Accuracy of the exported Hourly production

[CarlSEL]

Hi,

I am wondering what is the accuracy of the exported production in hourly format when using Meteonorm 8.2? It would be helpful if you have any figures.

Thanks

 

[developer_bh]

Hey CarlSEL,

there are two ways to read your question:

• Do you mean what is the accuracy of the hourly GHI values from Meteonorm in comparison to the minute values?
• Or do you mean what is the accurarcy of Meteonorm 8.2 overall?

The second question is hard to answer, because the evaluation is not simple. Meteonorm offers TMY's, which means they don't represent a specific year (which you could evaluate if you have the exact weather data + output data of that location), but a represantative year of at least 10 years. On top of that, Meteonorm uses a special synthesis method of satellite data which is part of their business secret.

But we evaluated the meteonorm 8.2 files internally in regards to plausibility, similiar to NREL studies [1].

MeteoNorm 3x3(COV).bmp

You can see that the meteonorm data for germany has a really low COV, which is a very good result and means that it's very plausible.

Best regards,

Ben

[1] https://www.nrel.gov/grid/solar-resource/assets/data/ases_47760_final.pdf

 

[CarlSEL]

Thank you for you response.

Yes what is the difference in accuracy when using the minute values against the hourly? 

Thanks

[hotline_oh]

Dear CarlSEL,

Thank you for your inquiry.

Synthesis of minute values: If required, the global radiation, which is supplied in hourly resolution, can be divided into minute values. In this process, 60 minute values are synthesized from each hourly value according to the Hofmann method.

Minute values of global radiation are particularly important in order to simulate yield losses due to inverter underdimensioning, inverter control or the interaction of PV system, consumers and storage systems as realistically as possible. Example: If your consumption profile has a resolution of 15 minutes, we recommend the simulation with minute values, as the consumption profile is otherwise averaged to 1h values.

We estimate that the accuracy of the simulation with hourly values deviates 3-5% from the simulation with minute values.