Weather downloading not compatible with Hercules runs

[genevievestarke]

The weather downloading functions are not quite lined up to work with the current Hercules input structure.

[paulf81]

Hi @genevievestarke thank you for flagging this! I would like to just tag @ejsimley and @misi9170 into this discussion to start.

Then also, I was using these scripts in a recent project and thought I would copy in my code to show how I go from what is downloaded to Hercules ready input. Perhaps these need to be folded back into the script:

# Download data from NSRDB
nsrdb_data = download_nsrdb_data(
    target_lat=target_lat,
    target_lon=target_lon,
    year=year,
    variables= ["wind_speed", "air_temperature", "ghi", "dni", "dhi"],
    plot_data=True,
    plot_type="map",
)

# If you want just one series can find a specific point of interest:
# find the index of the coordinate that is closest to the target
closest_index = np.argmin(np.sqrt((nsrdb_data['coordinates']['lat'].values - target_lat) ** 2 + (nsrdb_data['coordinates']['lon'].values - target_lon) ** 2))
coordinate = nsrdb_data['coordinates'].index[closest_index]
coordinate

# Organize into a dataframe those single entries
ghi_single = nsrdb_data['ghi'][[coordinate]]
dni_single = nsrdb_data['dni'][[coordinate]]
dhi_single = nsrdb_data['dhi'][[coordinate]]
temp_single = nsrdb_data['air_temperature'][[coordinate]]
ws_single = nsrdb_data['wind_speed'][[coordinate]]

ghi_single.columns = ['ghi']
dni_single.columns = ['dni']
dhi_single.columns = ['dhi']
temp_single.columns = ['temp']
ws_single.columns = ['ws']

df_solar = ghi_single.merge(dni_single, left_index=True, right_index=True)
df_solar = df_solar.merge(dhi_single, left_index=True, right_index=True)
df_solar = df_solar.merge(temp_single, left_index=True, right_index=True)
df_solar = df_solar.merge(ws_single, left_index=True, right_index=True)

df_solar = df_solar.reset_index()
df_solar = df_solar.rename(columns={'time_index': 'time_utc'})

# Rename columns for PVWatts
rename_dict = {
    "ghi": "Global Horizontal Irradiance",
    "dni": "Direct Normal Irradiance",
    "dhi": "Diffuse Horizontal Irradiance",
    "temp": "Temperature",
    "ws": "Wind Speed at",
}
df_solar = df_solar.rename(columns=rename_dict)

[genevievestarke]

Ah, nice, this is exactly what I was looking for! Maybe we can add some pre-processing functions into the resources folder (and workflow). I should be able to get something worked up next week!

[paulf81]

Awesome, thank you!

[paulf81]

#246 was a first step in this direction

[paulf81]

I think with #208 we can restart this effort by deciding how we want the 3 main modules (and any future to work) my ideas are:

1. Remove the year or start/stop input format and fix to start/stop only as more general
2. Allow specification that a grid of points filling a bounding box should be downloaded, a list of points, or a single points
3. Decide what Hercules-ready data should be. For wind this is clear, ws_000 etc for each turbine and wd_mean, but should we allow multiple input locations for solar? Hercules can't presently use these.

@genevievestarke and @misi9170 , maybe we need to have this all sketched out in a discussion before we can really dive into all this?