Menu
JAQForum Ver 19.10.27

Forum Index : Solar : Prediction of solar PV output for day(s) ahead.

Posted: 10:49pm
17 May 2026
Copy link to clipboard
Andrew_G
Guru

G'day,
I'm seeking views on the prediction of solar PV output for day(s) ahead.
(I'm new to this section of TBS, I normally lurk on the Microcontroller and PC projects section.)

We have had a new solar system for a few months (14kW of panels, 8kW inverter, 16.6 kWh battery).  
Last night was the first time we dropped to minimum SoC (set at 15%) and we had to import a fair bit of power.
At sunset we only had 90% in the battery - I could have held back on some consumption.
That prompted me to Google sites that helped predict solar output based on weather forecasts.
The one that caught my eye was 'OpenWeather'. I already use it for weather (current and forecast) as it outputs in JSON - which I can use via my Webmite.

Do any of you have any comments or suggestions?

Cheers,

Andrew
 
Posted: 12:59am
18 May 2026
Copy link to clipboard
Revlac
Guru


I did see https://forecast.solar/  but I think you have one that is probably better, however I don't think this is the total solution for this weather situation, your overnight consumption is high compared to your storage and production,   back later.
 
Posted: 01:46am
18 May 2026
Copy link to clipboard
KeepIS
Guru


This might get confusing with a post in both Solar and Micro forums.

You likely know that Windy.com has a solar Power index map layer, I believe you can pull data, I normally know how we will go by looking at predicated cloud cover.

I have 11kW of panels running at the moment, most are Flat, some South and East facing, only 8 are facing north, in South-East QLD.

With my panel configuration I keep single panel shading effects to a minimum and the flat and south panels are very effective in overcast rainy weather, where slope and direction are meaningless.

There are 7 solar chargers, one for each 4 panel array, it's raining and completely overcast today, currently solar index indicates 130W/m2, each solar array and charger are producing exactly 3.7A, South, North, East and Flat.

7 x 3.7A = 25.9A @ 53.2v = 1.38kW.

The solar index level indicator is real time tracking with the movable readout flag over my location. It's cycling between 92W/m2 and 296W/m2. Power is dropping to 700W under darker cloud cover with heavy rain.

I have over twice the battery capacity you have, and with the rain yesterday we are down to 64% SOC (Full electric home and workshop) another day of no real solar and we will be down to around 34%, it looks like the following day (WED) will be brighter with some lighter cloud cover, that should easily get us back to around 70% SOC. We have not used Mains for over a year.

I assume you have LiFePO4 batteries and are in a Southern state.
 
Posted: 04:16am
18 May 2026
Copy link to clipboard
Andrew_G
Guru

Hi Mike and Aaron,
Thanks for your replies.
I'm a complete novice on the solar front but have been playing with MMBasic for 10 years.
I posted to both forums as I assumed some shedders would have an interest in one field.
I know I hadn't been looking at the other forums.

We are 5k east of Melbourne, are fully electric with no gas. The house is to Passivhaus standards (heavily insulated, triple glazed and a heat recovery system).
It has a 'tin' roof, at 15deg facing north. All the unshaded roof has PV panels with micro-inverters. Yes it is a LiFePO4 battery (BYD).
With a full battery at sunset we typically dip to 40% and by playing with the the pool equipment get it back to 100% each day (except yesterday). We are not exporting yet.

I'll go looking at Windy.com.

Cheers,

Andrew
 
Posted: 04:45am
18 May 2026
Copy link to clipboard
TassyJim
Guru


I just did a test with the site Aaron referred to.
It gave good realistic readings for today (mostly sunny)
at 10am their 1315W compared to my 1295W

I suggest trying the public api for a few weeks to see how it compares with actual before getting to involved.
https://doc.forecast.solar/api:estimate

Jim

edit:
for the last two days, actual generated was ~7% lower than the forecast.solar estimate.
Early days but looking usable for short time forecasts.
Edited 2026-05-20 13:07 by TassyJim
 


To reply to this topic, you need to log in.

The Back Shed's forum code is written, and hosted, in Australia.
© JAQ Software 2026