Forecast.solar further than tomorrow e.g. 5 day forecast?

I’m a relative novice, trying to keep as much in the UI (rather than yaml) as possible. I am successfully using forecast.solar with 2 solar arrays and find it accurate. I’m aware the api will allow provision of a forecast up to a week ahead and I was looking for a simple (?magic;) way of accessing this data so I can have, say, the next 5 days energy production set out clearly on a dashboard to plan when to have a washing day etc. Does anyone have a template or anything that they have already created to do this? Many thanks, Phil, South Cheshire, UK.

Sorry - no magic available for this one.

The integration for forecast.solar calls the API based on the ‘free’ account, hourly, and pulls down the data for today and tomorrow.
First - this data is not stored anywhere within HA. It is used to update the integration entities, and the energy graph forecast dotted line plot, and that is all. Integration entity data is available from history, but that just tells you what the forecast for tomorrow was yesterday.
Second - the free account provides data for today and tomorrow only.

If you want more than +1 day you need a ‘paid for’ subscription. The professional level account apparently provides +6 days forecast. I don’t believe however that buying such a subscription will make any difference to the HA integration in that in HA you will still just get just today and tomorrow.

If you want the data for your own purposes, then you need to call the API directly and process/store the data yourself. I use Node-RED to call the API (free account) and maintain a graph for yesterday (history) today and tomorrow, plotted with actual PV generation.

If I want to plan a washing day, the HA weather integration is more than adequate to provide a broad brush idea of a good versus poor solar generation day. Strangely though, here in the UK, the weather is always going to be sunny in three days time, never tomorrow.

1 Like

Thanks so much for that very thorough answer to my question - although I did suspect this would be the case. I’ll have to work out a clearer screen using open weather or HA native weather to plan the domestic week. Will have to remember your “third day rule”!

I’m interested in your lux meter as I’ve looked for an LDR to measure high sun levels but they seem to hit max at relatively low levels and I have to cover them with an envelope or two to dull their exposure if am wanting to recognise strong direct sunlight (5506 and 5547 LDR’s). I’ve decided that interrogating solar pv output versus position of sun is probably the most accurate way of quantifying bright direct sun levels. Do you have any “bright sun” LDR’s / photoresistors?

Thanks again v much :slight_smile:

Long before I had solar PV and HA, I decided to go with Somfy for our new house blinds. They are all on RTS, so I have a Tahoma to manage that (which works very well indeed). I bought the Somfy thermis and the sunis devices too. These are native io rather than RTS, so more intelligent. Temperature one is in the kitchen/family living area, the sunis is on the outside wall, both facing west. The Tahoma (now upgraded to a newer Tahoma Switch) on its own runs basic routines, so that, between 12:00 and 17:00 if the kitchen is over 25 C and the sun is over 30,000 lux, the kitchen/living area blinds close. Works perfectly, and now my Tahoma is integrated into HA I have access to both sensors.

The sunis goes up to 100,000 lux I think, but is certainly very responsive and sensitive. One AA battery lasts for two to three years.

As my solar pv planes are on the 35 degree garage roof, facing ENE and WSW, the west side is in symapthy with the sunis, but the sunis being on a vertical wall means it is far more responsive to the (late) afternoon sun than the pv is - perfect for gauging insolation through the windows. This time of year, the sun angle is such that the pv does not respond as much as the sunis does, and in high summer, it tends to be the other way around. Horses, or in this case sensors, for courses.

I have got as far as tracking solar PV plane DC power output, trying to improve my solar forecast match, but not to try and match that with temperature, solar lux, wind direction, and the price of fish. Of course, as you live in Cheshire, you will be very conversant with both cheese and cats. Down here on the edge of the Bristol Channel we have to contend with the leftover weather coming in from Wales.

Oh, and my three day weather forecast for Wednesday is now showing as mostly cloudy. It will be ‘rain’ by Tuesday.

EDIT
Ok - very quick and dirty comparison. Here is my total solar pv output for the same period.

Visually, the right hand side (afternoon) will match better, the left not so - the sunis is in the shade until after midday, and naturally the pv planes will monitor the whole sky better even during cloudy periods.

Thanks for that - I got a free comedy sketch mixed into a highly informative message! So Sinis looked great - until I found that its over £100 for basically an LDR! I think I’ll plod on with a more refined version of the double white envelope cover for my north-easterly and south-easterly proposed new light sensors!

Of course the PV panels themselves are probebly the best sunlight sensors! May look at getting some small ones to both power a (battery for a) ESP32 and send a directional solar harvest figure through - just finding the lithium battery / small PV panel learning curve a bit to steep for my 58 yr old brain:(

Update - have found that Accuweather integration has “hours of sun” entities (??or attributes) for days zero to 4. I think these will do the job for me, along with what I already have from forecast.solar :slight_smile: