Take a look at our Powerwall charge level over the last couple of days.

In winter our Powerwall is set in ‘Time-based control’ mode so it charges up overnight on cheap electricity that we then use during the day – saving us money.

We reserve 20% of the useable battery capacity for use during grid outages so if the battery charges fully overnight that’s 80% * 13.5 kWh of energy from the battery which is 10.8 kWh.

On our current Octopus Go tariff that saves us 10.8 x (30.78p day – 8.25p night) – a potential £2.43 a day – ignoring battery conversion losses for now – need to investigate this in another post.

So far so good. Look at the 24th November – the battery starts charging around 1:30 am and charges to 100% – it starts discharging at 6:30am when day rate electricity starts and by around 19:00 it has discharged down to the reserve level. (A bit of Tesla madness means this 20% reserve shows as 24% on the above graph). The two steeper sections of the discharge curve are when the heat pump is running.

Now look at 25th November – the battery only charges to around 70% overnight – not using 30% of its capacity so less battery energy will be available today. What’s going on, is it broken? Look at the graph – the battery starts to discharge when the heat pump kicks in but unlike on the 24th the sun comes out – the heat pump turns off earlier as the sun warms our main room and the battery starts to charge reaching 100% just after 1pm. Around 3pm the sun has gone and the battery starts discharging. The graph below shows the solar production over the same two days.

Around 2kWh on the 24th, over 10 kWh on the 25th. The Powerwall must be using a solar forecast to control how much it charges over night – clever. Doing so saves a small bit of money – around 40p in this case but, hey, every little bit helps.

How much the Powerwall charges overnight is completely automatic – there is no user control. When it works as well as this who can argue…

…We can – can’t help feeling it’s being a bit too clever. The savings when it gets it right are tiny – the 40p mentioned above doesn’t take into account the additional export that would happen if the battery was fully charged, taking export into consideration means the saving is less than 30p. However if it gets it wrong the costs mount very quickly – if the sun hadn’t come out that 30% left uncharged would have cost us £1 importing day rate electricity. That risk reward balance means it needs to be right three times out of four.

It’s a difficult thing the Powerwall is doing – it has to cope with lots of different situations across the year – and it certainly got it right on 24th and 25th November. If it has a bad day expect another post!