Dear PV SOL users & team,

I usually design projects in Spain with PV SOL and I've noticed that reports usually show a lower specific annual yield than expected. For example, with equal or very similar data for a project in Barcelona, PV SOL shows 1.410,17 kWh/kWp, while SOLARGIS shows 1.487,9 kWh/kWp.

 

Can anybody please explain the difference?

 

Thanks for your help!

Hi Ricardo,

the model chain of PV*SOL is quite complex, I would recommend our help sites as a start:

https://help.valentin-software.com/pvsol/2018/calculation/

As you can see, there a really a lot of input data, parameters and algorithms that all play their part in the simulation of a PV plant. The complete comparison of the PV*SOL and the SolarGIS model chain could easily make up a Master thesis

Just to name a few key questions that strongly affect the PV output:

1. What meteorological input data do you use for irradiance and temperature?
2. Do you simulate in 1h or 1min resolution?
3. What models do you use for diffuse irradiance and transposition to the module plane?
4. What models for reflection, soiling etc?
5. How do you model the temperature and weak light behaviour of the PV module?
6. How is the circuitry and the interconnection as well as ohmic losses accounted for?
7. How detailed is the DC/AC inverter model?

But from the history of our two companies you could perhaps conclude that SolarGIS is very strong in providing meteorological input data, while we have a strong expertise in modelling the electrical behaviour of PV systems and rely on meteo data provider MeteoNorm (as you already know from a previous blog post).

Kind regards,

Martin

Hello again,

The question was not only about SolarGIS (that has been answered and quite clearly, thanks for that) but I've also heard from more experienced colleagues that PV SOL is usually a bit conservative in its estimates; that is, when PV SOL estimates a certain yield for a PV plant, once you actually build and operate it over the years your energy production will be higher than expected. This was mentioned by a German colleague, although I am designing plants mostly in Spain.

I by no means wish to question the validity of the model or algorithms used, I believe it is a very useful tool, but I was just wondering if this is just a rumour or if it indeed has been noticed and is being addressed.

Thanks and regards,

Ricardo

Hi Ricardo,

what you heard from your German colleague is actually true. We are known to model PV systems more on the conservative side. This prevents our customers from bad surprises That is, when you plan your PV system in PV*SOL, you can be very confident that you will reach and most likely surpass this yield in reality. In our opinion, this is a very important fact that system planners really appreciate in the long run.

But we can't say for example that our simulated yields are always 2% under the real yields. If you simulate 100 systems around the world with different technology and so on, the spread will be 5 to 0 % deviation, with an average of around 2%.

Hope that helps, kind regards,

Martin

It does, thank you, Martin!