catalyst-cooperative/rmi-ferc1-eia

2nd Pass / Validate Entergy FERC-EIA

aesharpe opened this issue · 4 comments

aesharpe commented
2nd Pass / Validate Entergy FERC-EIA
aesharpe commented

Whacky records that need a little group therapy:

Big Cajun 2

  • Jumps from 2014 to 2017. 2014 and younger records don’t have a good match. But the closet one is from an Entergy subsidy. Though not the same subsidy as the 2017 + matches…

Nine Mile Point

  • Seems to switch back and forth between plant part types in a way that is suspicious.

Sterlington

  • Record for 2019 has no good match. Consistency would say 1404_2012_plant_total_11241

Ouachita

  • Record for 2008 has no good match.

Perryville

  • Record for 2019 is questionable

Washington Paris

  • Only one record to match and not quite obvious whether it's right.

Link to spreadsheet:

Entergy_fix_FERC-EIA_overrides_ebfirstpass.xlsx

e-belfer commented

Big Cajun 2

Jumps from 2014 to 2017. 2014 and younger records don’t have a good match. But the closet one is from an Entergy subsidy. Though not the same subsidy as the 2017 + matches…

This is true, and weird! These are also two different sets of FERC records (2012-2014 and 2017-2020).

Nine Mile Point

Seems to switch back and forth between plant part types in a way that is suspicious.

This is true, but the FERC IDs are also changing (e.g., plant part types are consistent within IDs). Not sure how unusual this is, though!

Sterlington

Record for 2019 has no good match. Consistency would say 1404_2019_plant_total_11241

Oof. I'd say this and the 2018 record should both just match to plant_total (and not to the retired portion as well), since no combination of retired and unretired brings us closer to matching the actual capacity and this record is both most consistent and closest to 1/10 of the capacity.

Ouachita

Record for 2008 has no good match.

Personally, I think because none of the sub-components (or their combinations) is clearly a good match, going with consistency is the best here.

Perryville

Record for 2019 is questionable

Net gen matches both the 55620_2019_plant_total_11241 and 55620_1_2019_plant_unit_total_11241 records about as well. Even though capacity matches the 1 record more closely, the total record is more consistent to both 2018 and 2020 records, where it is clearly the best match. For that reason I'd suggest going with the total record here.

Washington Paris

Only one record to match and not quite obvious whether it's right.

I think the capacity and install year match quite well here, and net gen is missing. The rest of the records have a lower capacity or are "proposed" so I think this is our best choice here.

aesharpe commented

Updates:

  • Big Cajun 2 - Kept matches for different subsidies.
  • Nine Mile Point - Kept as is (with switching plant parts) mostly because 2015 was a weird year.
  • Sterlington - Matched 2019 record with plant total + retirement for consistency
  • Ouachita - Kept 2008 record for consistency.
  • Perryville - Matched to plant total for consistency and because net gens for that and the unit record previously matched were similar.
  • Washington Paris - Kept match as best option.