Missing from Uncertainty Analysis: Weather Data Uncertainty?

[kjs55]

The weather (solar resource) data uncertainty contribution (e.g., solar irradiance measurement uncertainty) appears to be missing from the uncertainty propagation in PVsyst (?), thus leading to significantly overly aggressive P99s. Please consider incorporating this #1 source of uncertainty into the uncertainty analysis in PVsyst. Thanks.

Edited March 7, 2023 by kjs55
Remove commas

[dtarin]

This is typically added under the custom field and renamed. All fields should be set by the user, and not dictated by the software as it is site and project specific.  

Edited March 7, 2023 by dtarin

[kjs55]

@dtarin: "Typically" according to whom? Please provide a reference (preferably written by the PVsyst team, unless you work for PVsyst in which case I accept your reply).

Edited March 7, 2023 by kjs55
Incomplete post

[dtarin]

Typically by users. It is a custom field, so use it as you need to. Whether or not there is a specific label in the software does not matter. The software takes all uncertainties entered and combines them (square root of sum of squares). The label you give it shows up in the report.

[Agnes Bridel]

Thank you for the suggestion. There is indeed uncertainty contribution from weather data. However, currently we do not have a way to know the uncertainty of all different data sources. Moreover, when using measured data as weather data source, this uncertainty will depend on many factors such as the measurement devices themselves. You are welcome to send us references if you have more details on the uncertainty. We may of course consider including this factor as one of the default uncertainties in PVsyst in the future, provided we have reasonable grounds for providing a "typical" uncertainty value.
The current workaround to include the uncertainty is by adding a custom variability.

Best regards,

[kjs55]

I think having a specific/dedicated, labeled placeholder (user input field) for it would make it clear to the PVsyst user that it's essential to include in the uncertainty propagation used to derive the P90s, etc. (being a primary, & often the #1, source of uncertainty in the propagation).

Edited April 4, 2023 by kjs55
Slight editorial