feat: process plant data to create data frame

[danielolsen]

Pull Request doc

Purpose

Parse HIFLD data into the beginning of a plant data frame, for use with a Grid object.

What the code is doing

build_plant is the entry-point function, which will eventually return a data frame with all information, and all lower-level functions are accessed from this function. build_plant effectively replaces the old plant_agg function, with different structure. The lower level functions are:

• map_generator_to_sub_by_location: look up substations within that interconnect and ZIP code, or if there are none, look within that interconnect and 'nearby' ZIP codes (+- 100). The logic of 'nearby' could probably be improved, but this is what was implemented in the collaborators' code.
• map_generator_to_bus_by_sub: pick the lowest voltage bus within a specified substation.
• add_locations_to_generators: use the locations from Power_Plants.csv, map them to the generating units as best as possible.
• map_generators_to_interconnects: look up the county for each substation using the Power Plant mapping, then determine the interconnect.
• add_pmin_to_generators: look up Pmin values from EIA Form 860.
• helpers.map_state_and_county_to_interconnect: look up information from const.py to map a state/county to an interconnection. We also call this from data_process.transmission so that we can access this information for substations from within map_generator_to_sub_by_location.

There are also placeholders for:

• estimate_heat_rate_from_ampd
• add_heat_rate_to_generators

These will need further research into the best way to implement, but they are independent and so can wait for a follow-up PR.

Usage Example

>>> from prereise.gather.griddata.hifld.data_process.transmission import (
...     build_transmission,
...     create_buses,
... )
>>> from prereise.gather.griddata.hifld.data_process.generators import build_plant
>>> lines, substations = build_transmission()
>>> bus = create_buses(lines)
>>> plant = build_plant(bus, substations)
>>> plant.isna().sum()
ID                0
NAME              0
CITY              0
STATE             0
ZIP               0
TYPE              0
STATUS            0
COUNTRY           0
NAICS_CODE        0
NAICS_DESC        0
SOURCE            0
SOURCEDATE        0
PLANT             0
PLANT_CODE       86
OPERATOR          0
OPERAT_ID         0
NAMPLT_CAP        0
SUMMER_CAP        0
WINTER_CAP        0
PRIM_FUEL         0
FUEL1_TRAN        0
SEC_FUEL          0
FUEL2_TRAN        0
PRIM_MVR          0
RUN_DATE          0
RET_DATE          0
lat             402
lon             402
interconnect    402
sub_id          687
bus_id          687
Pmin              0
dtype: int64

Not all plant codes can be mapped successfully, and same with lats/lons, and therefore sub_id & bus_id, but we're only missing about 3%. Suggestions for improvement welcome.

Time to review

1 hour.

[rouille]

Nice