Earth Notes: Bulk Electricity Storage @Home (2018)

How to get electricity storage behind the meter at home in 2018 without too much pain.

While I haven't been looking, suddenly a whole bunch of interesting AC-coupled behind-the-meter, PV storage options seems to have arrived in mid 2018.

The device that initially piqued my curiosity this time was the Enphase AC Battery which looked as if it might be as little as £1k to get 1kWh of storage, in a very small box that I could tuck beside my consumer unit and have deal with trickle/constant loads such as the fridge overnight. (Turns out to be more like £3k, installed, with a supporting control/meter box, but still, very interesting. Note that a Tesla Powerwall 2 might only cost a little over double, but with over ten times the storage and over twenty times the power capability.)

I've fired off a number of requests to potential suppliers, of the form:

Current PV system: 5.2kWp grid-tie system over 4 x SunnyBoy SMA inverters/strings, 3 strings on one meter/tariff, 1 on a newer tariff. Import, export and 2x generation meters installed.

(Also, separate off-grid system, ~700Wp with 2kWh lead-acid storage.)


A/c coupled storage system to increase self-consumption by:

  • covering night loads to avoid night-time grid draw (~70W/1kWh) spring/summer/autumn
  • and/or to cover up to 2kW to 3kW of peak load to avoid any grid draw during day during all but winter months.

Note: typical gross consumption per day 6kWh of which ~25% is covered direct from the PV across a year.

Would very much like to monitor (via WiFi / MODBUS or similar) and/or have the opportunity to manage from existing control hub to control overall charge/discharge regime. Would prefer all control to be local, ie no dependency on live Internet connection/account.

Please give me an indication of cost and space requirements, and lead time.

The lead time for an installation seems to be nearer months than weeks!

Motivation and Discussion

I think it's worth being clear about my motivation for this.

I'm not doing this to save money. I'm doing this partially experimentally and mainly as a (small, expensive) carbon-reduction exercise, reducing grid losses/strain shipping energy in and out of the house.

I want to develop ideas about how to manage storage in an efficient (carbon-cutting) way, and this would give me scope to tinker. Fiddling with my off-grid RPi system power management C++ code has been educational, for example.

Note that storage used in this way is absorbing some of the intermittency (spikiness) of demand as well as of PV generation. Normally the grid does a good job of evening out very short-term demand spikes by being a grid, summing demands of many millions of consumers, but those back-and-forth flows do have losses.

I kicked off a Green Building Forum thread on this topic...

Below are some selected posts from the thread.



The market has moved on, and though I'd still be paying more in raw storage terms than for my lead-acid gel off-grid battery storage of that time, this is packaged, installed, with data access, etc. And done right meets G83 and G100 and shouldn't upset my current FiT arrangements.

I'm not doing this to save money. I'm doing this partially experimentally and mainly as a (small, expensive) carbon-reduction exercise, reducing grid losses/strain shipping energy in and out of the house.

There may be a bigger store added in parallel with this later, even up to seasonal store.

... If the council doesn't pull my house down under CPO in the interim.


Damon, wouldn't it be better for the environment if you exported the solar energy and displaced some fossil generation?

The premise of the FIT and export tariff is that those of us without solar, should subsidise those of you with solar to export it, to everyone's net benefit. Grid tied storage feels like that bargain is broken, though not as bad as immersion diverters.

What does the carbon intensity data on this say? Haven't checked for a while.


It *is* best to displace ff generation on the grid with solar.

This doesn't change the amount of that displacement that happens (ie net exports are the same), it simply stops some of the back-and-forth across the local distribution system which is lossy and puts a strain on the local grid. This can also shave some peak-time loads by shifting when residual load is presented to the grid.

The losses in grid vs the battery are probably a wash, even with an efficient lithium-chemistry system.

But avoiding using the grid as if it were a battery is a good thing. Shaving peak demand is a good thing, and almost certainly a (carbon) saving one way or another IMHO.


Hi Damon, shouldn't people export solar during the daytime peak when grid intensity is high, then import at night when intensity is lower? You have written extensively on this.

Do people still claim Export tariff at deemed 50% of generation, even though they are not exporting anything? Paid for by me! but also by the folk on the sharp end of the energy poverty wedge. This encourages people to dump publicly-subsidised solar power into immersion heaters.


So, I would still (in winter, say) be running my dishwasher overnight and importing the bulk of that then. (I might only be able to supply ~25% of peak ~2kW demand from local storage.)

The question of what actually fires up to deal with an extra 1kW of grid demand, eg because someone stopped spilling PV generation to grid, remains open. It's probably CCGT most of the time, but when there is no pumped storage being released to grid I take it that shifted grid demand at that time is effectively extra grid storage. But also look at:

I'm still in several minds about diverting PV to resistive heating for many of the reasons you outline and that I have raged on before. But if that storage is efficient (eg such as Sunamp's v low losses compared to an actual hot water tank) or you are relieving the grid of peak flows around solar noon, then it may be less bad though to actually good. (If that heat storage were via a heat-pump preserving exergy, then I see that as being likely much better.)

But generally doing whatever is possible to reduce peak demand (and flows) by importing less or exporting more is good. A flat grid demand (or matching intermittents' availability) is a cheaper and lower-carbon one, I think.



Damon, you mentioned storing solar power to run dishwasher and washing machine overnight. Would you get the same effect if you just run them during the daytime solar generating period? maybe develop some software using the weather forecast to decide when the optimum time would be, interface the delayed start function of the appliance?

You also mentioned storing power to run the fridge/freezer overnight. Would you get the same effect by chilling them down during the day then switching them off overnight, maybe with some freezer blocks or bottles of salty water to act as stores of cold? Use a time switch or a pv diverter to switch them back on during generation times?


When it's sunny I do try to run the dishwasher during the day.

However if it's at all cloudy it can be hit and miss. If the generation is not >> 2kW for the entire first hour or so then the couple of spikes of consumption while heating water may not be fully covered by generation. Batteries could partly or fully fix that.

Also, when it is sunny, I do attempt to take the opportunity to load up the fridge with beer or wine to store cool, yes, but that's going to be small fry. And I'm not going to risk damaging the fridge or our health messing about with turning them of or off. I already bought the most efficient I could that fitted our requirements:

The washing machine timing is more tricky because it has to fit with drying days, etc, but we tend to run it to finish when we get up, which while not minimum grid demand and CO2, is better than peak. And we run very few hot washes; a fair percentage are run cold:

PS. I used to have my system automatically start my old dishwasher at a 'optimum' time by turning on the power. But the new one with electronic controls starts in a silly state sometimes and that is not possible. Something like Green Bean would be lovely, but I don't have access to any such thing.

Thread continues...

See also a Fieldlines thread on this topic.

FiT for Purpose

My FiT administrator says that I should be aware of recent Ofgem guidance and "whether the installation of the battery effects the generation meter."

One point from the Ofgem document is that "For the purposes of the RO and FIT schemes, we consider that in most cases a co- located storage facility would not be considered part of the RO generating station or the FIT installation. This is because storage is not directly referenced as an eligible generating technology under either of the RO or FIT schemes, and in most cases, the storage facility will not be essential to the operation of the generating station or installation."

Also "If the storage facility can only be charged by the FIT installation or the storage facility is installed after the generation meter, then it would be clear that the generation meter only measures generation from the accredited FIT installation." I am proposing to have the storage "after" the generation meter, connected to the consumer unit.

However, by default my export meter would be between the storage and the grid so if the storage were to cause export to the grid that export would not be eligible for FiT export payments. So either I could attach the storage between the (inner, house-side) export meter than the (outer, grid-side) import meter, causing any exports from the storage to be disregarded. Or maybe I'd be permitted to switch from metered exports to 'deemed' (50% of generation), but retaining the meter for my own records.

2018/07/20: The latest communication from my FiT administrator, after sharing with them a 'standard' (Enphase AC Battery) connection schematic into the consumer unit, was reassuring:

Providing the battery is between the Generation meter and Export capable meter as shown in the schematic you've provided, this should not affect your current Feed-in Tariff. As your generation meter will still clock the true amount of generation with the battery downstream of it, and the export meter will still clock true export with the battery upstream of it.


A sizing suggestion, ie rule of thumb, made to me by a potential supplier, was that for each 1kWh/day of gross consumption, there should be 2kWp of PV and 2kWh of storage. For us that would mean doubling our PV size and ~12kWh of storage (roughly a Powerwall 2), but interesting in any case!

A typical night may have one or both of the dishwasher and washing machine running, the former just under 1kWh and the latter around 1kWh if not a cold wash else much lower. Then the fridge/freezer (~1kWh/d, so 0.3 to 0.6kWh) and a few tens of watts extra 'vampires'.

The Enphase (~1kWh/260W) would cover fridge/freezer plus constant small loads, but not the bigger overnight loads above, currently scheduled then in part to be kind to the grid already.

A meatier system with ~2.5kW or greater power capacity could support the dishwasher or washing machine (~2kW peak while heating water), plus the loads that the Enphase could carry. ~2.5kWh energy capacity would keep them off the grid. Maybe another 0.5kWh would support lights and TV and some hot drinks while the sun is down.

Note that our kettle is 3kW, so 3kWh/3kW could potentially cover most incidental loads outside winter when the sun is down.

YouGen suggests making sure that the power output is at least 3.6kW.

In any case, continuing to schedule kitchen loads overnight means that when/if the storage runs out then residual demands on the grid are at a relatively good time.

For covering (low) night loads the trigger import load to start generating will need to be checked. If too high then the night loads simply won't be satisfied from the storage.

The biggest single always-present night-time load is our fridge/freezer. Note from a .csv spreadsheet/data of 1 minute samples that the load mean was 42W (1.02kWh/d), max 356W, typically 0W or 75W. So triggering reliably at/under 75W would seem to be key. Our TV and cable box together come in at ~40W, which might typically be accompanied by 20W of lighting after sun-down. To cover that, a reliable trigger at/under 60W would be needed.

Brands Considered

Brands/products considered so far, at least to some degree:

Enphase AC Battery
May be able to inject energy precisely enough to prevent imports as low as 5W. Not obvious if this will work fully given that we have two separate solar generation meters/feeds and the Envoy S seems to only monitor one. Not obvious if data downloads to my server can be automated. Round-trip cell efficiency 96%. Round-trip inverter max efficiency 97%. 1.2kWh battery unit size/spacing 45x31x22cm with 30cm above and below.
Sample UK vendor 1 £4495 2.4kWh including fitting, commissioning, VAT 2018/07/13.)
Sample UK vendor 2 £3937 2.4kWh kit including VAT + UK mainland delivery 2018/07/18.)
Tesla Powerwall 2
1150 mm x 755 mm x 155 mm. Round-trip efficiency > 90%. 13.2kWh capacity. 3.68kW charge/discharge power.
OutBack GridZero (or Mini Grid) mode inverter chargers
GridZero import trigger load seems to be 240W/1A.
Powerflow Sundial M2/S2
Master and slave 2kWh nominal (1.6kW actual), max output power ~500W. Size ~44/30/28cm. Minimum trigger power on import is 170W (confirmed in email by manufacturer), which would fail to engage for our typical night loads.
Powervault 3
120x98x25cm for 4kWh and 8kWh units; 20cm space needed for airflow. 3.3kW max charge rate for both. 4.6kW max continuous power/discharge for 4kWh unit, 5.5kW for 8kWh.
Sonnen eco 8.0 and 9.43
The 2kW (8/2) version is the wall-mounted unit alone, which I could accommodate, but everything else is huge and I couldn't. Would stuggle to cover my 2kW/3kW peak loads for that use case. Trigger limit 30W (and same for eco 9; email from sonnen UK).
The eco 9.43/2,5 capacity 2.5kWh (and /5 is 5kWh); DoD 90%. Charge/discharge power 1.1kW and 2.5kW. Cabinet 1 size 88/67/23cm. Trigger limit 30W (as for eco 8; email from sonnen UK).
Tesla Powerwall 2
1150 mm x 755 mm x 155 mm. Round-trip efficiency > 90%. 13.2kWh capacity. 3.68kW charge/discharge power.
Victron ECOmulti
The marine blue may prove too industrial a look. Have previously considered Victron in other configurations to achieve a similar effect, but tied to lead-acid storage.

Note this helpful insight from Matei Predescu of

[Re Powerflow] ... I've suggested it because, like the others, it's one of the few AC batteries out there. Most of these battery companies are relatively new and not that different, since the battery technology inside is mostly the same and comes from the same few companies (Samsung, LG, etc). The only thing differentiating them would be the BMS and how well it was thought out, since it is responsible for basically keeping the battery as healthy as possible, for as long as possible.

2018/07/18: Deposit Down

Excitement: I have today put down the deposit on a modest system. I should have a shiny new storage testbed within a month of starting the search for storage this time around. Hurrah!

Oh, and yet more data, live and slightly-less live. In particular accurate import and export and thus gross consumption data. I may get away with less manual meter reading in future! Indeed, with gas, import and export and generation all electronic I may no longer need to read 6 meters each day.