# Determine the solar installation size based on W/m² solar irradiance data

I have access to live solar irradiance data (in W/m²) at each minute of the day in my location. How do I make use of these datas to determine the solar installation size (in kW)?

I guess what I don't really understand here is what really happens within the solar system the moment the sun radiate a power of say 900W/m² at that minute. Does the panel immediately generate its potential power at 900W/m² solar irradiance at that minute or am I completely missing the point here?

I hope the kind fellows who will answer this question do not suggest to look for datas in kWh/m²/day to determine installation size as the internet has a lot of articles on it already.

The conversion of light into electricity happens pretty much instantaneously. Well, it takes time, but such a short amount of time that we can't perceive it without some serious instrumentation.

So if the insolation on the panels themselves at one second is 500W/m², and your system's efficiency is 20%, you'll get 100W for every m² of panel in that second.

If cloud passes over and insolation halves to 250W/m² for ten seconds, then your power will also halve for ten seconds (assuming it's a small installation, so that the entire array effectively experiences the same cloud at the same moment).

For grid-connected systems, this doesn't really make any difference: the rest of the grid acts as a massive storage buffer that smooths everything out. So in general, the number that interests people is how much energy will they get on average per day on an average day of the year. The system designer still has to account for the maximum generation at any one moment, to pick the correct inverter.

If, however, it's a standalone system, not grid-connected, and so does not have the luxury of that storage buffer, then it usually has a battery connected: that battery provides the smoothing. But now the system designer needs to do an hour-by-hour calculation for, say, 3 years, as well as accounting for peak momentary power in and out, to size the battery (and inverter if required) correctly for the application. As the battery is going to store enough energy for at least several hours of load, the modelling doesn't need to be any more detailed than hourly. But if the particular application had little or no storage, then you'd have to model down to the level of individual seconds, or possibly even sub-second intervals, to get a correct system design.

• Thank you for your detailed answer. In that case, is there a way to convert the insolation data (W/m²) of each minute to Wh/m²/day? Jan 5, 2016 at 0:57
• Yes there is - would you like to ask that as a separate question, so that others will find it and its answer, rather than leaving it buried in comments (comments can get deleted at any time, they're temporary and ephemeral) Jan 5, 2016 at 7:31
• Alright I've asked, thanks for pointing that out. Jan 5, 2016 at 13:01