Is a home battery and solar system worthwhile
now?
So I took the solar data from my location which I got from Solcast API (
https://solcast.com/). I then fit this against a reasonably pessimistic model of a 4kWp solar system on an ideal roof orientation. The system is modelled to generate 3,970kWh per year which is about average.
The Python script takes the insolation data, calculates solar generated in half-hourly intervals, and models battery charge and discharge. If there's a shortfall in the first iteration (it appears that there is insufficient solar for that day) it adds additional charging during an off peak period (currently set to 2:30 am to 6:30 am, though it could vary.) This is a little inefficient as it should aim to top the battery up in full by the end of the off peak, probably by varying charge current.
This is a very simplistic algorithm, I haven't attempted to add in any kind of iterative solution solver to improve upon the "first iteration". I'm reasonably sure that if you know, or can estimate, the sun and cloud cover for tomorrow, and have a model of daily load and know the starting state, that it's possible to perfectly calculate charging times needed. For now, I make a very simplistic optimisation: if the projected solar energy for that day is insufficient, I charge a number of 30 minute slots in that off peak period. The reality is that there are a few days where the energy from the day before is actually enough that this charging is not necessary. But it is better to over-estimate expenditure than to under-estimate it.
The results are very interesting (to me, at least!) because they justify such a system rather easily.
Assumptions:- Sun is similar to 2021-2022 season in any given year
- Panel efficiency 20%, panel area 20m^2
- On peak cost of 60p/kWh, off peak cost of 15p/kWh (assuming the 'Octopus Go' sweetheart rate does not remain forever at 7.5p/kWh)
- Off peak charging time available is 4 hours (charge rate ~3kW)
- Daily usage of 12kWh electricity excluding EV (charged in off peak period) and no heat pump
- Daily usage is assuming timed high loads can be shifted into off peak (e.g. dishwasher, tumble dryer) - in other words, it is the "unshiftable" demand
- Battery size 9.5kWh (GivEnergy LiFePO4) with charge efficiency 91%
- Battery does not require heater in winter (mount in insulated garage >5C year round)
- The 5kW output of the battery is enough to cover ~98% of usage and therefore I don't need to account for forced grid use here
Pre-system cost is
£2,591 Post-system cost is
£273 Yes, it saves nearly £2,300 per year for an ordinary user - or about the same as the price cap will rise by this year for both electricity and gas. If the system costs £10k, then it pays back within 4.5 years.
In addition to the 'savings', it contributes 911kWh of energy into the EV battery in the case when there is too much sun that day to go into the home battery. (This assumes that all of this can be used. Probably it could also be diverted into the hot water tank or something else if needed.) That's around 3,000 miles of 'emissions and cost free' driving.
If I change the parameters somewhat:
- Off peak 30p/kWh, on peak 90p/kWh
Pre-system cost is
£3,917 Post-system cost is
£535Saving
£3,382Reflecting a short term high price but still a low off peak price.
A payback time of
under 3 years.
However, if the off peak price rises considerably, say:
- Off peak 60p/kWh, on peak 90p/kWh
Pre-system cost is
£4,040 Post-system cost is
£1,036Saving
£3,004Surprised me that the benefits are still there. It's really all down to the solar. The off-peak is nice, but the solar is *great*.
Now let's assume things return to 'normal':
- Off peak 6p/kWh, on peak 17p/kWh
Pre-system cost is
£699 Post-system cost is
£106Saving
£593As can be seen, the savings vanish if the on peak price is low. The payback time is going over 15 years in this case, well beyond the warranty of the battery. Perhaps it would make more sense if there was a heat pump, with high daily consumption, though extra batteries would likely be required. The pre-cost does not include any benefit in the summer from EV charging which may change the maths a little bit, but probably not substantially.
I have attached the script in Python if anyone is interested in playing around with it. The solar data can be fetched from Solcast, you will need to sign up for a trial account and input your address to get solar data.
Now to convince the other half to spend £10k on this kit ... that may be the biggest challenge!
Attached outputs are for the 60p/15p scenario.