non UTC timezone systems

[peterdudfield]

Looks like

pv system from pvout.org are in Europe/London time not UTC (36150, 36386, 43629, 60211)

We can switch them off, or adjust the timezone

[peterdudfield]

Taken at 15:29 UTC

[peterdudfield]

select distinct on (pv_yield.pv_system_id) * from pv_yield
join pv_system on pv_yield.pv_system_id = pv_system.id
where datetime_utc > '2022-06-23 12:00'
and datetime_utc > pv_yield.created_utc + '00:30:00'
order by pv_yield.pv_system_id, datetime_utc

• 36150
• 36386
• 36482 (not using)
• 37107 (not using)
• 43629
• 60211
• 62236 (not using)

[peterdudfield]

Will set them, so ML model deosnt use them for the moment, then we can we think about adjusting them later

update pv_system
set correct_data=false
where pv_system_id in (36150,36386,43629,60211)

[peterdudfield]

The 4 following systems, do seem to make a shift in the solar profile

[peterdudfield]

Just applied this on 'development' at the moment

[JackKelly]

pv system from pvout.org are in Europe/London time not UTC

That's my understanding too. I think BB at PVOutput.org confirmed this by email a few months ago?

(Related question: Will the PV data be in UTC when it gets to the ML model in production?)