@guystewart You wrote a blog post (almost 2 years ago!) about the prediction model ( Introducing Solar Production Forecast ) - great work, congrats to the team, its a really nice feature to have. It goes some way to helping customers who have not commissioned their system correctly and have no idea what “normal” or “expected production” is - but as the model does take into account panel degradation, this is somewhat moot. It will help users who have a break in their parallel string (and therefore lose half or a third production) but it won’t help users who for example have water ingress in a connector that slowly goes bad. However i digress …
You noted that;
“A future update for additional advanced VRM widgets will show the estimated production for each individual MPPT tracker.”
My question relates to constant power sources, such as an mppt controller that take power from a micro hydro turbine - In this customer’s case, its supplying 400w 24h of the day.
Does the system get messed up by the non-circadian nature of constant power supplies like this?
Once the update hits (or already has?) will constant sources be detected and then treated differently?
I imagine its very hard to predict the existence or not of a supply like a hydro where its often only after a long period of no rain that the turbine might stop, whereas solar is orders of magnitude more reliable even if overcast days make it difficult to be precise.
Also: you mention at the bottom of the article;
Limitations
Sites with these RS products may show less accurate estimates if there are other sources of solar production, or no forecast at all if they are the only source of solar production. We are working now to add full support for them in the near future.
but you don’t mention any RS products in the article. Is the inference that the data generated by RS controllers (RS 450/100 and RS 450/200) is not usable by the prediction system? That would seem unlikely. Can you elaborate on that limitation?