You might recall the comment from my Solar PV blog post that exporting unused generated energy back to the grid is the biggest annoyance of a Solar owner.
The obvious solution to this is battery storage. I’d ummed and arghed about it for a while and finally took the plunge last month.
There are quite a few options out there in the market that I explored.
PowerVault, Tesla Powerwall and Moixa all provide the battery and inverter in one nice single box package. So you need the physical space to be able to place one of these. Not everyone will have this amount of space available, I certainly didn’t.
The other route you can go down is a separate inverter that you can then couple with modular batteries. So you can place the two elements separately.
But when it comes to modular batteries, there is only one name that stands out; Pylontech.
The most common configuration of battery that Pylontech supply are single batteries (model US2000) at 2.4kWh capacity each, which you can then string together to give the capacity you require.
After speaking to people, reading forums and generally hunting around I decided to go for the Lux AC3000 inverter. I’d heard good things about the actual unit with its 10 year warranty and also the future connectivity options which i’ll touch on shortly.
As you can see from the above, the batteries have been installed up in the loft space. The power cable and CT clamps run down the outside of the house in conduit to the consumer unit.
Who and where from?
I got the Lux inverter and Batteries from Big Green Beard, who are the sole distributor of Lux in the UK.
When you get in touch with them for a quote, ask for Jake and be sure to mention that the energy-stats website sent you. He should hopefully look after you.
My installation was done by Mark Skelton (@markskelton72 on Twitter). An electrician based out of Wakefield, but willing to travel nationwide.
How much battery to get?
The battery capacity to initially buy was a tricky choice. Did I go just one (2.4kWh), two (4.8kWh), three (7.2kWh) or even higher? In the end I decided to dip my toe in with two batteries for a capacity 4.8kWh, of which 90% is usable via the Lux inverter.
The beauty of going with a modular system like the Lux and Pylontech combo is that you can easily add batteries as and when. It only takes two short cables to add more batteries and is a simple DIY job.
Ideally I’d like to get to 7.2kWh which alongside my Solar should give me the sweetspot between sunset and sunrise. So using a combination of Solar in the day and batteries at night.
I’ve looked back though my Solar generation and electricity export history. 4.8kWh will likely get me through most of the nights in the Summer months, but its when you get into Spring/Autumn and Winter when the days are shorter then I would need 7.2kWh and maybe even more. Yes, you could spend a fortune on 9.6kWh and upwards, but for our use case I don’t think we’d make enough use of loads of batteries, not from looking at the figures anyway. But your usage may vary, so you need to decide what is going to be right for you.
Yes, there are days in winter where we don’t generate much via Solar at all. But i’d then be looking to the likes of Octopus Energy Agile tariff to fill up with overnight cheap energy and then use the stored electricity through the day, ensuring i’d got enough stored to get me through the 4pm to 7pm pricing peak of Agile.
Some days, even with batteries, you’re going to have to import some electricity! Unless you buy a 13.5kWh Tesla Powerwall system for about 8 grand!!
Lux provide an online portal from your web browser as well as both iOS and Android apps. Here’s a screen grab from my portal one sunny afternoon. The app has an almost identical look and layout.
Firstly, the Lux Inverter and Pylontech batteries are virtually silent. The only noise you can hear is the faint click of a relay switching at various times. This is a bonus as the loft space where ours is installed is right next to one of the kids bedrooms.
Below is a screen grab from my emoncms.org account. This is another bunch of monitoring I’ve got going on that i’ll talk about in another blog post at some stage. Basically, its a clamp based monitoring system based around a Raspberry Pi. Find out more about the emonPi and emonTX over at https://openenergymonitor.org/
Part of the emoncms software suite is an app that shows your historical energy usage and maps it against Octopus Agile pricing.
As you can see here from a week in early March, I’m filling the battery:
- overnight when Agile pricing is cheap (blue)
- using solar though the day (yellow)
- on March 16th I pulled a few kWh cheaply at 3pm so that I could get through the 4pm to 7pm pricing peak.
As you can see from the summary at the bottom of the screen grab, by using targeted energy imports alongside our solar, the average unit price of everything I imported that week was just 5.1p per unit.
And in this next output below from my PVOutput monitoring you can see the following:
- We imported some cheap Agile electricity into the battery at midnight to cover the early morning pre-solar period
- Solar fills the battery through the day
- The battery then covers the 4pm to 7pm pricing peak and well into the night
So on that day I only imported less than 2kWh all day. Which according to emoncms.org was at an average of just 4.7p per unit.
The orange line is the battery percentage full (using the scale on the left).
Limitation of only having two batteries
Because the batteries are modular, the charge and discharge rate depends on how many batteries you have.
|No of Batteries||Max Charge / Discharge Rate|
Note: apparently if you link 2 x Lux AC3000 inverters together you can get upto 7.2kWh in parallel.
So because I’ve got 2 batteries, I’m limited to 2.4kWh both into the battery and coming back out from the battery. You can see this clearly below from 10:40am onwards where the solar was generating over 3kWh but I was only able to charge the batteries at just over 2.4kWh (the grey bit).
The same would be true going out. If you stuck the kettle on alongside another high use item and needed say 3.3kW, you’d only be able to pull 2.4kW from the batteries, the other 900W would need to be imported (or come from Solar)
3.6kWh is the maximum limit of the Lux AC3000 inverter. Even if you had 5 batteries you’d still be limited to 3.6kWh.
This is another reason why i’d like to upgrade to 3 batteries.
Agile API Integration
One of the main reasons I went with the Lux is current and upcoming Octopus Energy API integration. I think every battery provider needs to be working on this. Powervault and Moixa are definitely working on this and there are hobbyist Tesla Powerwall integrations out there too.
I’d heard rumours from a couple of sources that Lux are working with Octopus on official API integration. How quickly that materialises, who knows. But the fall back is the fantastic Octolux by Chris Elsworth.
Octolux is a collection of scripts that need to run on a linux machine on your home network. So this could be something as cheap and accessible as a Raspberry Pi Zero W from Pimoroni. Anything that you can run install Ruby on really.
The scripts talk directly to your Lux inverter as well as the Octopus API and decisions are made to charge / discharge etc based on the rules you set as well as the current Agile pricing.
- For example you can set your required charge percentage overnight (say get to 60%) and the script will choose the cheapest 30 mins slots and charge for you.
- It will also check heading towards 4pm how much charge you’ve got. If you’re running low it will invoke a quick top up using cheaper electricity ahead of the 25p per unit peak.
- The script can also be set to charge regardless of any schedule if say the Agile price is less than 2p.
I’ve been using Octolux for a a while now and I absolutely love it. Chris has done a tremendous job with this and I heartily recommend it.
It will certainly do for me whilst we wait for any formal API integration from Lux.
Autumn 2020 Update: I wrote a new blog on the site regarding Agile Battery Integration: https://www.energy-stats.uk/lux-batteries-and-octopus-agile-integration/
How to stop your EV discharging your battery
Sadly, in most cases, batteries and EV chargers know very little about each other, especially when you’re trying to charge your EV.
Unless you do something to stop it, when you charge your EV it will take electricity from your battery first, grid second.
With the Lux setup, if you’re charging the battery, then it obviously can’t discharge. So one way to solve the problem is to always charge the batteries for the same time frame as you charge the EV.
But with differing battery sizes this is almost impossible to line up. Agile then makes things even worse as the charging can be stop/starting every half hour slot.
The easiest way to solve this on the Lux is to set a non-charging period, which Lux call Charge Priority.
In the example below, I set a small 5 minute period between 09:00 and 09:05 where the battery would not discharge.
Obviously, if you were on on Go (or similar fixed time period tariff), you could just set the charge priority to 00:30 to 04:30 for example.
It’s Agile that complicates matters.
Another way to do it would be use Octolux scripts as mentioned earlier. You could set the minimum discharge price to be slightly higher than the highest price you’re gonna charge your EV at that night.
So the max slot price on Agile might be 5p, so you’d set the minimum discharge price in Octolux to be 5.1p.
I’ve used this method quite a few times and it works well. Although it does need some manual fiddling with text files.
Anyone that has considered battery storage has likely looked at the return on their financial investment. Unfortunately, as prices stand in 2020, you are definitely in this for the long haul as battery storage is no get rich quick scheme.
- Lets say your 4.8kWh system costs you a ballpark figure of £2500 installed.
- 90% of that capacity is usable (round that to 4.3kWh).
- You fill the batteries for free every day via your solar (granted, not always possible but lets go with this). That’s 365 x 4.3kWh = 1,570kWh per year.
- So that means you’re saving whatever the cost of that 1,570kWh electricity would had been if you had imported it instead.
Lets summarise a range of electricity unit prices and yearly summaries.
- 5p x 1570 = £78.50 per year (£785 saved in 10 years)
- 7.5p x 1570 = £117.75 per year (£1177 saved in 10 years)
- 10p x 1570 = £157 per year (£1570 saved in 10 years)
- 15p x 1570 = £235 per year (£2350 saved in 10 years)
- 20p x 1570 = £314 per year (£3140 saved in 10 years)
So as you can see, to get your money back in 10 years you’d need to be saving somewhere around 15p per unit (think Agile peak pricing and higher). I’ve also not even mentioned the round trip losses of filling and discharging the batteries. ie you don’t get back all of what you put into the battery. A small percentage is lost as part of the pushing and pulling.
Don’t have solar? Well you could still ‘buy cheap’ overnight to fill the batteries and then use during the day. But then the calculations become even less attractive.
Bottom line, you’re unlikely to choose battery storage to be a money spinner. But there are other reasons you may want a battery; load shifting, buying greener/off peak energy, at lower co2 intensity times or even trying to go off-grid entirely.
Hell, you might even think stuff the economics, I just want a battery so I can feel good about my green credentials and see more graphs and stats. LOL.
That certainly does remind me of someone!! 🙂
Autumn 2020 Update
Well I succumbed to temptation. I upgraded to a third 2.4kWh battery, going from 4.8kWh to 7.2kWh in total.
With 90% of the capacity available that gives me 6.48kWh usable.
There were 3 main reasons for this
- More capacity to make better use of any autumn and winter sunshine rather than export back to the grid.
- More capacity heading into autumn and winter so that I could fill the batteries overnight when cheaper to cover a lack of Solar in the day.
- To allow those 3.6kWh charge and discharge rates.
So in this example below, I’m able to throw 2 x 30 min slots overnight at 3.6kWh charge rate.
From the Solcast solar forecast I knew there was going to be some winter sunshine, so I didn’t fully charge the batteries overnight.
The sun did its job during the day by helping top them up (7kWh generated) and we were able to ride into the evening on battery.
Looking back on the Octopus Watch app by Smarthound, my average import price that day was 8.48p per unit.
On for context, here’s the Agile pricing chart for the 1st of December.
In essence, then it is just rinse and repeat this process all winter. If there is no sun forecast the next day, fill the batteries to 100% overnight on the cheapest slots and see where that gets us in the day. Then worst case, if we approach the 4pm Agile peak and we are low on battery, give them a little top up to save us not having to pay 30 to 35p rates.
And for those that are worried about tinkering, I barely have to lift a finger, the software makes all the decisions.