
| 1 Hour | 18.7 | 10.7 | Power | 69 | $69,000 |
|--------|-------------|------------|---------|--------------|------------|
| 1 Day | 149.3 | 64.7 | Energy | 138 | $138,000 |
|--------|-------------|------------|---------|--------------|------------|
| 1 Week | 428.7 | 349.8 | Energy | 397 | $397,000 |
|--------|-------------|------------|---------|--------------|------------|
| 30 Days| 1310.1 | 917.5 | Energy | 1213 | $1,213,000 |
|--------|-------------|------------|---------|--------------|------------|
| 90 Days| 2907.0 | 1716.3 | Energy | 1588 | $1,588,000 |
|--------|-------------|------------|---------|--------------|------------|
|270 Days| 1265.0 | 2872.5 | Energy | 2873 | $2,660,000 |
---------------------------------------------------------------------------
Conclusions
I think it is clear that it is not reasonable to use Li-ion batteries to prevent the flow of electricity to or from my home onto the electricity grid. This is true even if I am just trying to handle the worst-case one-hour or one-day period that occurs through the course of a year.
Often when I refer to net metering I call it "the magic of net metering." I use this term because I know that it is absurd to try to store 3 MWh from summer until winter using existing storage technologies. What I did not fully comprehend was how difficult it would be to handle only the worst-case hour, day or week of a given year using only batteries.
So it seems that looking at worst-case numbers leads to extremely-expensive battery solutions. Perhaps a more reasonable question with the current technology would be, "How much could my grid consumption be reduced if I add X amount of storage to my system?" That is a much more difficult question to answer because it requires a form of simulation to answer. Another interesting question might be, "What can be accomplished using a combination of additional PV in addition to AC batteries?" In addition to some form of basic simulation, this will require matching up hourly data from the PV array with the net meter data to determine whether the picture can be improved significantly.
Ultimately, it seems clear that heating with electricity is a challenging task for a renewable energy source which produces the least whenever the most heat is needed. This picture gets much more challenging as the latitude increases.