Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Optimising Solar PV / PowerWall to Use instead of Export

This site may earn commission on affiliate links.

WannabeOwner

Well-Known Member
Nov 2, 2015
9,170
5,337
Suffolk, UK
I've added 32 panels to the 16 I already had and I'm now in the (new for me) position where I am generating more Solar PV than I can use, and PowerWall is filling up quiet early in the day.

Whatever back-of-envelope sums I did, which caused me to think that what I have installed wouldn't be enough, may have been wrong.

So looks like I need to figure out how to use up any excess.

I have added a Zappi, so I could just let that dump to car ... but I'm wondering if I could be more ambitious than that. I'd be interested to hear what folk have cobbled together via API etc.

My thought is:

I need a bit in the PowerWall to cover until the sun is up. I'm (currently) on E7, so just from 7AM until sufficient sun available to charge PowerWall. This morning that started at 06:40. In shoulder months that might be later, and on days which are overcast etc.

So some sort of weather-prediction for how good the sun will be tomorrow, and whether usage will be abnormally high (my scenarios for that are "Baking cookies" or perhaps a setting for "House full, Party mode"). I was thinking that I could raise the Reserve Percentage at the start of Off Peak, and then drop it to 10% when Peak starts

Then my thinking was, rather than let the Zappi react to export, and have to wait until enough spare AMPs to make it worth charging the car, how about charge the car when PowerWall is at, say, 95% and stop when PowerWall falls to 90%. Charge the car at e.g. 5AMPs, if the PowerWall gets to 96% then increase that to 10AMPs and so on. (Obviously once I have some better data then choose a suitable default rate).

Change the lower limit once the end-of-day is approaching - so, ideally, at end of day the PowerWall will be filling up towards 100% rather than being at 90% when the sun sets.

Usually the PowerWall at 100% at 18:00 provides enough to run the house until midnight, when I would then have E7 rate available.

Although I'm starting to wonder if I need a 3rd PowerWall (several servers here so background rate is typically 1.5 - 2.0 kW, and 2x PowerWall won't run the house all day from an overnight off-peak 100% charge)

Panels = 48
PowerWall x2
Tesla Wall Charger
Zappi

Raspberry PI lying about (I'm a database programmer by trade, but very little time, I'd prefer to buy an APP that I can use / adapt. But happy to process whatever data is available, less keen on API programming because of the debugging time that it will need)
 
Your best bet for off the shelf would be something like Home Assistant and/or Node Red, both can be used for automating tasks based on events within the home. However they don't natively support weather prediction, just reporting what the forecast is (although Node Red can run ML inference models).

I'm planning to write my own ML script to automate charge levels in my PylonTech system based on forecasted PV production, but I was planning to leverage an irradiance meter as a measurement reference against a forecast. Probably wouldn't suit your usecase, and unfortunately will be at least a couple of months before my install and I can start writing it!
 
  • Like
Reactions: WannabeOwner
Not quite the same, but I use a RPi to switch on a Tapo plug connected to an electric heater when the export is >800W or so. Provides some nice local heating in the “shoulder” periods of the day / year. It cost me about £60 to setup but I already had an RPi and EcoEye monitor.

Probably won’t be useful for your large PV + Powerwall setup, but might be of interest to other readers of the thread. :)
 
  • Like
Reactions: WannabeOwner