PV LIB results x IES VE results

[Ricardo Filho]

Hi all, 

i am trying to create a database with 672 variations of PV Panels systems, and to validate the results from PV LIB i am using IES VE and the Solar Atlas website

has anyone ever done anything similar?

I am getting very different results when I compare the 3 sim tools

example

for a 0 degree inclination 0 azimuth 5Kwp system

the VE gives me 4.2 MWh, solar atlas 4 and PV lib 9.1

Happy to share the code 

[cwh...@sandia.gov]

What location? 4MWhr for a 5kwp system seems very low.

[Ricardo Filho]

Dublin, Ireland

not the best location

[cwh...@sandia.gov]

Right, then 4MWhr is credible, and 9.1MWhr is suspect.

Can you share your pvlib code?

[Ricardo Filho]

import pvlib
import pandas as pd
import numpy as np

# Define PV Types and Locations
pv_types = {
    'Amorphous Silicon': {'Efficiency': 0.0990, 'NOCT': 45.0},
    'Monocrystalline Silicon': {'Efficiency': 0.2010, 'NOCT': 42.0},
    'Polycrystalline Silicon': {'Efficiency': 0.1700, 'NOCT': 44.0},
    'Other Thin Films': {'Efficiency': 0.2285, 'NOCT': 44.0},
    'Thin Film Cadmium-Telluride': {'Efficiency': 0.1200, 'NOCT': 42.8},
    'Thin Film Copper-Indium-Gallium-Selenide': {'Efficiency': 0.1300, 'NOCT': 42.8}
}

# Load EPW file for local weather data
epw_file_path = 'IRL_Dublin.039690_IWEC.epw'  # Adjust the filename as needed
weather_data, metadata = pvlib.iotools.read_epw(epw_file_path, coerce_year=2023)

# Extract necessary data
times = weather_data.index
ghi = weather_data['ghi']
dni = weather_data['dni']
dhi = weather_data['dhi']
dni_extra = pvlib.irradiance.get_extra_radiation(times)  # Calculate dni_extra
latitude, longitude, tz = metadata['latitude'], metadata['longitude'], metadata['TZ']

# User inputs for PV type and output type
print("Select the PV type:")
for idx, pv_type in enumerate(pv_types, 1):
    print(f"{idx}. {pv_type}")
pv_type_index = int(input("Enter the number corresponding to the PV type: ")) - 1
pv_type_name = list(pv_types.keys())[pv_type_index]
pv_type = pv_types[pv_type_name]

output_type = input("Enter 'timeseries', 'total', or 'monthly': ").strip().lower()

# Simulation settings
powers = range(5, 125, 5)
inclinations = range(0, 35, 5)
orientations = [0,180, 225, 135]  # South, South-West, South-East

# Prepare data collection
all_data = {'Datetime': times}
column_names = []

# Perform the simulation
location = pvlib.location.Location(latitude, longitude, tz)
solpos = location.get_solarposition(times)

for power in powers:
    for inclination in inclinations:
        for orientation in orientations:
            poa_irradiance = pvlib.irradiance.get_total_irradiance(
                surface_tilt=inclination,
                surface_azimuth=orientation,
                solar_zenith=solpos['apparent_zenith'],
                solar_azimuth=solpos['azimuth'],
                dni=dni, ghi=ghi, dhi=dhi, dni_extra=dni_extra,  # Include dni_extra here
                model='perez'  # Using Perez model for better accuracy
            )
            system_output = power * pv_type['Efficiency'] * poa_irradiance['poa_global']
            col_name = f'{power}_{inclination}_{orientation}_{pv_type_name}'
            column_names.append(col_name)
            all_data[col_name] = system_output

# Convert dictionary to DataFrame
results = pd.DataFrame(all_data)
results.set_index('Datetime', inplace=True)

# Handle output based on user choice
if output_type == 'timeseries':
    results.to_csv('solar_simulation_time_series_results.csv')
elif output_type == 'total':
    total_generation = results.sum().to_frame(name='Total Generation')
    total_generation.to_csv('solar_simulation_total_generation_results.csv')
elif output_type == 'monthly':
    monthly_totals = results.resample('M').sum()
    monthly_totals.to_csv('solar_simulation_monthly_totals.csv')

print(f"Simulation completed and results are saved according to selected output type: {output_type}.")


weather file from Energy Plus

[cwh...@sandia.gov]

Fairly sure the problem is with the irradiance-to-power calculation:

            system_output = power * pv_type['Efficiency'] * poa_irradiance['poa_global']

I assume power "5" means 5kW at 1000 W/m2 (rating condition). The conversion efficiency and area are implicit in the rating. 

Instead of the above, you should calculate

             system_output = power * poa_irradiance['poa_global'] / 1000 W/m2

Or better, if you have air temperature in the EPW file (I think it should be there, haven't looked), calculate a module temperature using one of the models in pvlib.temperature, then use pvsystem.pvwatts_dc to get the power. Efficiency decreases as temperature increases; the above equations assume the modules are always at the rating condition of 25C.

I would also check that the poa_global is zero at night.

[Ricardo Filho]

Does it look better now?

import pvlib
import pandas as pd
import numpy as np

# Define PV Types and Locations
pv_types = {
    'Amorphous Silicon': {'Efficiency': 0.0990, 'NOCT': 45.0},
    'Monocrystalline Silicon': {'Efficiency': 0.2010, 'NOCT': 42.0},
    'Polycrystalline Silicon': {'Efficiency': 0.1700, 'NOCT': 44.0},
    'Other Thin Films': {'Efficiency': 0.2285, 'NOCT': 44.0},
    'Thin Film Cadmium-Telluride': {'Efficiency': 0.1200, 'NOCT': 42.8},
    'Thin Film Copper-Indium-Gallium-Selenide': {'Efficiency': 0.1300, 'NOCT': 42.8}
}

# Load EPW file for local weather data
epw_file_path = 'IRL_Dublin.039690_IWEC.epw'

weather_data, metadata = pvlib.iotools.read_epw(epw_file_path, coerce_year=2023)

# Extract necessary data
times = weather_data.index
ghi = weather_data['ghi']
dni = weather_data['dni']
dhi = weather_data['dhi']
dni_extra = pvlib.irradiance.get_extra_radiation(times)

latitude, longitude, tz = metadata['latitude'], metadata['longitude'], metadata['TZ']

# User inputs for PV type and output type
print("Select the PV type:")
for idx, pv_type in enumerate(pv_types, 1):
    print(f"{idx}. {pv_type}")
pv_type_index = int(input("Enter the number corresponding to the PV type: ")) - 1
pv_type_name = list(pv_types.keys())[pv_type_index]
pv_type = pv_types[pv_type_name]

output_type = input("Enter 'timeseries', 'total', or 'monthly': ").strip().lower()

# Simulation settings
powers = range(5, 125, 5)
inclinations = range(0, 35, 5)

orientations = [0, 180, 225, 135]  # Cardinal orientations

# Prepare data collection
all_data = {'Datetime': times}
column_names = []

# Perform the simulation
location = pvlib.location.Location(latitude, longitude, tz)
solpos = location.get_solarposition(times)

for power in powers:
    for inclination in inclinations:
        for orientation in orientations:
            poa_irradiance = pvlib.irradiance.get_total_irradiance(
                surface_tilt=inclination,
                surface_azimuth=orientation,
                solar_zenith=solpos['apparent_zenith'],
                solar_azimuth=solpos['azimuth'],
                dni=dni, ghi=ghi, dhi=dhi, dni_extra=dni_extra,

                model='perez'
            )
            # Calculate system output based on normalized irradiance

            system_output = power * poa_irradiance['poa_global'] / 1000

            col_name = f'{power}_{inclination}_{orientation}_{pv_type_name}'
            column_names.append(col_name)
            all_data[col_name] = system_output

# Convert dictionary to DataFrame
results = pd.DataFrame(all_data)
results.set_index('Datetime', inplace=True)

# Handle output based on user choice
if output_type == 'timeseries':
    results.to_csv('solar_simulation_time_series_results.csv')
elif output_type == 'total':
    total_generation = results.sum().to_frame(name='Total Generation')
    total_generation.to_csv('solar_simulation_total_generation_results.csv')
elif output_type == 'monthly':
    monthly_totals = results.resample('M').sum()
    monthly_totals.to_csv('solar_simulation_monthly_totals.csv')

print(f"Simulation completed and results are saved according to selected output type: {output_type}.")

[cwh...@sandia.gov]

>  Does it look better now?

Yes, better. What annual energy do you get?

Since you have a NOCT value for each module type, you could make a crude adjustment for temperature: 

system_output = power * poa_irradiance['poa_global'] * (1 - 0.004 * (NOCT - 25))

This assumes a temperature coefficient for power of 0.4%/C, which is typical for silicon, maybe a bit high for CdTe. You can find temperature coefficients on module datasheets, so I'm assuming you would have values for the 672 modules you want to model.

[Ricardo Filho]

Based on the code below I am getting 4.3MWh against 4.1 from the VE and 3.9MWh from solar atlas

not sure if this is an Ok error 

import pvlib
import pandas as pd
import numpy as np

# Define PV Types and Locations with additional temperature coefficients
pv_types = {
    'Amorphous Silicon': {'Efficiency': 0.0990, 'NOCT': 45.0, 'TempCoeff': 0.0043},
    'Monocrystalline Silicon': {'Efficiency': 0.2010, 'NOCT': 42.0, 'TempCoeff': 0.0040},
    'Polycrystalline Silicon': {'Efficiency': 0.1700, 'NOCT': 44.0, 'TempCoeff': 0.0042},
    'Other Thin Films': {'Efficiency': 0.2285, 'NOCT': 44.0, 'TempCoeff': 0.0029},
    'Thin Film Cadmium-Telluride': {'Efficiency': 0.1200, 'NOCT': 42.8, 'TempCoeff': 0.0025},
    'Thin Film Copper-Indium-Gallium-Selenide': {'Efficiency': 0.1300, 'NOCT': 42.8, 'TempCoeff': 0.0027}

}

# Load EPW file for local weather data
epw_file_path = 'IRL_Dublin.039690_IWEC.epw'
weather_data, metadata = pvlib.iotools.read_epw(epw_file_path, coerce_year=2023)

# Extract necessary data
times = weather_data.index
ghi = weather_data['ghi']
dni = weather_data['dni']
dhi = weather_data['dhi']

temp_air = weather_data['temp_air']
wind_speed = weather_data['wind_speed']

dni_extra = pvlib.irradiance.get_extra_radiation(times)
latitude, longitude, tz = metadata['latitude'], metadata['longitude'], metadata['TZ']

# User inputs for PV type and output type
print("Select the PV type:")
for idx, pv_type in enumerate(pv_types, 1):
    print(f"{idx}. {pv_type}")
pv_type_index = int(input("Enter the number corresponding to the PV type: ")) - 1
pv_type_name = list(pv_types.keys())[pv_type_index]
pv_type = pv_types[pv_type_name]

output_type = input("Enter 'timeseries', 'total', or 'monthly': ").strip().lower()

# Simulation settings

powers = range(5, 125, 5)  # in kW

inclinations = range(0, 35, 5)
orientations = [0, 180, 225, 135]  # Cardinal orientations

# Prepare data collection
all_data = {'Datetime': times}
column_names = []

# Perform the simulation
location = pvlib.location.Location(latitude, longitude, tz)
solpos = location.get_solarposition(times)

for power in powers:
    for inclination in inclinations:
        for orientation in orientations:
            poa_irradiance = pvlib.irradiance.get_total_irradiance(
                surface_tilt=inclination,
                surface_azimuth=orientation,
                solar_zenith=solpos['apparent_zenith'],
                solar_azimuth=solpos['azimuth'],
                dni=dni, ghi=ghi, dhi=dhi, dni_extra=dni_extra,
                model='perez'
            )

            temp_module = pvlib.temperature.pvsyst_cell(temp_air, poa_irradiance['poa_global'], wind_speed, pv_type['NOCT'])
           
            system_output = power * poa_irradiance['poa_global'] * (1 - 0.004 * (pv_type['NOCT'] - 25))

            col_name = f'{power}_{inclination}_{orientation}_{pv_type_name}'
            column_names.append(col_name)
            all_data[col_name] = system_output

# Convert dictionary to DataFrame
results = pd.DataFrame(all_data)
results.set_index('Datetime', inplace=True)

# Handle output based on user choice
if output_type == 'timeseries':
    results.to_csv('solar_simulation_time_series_results.csv')
elif output_type == 'total':

    total_generation = results.sum().to_frame(name='Total Generation kWh')
    total_generation['Total Generation MWh'] = total_generation['Total Generation kWh'] / 1000

    total_generation.to_csv('solar_simulation_total_generation_results.csv')
elif output_type == 'monthly':
    monthly_totals = results.resample('M').sum()

    monthly_totals['Total Generation kWh'] = monthly_totals.sum(axis=1)
    monthly_totals['Total Generation MWh'] = monthly_totals['Total Generation kWh'] / 1000

    monthly_totals.to_csv('solar_simulation_monthly_totals.csv')

print(f"Simulation completed and results are saved according to selected output type: {output_type}.")

[cwh...@sandia.gov]

Given the differences in input weather, the unknown weather-to-power modeling in the Solar Atlas and in VE, and the unknown assumptions about module parameters, I'd say 4.3 / 4.1 / 3.9 are reasonably close.

Close enough depends on what you are trying to learn from the comparison.

[Ricardo Filho]

Thanks!
my goal is to create a database with 672 variations power from 5 to 120kWp, various inclinations and orientations 
and then get an app based on user timeseries data for consumption  x a given scenario from the database

[Ricardo Filho]

Is this the parameter you mentioned before?

[cwh...@sandia.gov]

Yes. In the implementation, be sure that temperature > 25C reduces power. 

[Ricardo Filho]

Got even closer now, the VE assumes a degradation factor of .99 and conversion losses of 4%
factor these in I am getting 4.2MWh on PVLIB and 4.12 on the VE with the same weather file