0

I am trying to generate irradiance data using PVLIB Python, on a vertical plane with data input on a horizontal plane collected from the PVGIS website. The downloaded files are in .csv format. and contains irradiance values in the hourly frequency

# creating a lambda function to parse the date in csv file
mydateparser = lambda x: pd.datetime.strptime(x, "%Y%m%d:%H%M") 

irrad_horz_ERA5 = pd.read_csv(
    (r'C:\Users\dpaul\OneDrive\1st part\Downloaded csv\PVGISERA5\Horizontal'
     r'\Timeseries_45.467_9.191_E5__0deg_0deg_2010_2016.csv'),
    parse_dates=['Date'], date_parser=mydateparser)

latitude = 45.466
longitude = 9.183
altitude = 120
city = 'Milano'
timezone = 'Europe/Rome'
ground_albedo = 0.2 

times = pd.date_range(
    '2010-1-1 00:00:00', periods=52584,
    freq='1 h', tz=timezone) # 2010 to 2015


solpos = pvlib.solarposition.get_solarposition(
    times, location.latitude, location.longitude) # solar position
columns = ['apparent_zenith', 'azimuth']
solpos[columns] = solpos[columns].apply(pd.to_numeric, axis=1)

cols = ['ghi', 'Bi', 'Di']
irrad_horz_ERA5[cols] = irrad_horz_ERA5[cols].apply(pd.to_numeric, axis=1)

westERA5 = pvlib.irradiance.get_total_irradiance(
    surface_tilt=90, surface_azimuth=270, solar_zenith=solpos.apparent_zenith,
    solar_azimuth=solpos.azimuth, ghi=irrad_horz_ERA5.ghi,
    dni=irrad_horz_ERA5.Bi, dhi=irrad_horz_ERA5.Di, albedo=ground_albedo)

westERA5.head(20)

This is generating continuous NaN values like in the attached. I am not sure if the pvlib function I am using is either incorrect or anything else.

pvlib output is all NaN

Mark Mikofski
  • 19,398
  • 2
  • 57
  • 90
Debayan Paul
  • 95
  • 2
  • 9
  • Your code fragment is pretty hard to read. Perhaps you could put some spaces and newlines in? – adr Nov 19 '19 at 19:16
  • Can you please post the contents of `Timeseries_45.467_9.191_E5__0deg_0deg_2010_2016.csv` and `irrad_horz_ERA5` to determine if the file is correctly parsed by pandas? Also try using the values of `ghi`, `Bi`, and `Di` instead of series -- is it possible that `irrad_horz_ERA5` has different timestamps than `solpos`? That may be the cause of the `NaN` if `irrad_horz_ERA5` is correctly parsed, but has different times. – Mark Mikofski Oct 14 '20 at 19:03

0 Answers0