Solar Discussion - Should we start using PVGIS-5 ?

zeupater

Hi

I've just read through a brief review of the features and accuracy which the latest version of PVGIS offers and it looks to be pretty interesting .... The Vriconian site suggests that ...

... taking the system's total generation since install against the PVGIS5-CMSAF dataset the discrepancy is 1.4% (high) before making allowance for minor winter shading, which looks remarkably good compared to the level of underestimation apparent on all previous model estimates ...

Solar PV Generation Estimation: PVGIS-5 (Beta Release)

If that degree of accuracy is repeatable, should we be looking to review our targets and start to recommend to others that future targets are set using PVGIS-5 with the climate model (CMSAF) ?

Thoughts ?

HTH
Z

Martyn1981

I finally changed all my figures for Jan 1st ....... now I gotta do it again. Boo!

Merlin139

http://re.jrc.ec.europa.eu/pvg_tools/en/tools.html

Just been on it and played about.

It increased mine by about 7% over the year

ASavvyBuyer

By using the latest PVGIS, do we actually get more generation from our panels ?

zeupater

Merlin139 wrote: »

http://re.jrc.ec.europa.eu/pvg_tools/en/tools.html

Just been on it and played about.

It increased mine by about 7% over the year

Hi

I checked ours earlier & against PVGIS4 it's raised the annual target by around 5.5% (~WSW array), bringing it much closer to our long term average so I'm minded to reset our targets .... might leave it a while yet though to see if there's any 'tweaking' between beta and the final release, but it does look much more accurate from here ...

HTH
Z

cjmillsnun

Looked at it and reset my targets. It looks more realistic.

Martyn1981

zeupater wrote: »

Hi

I checked ours earlier & against PVGIS4 it's raised the annual target by around 5.5% (~WSW array), bringing it much closer to our long term average so I'm minded to reset our targets .... might leave it a while yet though to see if there's any 'tweaking' between beta and the final release, but it does look much more accurate from here ...

HTH
Z

Sounds like a good plan. Perhaps I'll change my spreadsheet (again) for next Jan 1st.

With all my shading I shouldn't be breaking even, so averaging 90-95% seems a lot more rational, though perhaps a small dagger to the heart each month.

zeupater

Martyn1981 wrote: »

Sounds like a good plan. Perhaps I'll change my spreadsheet (again) for next Jan 1st.

With all my shading I shouldn't be breaking even, so averaging 90-95% seems a lot more rational, though perhaps a small dagger to the heart each month.

Hi

On shading, have you ever tried creating & uploading a 'horizon file' for your location within PVGIS? ... I tried it out some time ago to see how it worked, but with our level of (deep winter) shading it wasn't worth the effort ... from memory it was quite easy to create the file, but far more complex to gather the data!

HTH
Z

EricMears

Afraid I'm not a big believer in running any beta software. Look forward to an announcement that a fully tested product has become available.

zeupater

EricMears wrote: »

Afraid I'm not a big believer in running any beta software. Look forward to an announcement that a fully tested product has become available.

Hi Eric

The system pretty much looks like a finished product release candidate seeking interface feedback than a traditional software beta release, so I doubt that there'll be anything more than minor 'tweaks' to the actual model ..... after-all the basic (/default) irradiation data (where we are) is still CMSAF as used in PVGIS4, so it's really down to overlaying an averaged spectral & wind cooling effect dataset to produce the new estimates ....

HTH
Z

EricMears

zeupater wrote: »

Hi Eric

The system pretty much looks like a finished product release candidate seeking interface feedback than a traditional software beta release, so I doubt that there'll be anything more than minor 'tweaks' to the actual model

Pleased to hear that; hopefully it won't be too long before I'll be able to rely on their testing.

Alas, 'minor tweaks' to any piece of software can have unexpected results. Since it won't affect what I actually generate, there doesn't seem any point in rushing to implement a provisional system