To maximise the solar energy, is there a facility where the capacity of the overnight charging to the battery can be automatically determined? For example, if the next day is cloudy you charge to 100%, and if sunny you charge to 50%.
Simonhewitt yes, plenty of automations available if you're comfortable with configuring software. Mine is at www.github.com/salewis38/palm
stevelewis looks complicated. Is there any guidance in how to set the system up?
Simonhewitt I suppose it all comes down to how skilled you are with coding etc and whether the effort spent automating this is more or less than the effort of doing it manually. And there is always the fun of doing the project.. In theory you should be able to control that from the web console each evening, probably no more than a couple of minutes to login and make the changes. Although, if the process is anything like that for controlling the battery discharge times (when and to where) then it could be a furstrating manaual process as things may not work as you might expect them to.
Simonhewitt in the notes on the same page. It's fairly straightforward if you have experience with a Raspberry Pi.
Maxwell Yep! I had automations set up but frequently went in manually to override them based on my own experience at my location. In the end I now do it all manaully. A couple of minutes once or twice a day depending on the stability of the weather. One problem we all face in the UK is that barring those times when there's a big high pressure area sitting over us or scandinavia our weather is so changable and forecasts unreliable. I use a combination of BBC local weather (farming forecast is best), Solcast and my local knowledge to set a charge % target. I usually review this just before going to bed to see if I want to tweak it. Mostly, I am tweaking during the winter months when generation hours are short. A summer comes along tweaking is only needed during dull and rainy periods as most days will have enough generation to not charge overnight at all. So, overall, I found automation a fun task and a great learning exercise but it ignores local weather patterns which can have huge impact on regional forecasts.
Home Assistant?
cluelesscris I've been a long term user of Accuweather and feel it isn't strong enough for cloud data, so I have been trying a few alternatives. The two I'm currently using with some success (as standalone apps, not APIs) are 'Clear Outside' and 'Windy'. The former is designed for astronomers and gives a week ahead cloud data for a nominated location which you'll need to select each time you run it. Windy is newer to me but the 'Display as 'Meteogram is very close to helping cloud prediction and the intent is to see if this can help predictions in the coming weeks. Both apps are on iOS and Android and give cloud predictions at three levels.
Finally, I have got Home Assistant running happily on a Raspberry Pi2, it's gathering data from Giv and Myenergi sensors; the go-to weather prediction seems to be from Solcast's API but I've not gone down this route yet without more understanding of forecast and actual effect 🌥️
Thank you for all your helpful advice, it is much appreciated.
I use HomeAssistant with Solcast solar forecast and everyday at midnight AC battery charge target is set based on the solar forecast for the next day and few other factors.
This works great most of the time - I am referring here to the GE battery SOC issue that is adding the unpredictability factor to the whole equation - but my formula assumes that daily battery charge loss due to that problem is 1kWh unless the battery is discharged to 0%