Earth Notes: Bulk Electricity Storage @Home (2011)

Should you consider serious electricity storage at the household level, behind the meter?
example 8-day UK electricity demand graph for 2007/10/11 c/o National Grid

2011: Can it be done? Should it be done? Why not just build more microgeneration or keep using the grid?

What energy do we already store at home?

Those of us with laptops and torches are storing energy for a few hours' use. And early adopters' electric vehicles contain substantial storage too.

Almost all of us have a fridge or a freezer that stores 'cool' for at least a few hours: if you turn the power off the food does not immediately warm up and spoil.

Those of us with hot-water tanks or night-storage heaters can store heat that way for maybe a day, quite cheaply.

Households with off-grid central/space heating such as wood or a fossil fuel such as oil or LPG typically have substantial storage for many weeks, or even a year or two with wood stacked and drying... 'Seasonal' thermal stores of heat can last the winter.

But storing substatial electricity in a form that (say) a washing machine or dishwasher or fridge-freezer could use is still hugely expensive. For example, a typical UK household uses about 10kWh (10 'units') per day of grid electricity; storing that in Lithium batteries similar to a laptop might cost £10,000 ignoring the electronics to get it in and out.

Reasons To Be Storing Joules: Here's 3

  1. Managing intermittency and demand.
  2. Daily load-shaping and cost minimisation.
  3. Off-grid and seasonal storage.

There are various reasons to store energy, from helping the grid manage intermittent renewable sources such as wind, all the way to being off-grid thus needing to hold energy from local sources, usually to cover a few days' poor generation, but maybe also to carry energy from summer to winter. I shall investigate some of these below, but note right up front that electricity is the most expensive form in which to store energy by far, compared to heat or fuel (eg: wood, oil).

Little House on the Cemetery

To keep this thought experiment concrete I'll consider both our house circa 2010 and a typical UK household (heated with mains gas rather than electricity). For simplicity I assume the system to have low-enough round-trip losses (getting energy in and out of the batteries) to generally ignore, less than 10%.

As of the end of 2010 we annually (based on this year's figures mainly):

and we are roughly zero-carbon overall, and if our use of gas for heating (with a heat-pump) and cooking was switched to electricity, we could drop the gas connection and still just about be neutral or an exporter of electricity.

But we still need to be connected to the grids for now. In particular, even in summer when our PV generates four or five times the electricity that we consume each day, we must import at night and even during parts of the day when consmption transiently exceeds generation.

Managing intermittency and demand

Some industrial customers get cheaper electricity in return for being disconnected from the grid with little or no notice when there is a shortage of generation (eg when a generator goes off-line unexpectedly). As of 2010 there's something like £100k/year available for every 1MW of load that can be removed automatically within a fraction of a second, for which good candidates can be air-con and cold-store refrigeration. There are different rewards for changing load more slowly (lower-slew rate) or for simply scheduling in 'off-peak' hours. This is called demand control or demand management.

As more intermittent generation is added to the grid, where we cannot choose when it generates (it's not 'callable') we can either back it up with conventional 'thermal' gas plant or we can add more demand management. And the times when demand and generation don't match are likely to become less predictable than always relying on excess in the wee hours of the morning.

The 'smart grid' will likely bring some of this to retail customers (there are already off-peak tariffs such as Economy 7 in the UK, and considerably more complex arrangements elsewhere such as in France). Shaping the load dynamically to fit available generation could happen through price signals and/or storage on the grid or at/near customer premises.

Storage near the customer already happens in some places where the local grid is weak and has trouble coping with peak demand and possibly even with peak generation where there's lots of local microgeneration installed. Local storage can reduce the need for building of new transmission and distribution (ie wires in the air and ground).

Daily load-shaping and cost minimisation

From about March to October we generate more electricity than we consume, so a battery system of capacity no more than about one day's consumption (5kWh) would mean that we would usually never need import at all outside winter, and could export some extra energy when the grid carbon intensity is highest to maximise our dent in the UK's carbon footprint.

For a typical household with no microgeneration, once movable loads such as washing machine and dishwasher were moved to off-peak (or lower-carbon-intensity) times, a battery pack of (say) the remaining 8kWh could move the rest of their demand. For example, for Economy 7 rates at 1/3rd of standard rates, this could save 2/3rds of the bill or maybe £400 at current rates. Shifting the load to the lowest-carbon times could save on average about 30% of generation CO2 emissions minus the losses in the battery system.

The daily times of these cost or carbon savings may not stay fixed as more intermittent generation such as wind enters the grid in future. However, times of low-intensity and low cost are likely to remain together (ie well-correlated) and savings might be claimed by the retail customer or their supplier as aggregator.

Off-grid and seasonal storage

If we were to be off the electricity grid entirely (but still on the gas grid) we would need to ship excess from summer to winter in a huge battery, or put up enough extra solar PV to generate enough every day that a smallish battery (5 days to a week or so of very gloomy/snowy weather) would do.

Our current 5+kWp array (east/west facing) is only generating ~1kWp per day and none at all when the panels on the roof are covered in snow. Our consumption is ~6kWh/d at the moment (some goes to run the gas heating) so we'd need to generate on average an extra 5kWh/day which implies 5kWp (~40m^2, and ~£20k) of south-facing panels plus maybe storage of 6kWh/d for at least 7 days, ie over 40kWh of storage at about £40k in Lithium batteries) so getting on for 10 times the minimial load-shifting scenario. And even then it is possible to have no decent subshine for longer than a week. Note also that without a grid connection we'd be wasting (not using) the vast proportion of the sunshine falling on our panels in summer. If you think that the £50k bill that this implies is big...

If we we to try to shift the electricity from the summer to cover the ~90 days at 6kWh/day minus an average of say 2--3kWh/d generated on average, then it might be sufficient to carry forward ~300kWh of energy about 3--6 months and we would have plenty of excess already so could tolerate 50% or more self-discharge over that time if necessary. In Lithium-chemisty batteries that might cost £300k, but in Lead-acid (~3%/month self discharge) that might cost a mere £100k. Vanadium redox flow batteries and molten-salt batteries such as the Zebra and GE's Durathon are also interesting here since when in the inactive form (eg flow reagents removed, or cold/solid) there is essentially no self-discharge and the amount of battery brought on-line at once need only cover average daily power draw and consumption, ie these can be activated incrementally as required.

If we were also to shift our heating and cooking from gas to heat-pump and assume for simplicity that (say) ~2kWh/d cooking stays as is and the rest is converted from electricity to heat with a CoP of ~2.5, then total heat demand would be about 2800kWh (cf 5800kWh gas currently), of which about 900kWh (cf 1500kWh gas, 500kWh cooking, 1000kWh DHW) would be used in the 8 months outside winter that we'd be then above break-even on electricity generation, and the remaining 1900kWh would need to be carried over. The cool £2m price-tag is clearly ridiculous, and points to the fact that it would be better to store the heat as heat in some sort of thermal store.

Suppose to further stretch this idea we wanted to put up enough solar PV to generate enough to cover heat demand (peaking at ~40kWh/d, thus maybe 16kWh electricity) mid-winter we'd need to get out 'extra' solar up to maybe 20kWp (£80k) with truely vast excess in summer (and we don't have 160m^2 south-facing!), and the extra battery to support 7 days at over 20kWp is another £140k in Lithium chemistry or maybe £50k in lead-acid.

Note however that we'd still be nominally generating more than we use in a year, though system losses could be expected to make that moot.

It's already difficult to capture enough heat as heat with solar thermal to contribute much to space heating mid-winter in northern climes.

For these sums it would be better to (re)build a house to PassivHaus standards and not require any formal heating system at all and skip the bill for storing such enormous amounts of energy round the seasons. If it's hard to see how to do it for one house, it's not going to be easy doing it at a national level if people won't even wear a sweater and turn the heating down!

Note that a typical UK home uses twice the electricity and three times the gas/heat that we do, so inflate all the numbers appropriately, or at least think long and hard about retrofitting the best energy efficency possible into the existing housing stock.


  • 2011/02/28: Will Distributed Solar Drive Utilities into Bankruptcy?: bulk energy storage and numbers.
  • On 2011/01/10 400Ah@12V of lead-acid gel cost me £900 including VAT and delivery, ie close to £400 per usable kWh of storage, and at 120kg gives an effective energy density of ~16Wh/kg.
  • GE Durathon "NaMx" NaClFeNi battery tech. A type of "molten salt" battery.
  • Prudent Energy Vanadium Redox flow battery tech.
  • Sealed lead-acid (SLA) self-discharge of 1--3% per month dependent on type.
  • 2010/09/16: From the Roof, to a Battery, to the Grid.
  • US Federal Energy Regulatory Commission (FERC) June 2009 study "finds the potential for peak electricity demand reductions across the country is between 38 gigawatts (GW) and 188 GW, up to 20 percent of national peak demand, depending on how extensively demand response is applied. This can reduce the need to operate hundreds of power plants during peak times."
  • As of 2008/2009 the UK grid's balancing costs are of the order of £100k/MW/year, ie £1 per year for every 10W of firm-frequency-response deferrable load and that load on all year would cost maybe £13 on a typical domestic tariff (assume 15p/kWh).
  • GE Reinvents the Fridge (Treehugger 2008/10/30). VP Kevin Nolan explains: "Peak hour energy demand is growing faster than total energy demand. It is imperative that we begin to shift some of the energy load from peak hours to other parts of the day -- helping to avoid the need to build new power plants to meet the demand."
  • Toronto Hydro's peaksaver scheme to peak-shave a/c and heating load for a one-off CAD50 incentive.
  • TheWattSpot: ComEd (Commonwealth Edison) Residential Real-Time Pricing and automatic moderation of aircon demand in Illinois, US.
  • We Energies (Wisconsin, USA) Energy Partners programme gives up to a $50 credit each summer if you allow the utility to interrupt your air conditioning between noon and 11 pm, May 15 to September 15 each year, if necessary. ($50 for up to six (6) continuous hours in a single day, $40 for up to four (4) continuous hours in a single day, $12 for up to 45 minutes off with 15 minutes on for up to eight (8) hours in a single day.)