3 Feeder models debugged.

[dpinney]

GOAL We need 9 prototype "models". These will be instances of the ones we wrote for each of the partners. I.e. one of (solarRates, solarFinancial, solarEngineering) for each of (GRE, Owen, Coserv).

[dpinney]

First draft: https://www.dropbox.com/s/tkogaq8y094nkpb/6.%20GOSED%20Task%203%20and%204%20Data%20Collection%20Plan.docx?dl=0

[dpinney]

Name updated. Data collection about done, now we need to run the 9 models.

[dpinney]

Data collection complete.

Data conversion to OMF formats almost done. GLMs are running, just need to calibrate to SCADA data.

[dpinney]

Completed in commit cc4872689047bad265f75b6086573976569426b0

[dpinney]

FULL DEBUGGING NOTES

• solarEngineering
  • translate and test code
  • Beckenham
  • Convert
  • SCADA in
  • Debug GLD running. Needed small manual edit to... transformer?
  • Calibrate. Flatline on chart. The years are different SCADA versus what's in calibrate.py. I updated calibrate.py to do dynamic date choosing, and it works!
  • JSON bundled version.
• Alberich
  • Fix conversion. Issue was unsafe array access on transformer lists, and also missing secondary transformer data with unsafe access. The std data is a joke
  • SCADA in--note we're using station level, and we have feeder readings if we need them
  • Debug GLD running.
  • First steps: needed to change substation transformer from 0 VA rating to 25000. Also change PT_phase and CT_phase on regulator to one phase (A) for BANK regulated transformers. Had to remove load object A and B = 0 loads when on C phase system. Still has phase problems. Had to take a non-sub transformer from 0 to 100 VA power_rating. Also had to replace all zero transformer impdeance j parts with nonzero values.
  • Separate code to check all phases.
  • Ask GLD team. Problem is center-tapped transformers have to connect to triplex nodes. Need to find/replace broken center_tapped ones. What does that involve? Replacing all the ones where XS is the phase.
  • New error: some conductors are missing. Need to walk the tree again.
  • New, new error: need to add back some center tap thingies.
  • New, new, new, new error: more phases entering than leaving. Deleted one. Error comes back. Changed phasing on nodeOTx200720OHx19973 from ABC to AC.
  • Another error: ERROR [2000-01-01 00:00:00 PST] : sync_node(obj=0;AUBREY 2): All nodes were not properly populated. GLD 3.1 says the problem is NODE150073. Deleted it and it works now.
  • Calibrate. Data in MW but we expected kW--fixed by multiplying SCADA * 1000.
  • JSON bundled version.
• TreePond
  • Windmil convert.
  • SCADA in.
  • Debug GLD Running. Have to replace zero reactances in impedance. Also needs to up some transformer ratings 10x (435666_A, 435666_C, 435666_B).
  • Calibrate. They've got 7 lines coming out of their sub! Quick hack to make it just 1.
  • JSON bundled version.
• Calibrate fixes:
  • Added additional figure creation so the graphs don't all get written on top of each other.~
  • Fixed issue with timestep being set statically to 2011-01-01.
  • Switch to complex numbers? Too hard because it would take a GLD solver rewrite.
• GLD shakedown test.
  • TreePond.
  • Alberich. Weird voltages. Sub had the wrong primary side rating (78.6 kV instead of 138kV like the source).
  • Beckenham. Needs serious debugging. First needed to change all the regulators because the band centers were wrong. No amount of transformer and regulator improvements helped (i.e. every tformer to huge rated_power, each regulator to bandwidth of 5 kV). Suspect wrong SCADA data. Or else fundamental improvements in powerflow math are needed. Cranked SCADA values down to 75% of total and we'll re-evaluate later.
• Bonus points.
  • Add weather?
  • Hack on Alberich a little more? Hack on Beckenham a little more?