Solar components

  • File name: solar_components.ipynb

  • Last edited: 2020-06-30

  • Created by: Stefan Bruche (TU Berlin)

Note:

The solar classes require the availability of the Python module pvlib. The module is not provided with the standard installation of aristopy. If you want to use the solar classes, consider installing the module in your current environment, e.g. via:

>> pip install pvlib

For further information and an installation guide, users are referred to the pvlib documentation.

import pandas as pd
import aristopy as ar

# Get data from csv-file
data = pd.read_csv('testdata.csv', sep=';', decimal='.', index_col=[0])

# Convert index to type DateTimeIndex and add required information about the time zone.
data.index = pd.to_datetime(
    data.index, format="%d.%m.%Y %H:%M").tz_localize(tz='UTC')

# Create a SolarData class instance with global (GHI) and diffuse (DHI)
# horizontal irradiation data for the location Potsdam/Germany.
solar = ar.SolarData(ghi=data['GHI [W/m2]'], dhi=data['DHI [W/m2]'],
                     latitude=52.3822, longitude=13.0622, altitude=81)


# Photovoltaic
# ------------
# Calculate the direct normal irradiation (DNI) from GHI, DHI and the solar positions
df_solar = solar.get_irradiance_dataframe()

# Append ambient temperature data (and wind speed if available)
df_solar['temp_air'] = data['T_amb [C]']

# Create a PV System (consisting of a module and an inverter).
pv_system = ar.PVSystem(module='Canadian_Solar_Inc__CS6X_300P',
                        inverter='Canadian_Solar_Inc___CSI_60KTL_CT__480V_')

# Calculate the feed-in of the PV system for specified conditions at a site
# (irradiation, temperature, collector tilt and azimuth).
pv_feed_in = pv_system.get_feedin(
    weather=df_solar, location=solar.location,
    surface_tilt=30, surface_azimuth=180,  # South
    modules_per_string=20, strings_per_inverter=10, scaling='peak_power')  # [W/Wp]


# Solar thermal collector
# -----------------------
# Calculate the irradiance components on the solar-thermal collector array.
poa = solar.get_plane_of_array_irradiance(surface_tilt=45, surface_azimuth=180)

# Set up a solar collector array (type: evacuated tube collectors).
solar_collector_data = ar.SolarThermalCollector(
    optical_efficiency=0.80, thermal_loss_parameter_1=1.1,
    thermal_loss_parameter_2=0.008, irradiance_data=poa['poa_global'],
    t_ambient=data['T_amb [C]'], t_collector_in=60, t_collector_out=90)

# Calculate heat output of the solar-thermal collector
solar_heat = solar_collector_data.get_collector_heat_output()


# Solar components in the model
# -----------------------------
# Create energy system instance
es = ar.EnergySystem()

# Add a photovoltaic component (electricity source)
pv = ar.Source(
    ensys=es, name='pv', outlet=ar.Flow('Elec', 'elec_sink'),
    time_series_data=ar.Series('pv_feed_in', pv_feed_in),  # [MW/MWp]
    capacity=100, capex_per_capacity=700e3, opex_per_capacity=0.025*700e3,  # [MWp]
    user_expressions='Elec == CAP * pv_feed_in')

# Add a Solar-thermal collector component (heat source)
solar_collector = ar.Source(
    ensys=es, name='solar_collector', outlet=ar.Flow('Heat', 'heat_sink'),
    basic_variable='Area', additional_vars=ar.Var('Area', has_time_set=False),
    time_series_data=ar.Series('collector_feedin', solar_heat / 1e6),  # [MW/m²]
    user_expressions='Heat == collector_feedin * Area',
    capacity=1e5, capex_per_capacity=300, opex_per_capacity=300*0.01)  # [m²], [EUR/m²]

# Add sinks for electricity and heat
elec_sink = ar.Sink(ensys=es, name='elec_sink', inlet=ar.Flow('Elec'))
heat_sink = ar.Sink(ensys=es, name='heat_sink', inlet=ar.Flow('Heat'))

# Run the optimization
es.optimize()

SolarData class

Read the input data from a csv-file for one year in hourly resolution (8760 time steps) with pandas.

[1]:

# Import the required packages (jupyter magic only required for jupyter notebooks)
%reload_ext autoreload
%autoreload 2
%matplotlib inline

import pandas as pd
import matplotlib.pyplot as plt
import aristopy as ar

# Get data from csv-file
data = pd.read_csv('testdata.csv', sep=';', decimal='.', index_col=[0])

The index of the solar data needs be of type DateTimeIndex and must contain information about the time zone. The Data from the reference “Deutscher Wetterdienst” (DWD) is provided with a UTC timestamp. The same is true for the Photovoltaic Geographical Information System (PVGIS) data of the JRC. So, we convert the index to a datetime and localize it to the UTC time zone. Alternatively, we could use the pandas method date_range to create a new index with time zone information included.

[2]:

data.index = pd.to_datetime(
    data.index, format="%d.%m.%Y %H:%M").tz_localize(tz='UTC')

[3]:

# print the time zone of the data index
data.index.tz

[3]:

<UTC>

[4]:

# print first rows
data.head(3)

[4]:
T_amb [C] DHI [W/m2] GHI [W/m2]
2018-01-01 00:00:00+00:00 11.3 0.0 0.0
2018-01-01 01:00:00+00:00 10.9 0.0 0.0
2018-01-01 02:00:00+00:00 11.1 0.0 0.0

A SolarData class instance is created with provided specifications for the location and irradiance data. The global (GHI) and diffuse (DHI) horizontal irradiation time series are required input arguments. The direct normal (beam) irradiation (DNI) can also be specified or is internally calculated based on GHI, DHI, and solar positions. The selected location is Potsdam/Germany. Later, the SolarData instance is used to calculate and return values for all irradiance time series get_irradiance_dataframe, and the plane of array irradiance (POA) get_plane_of_array_irradiance.

[5]:

solar = ar.SolarData(ghi=data['GHI [W/m2]'], dhi=data['DHI [W/m2]'],
                     latitude=52.3822, longitude=13.0622, altitude=81)

PVSystem class

First, we calculate the direct normal irradiation (DNI) from GHI, DHI, and solar positions of the respective location, and return the data in a pandas DataFrame.

[6]:

df_solar = solar.get_irradiance_dataframe()

If available, ambient temperature and wind speed data can be appended to the DataFrame. In this case, pvlib requires the column names: ‘temp_air’ and ‘wind_speed’. This step is optional. Default values 20°C and 0 m/s are used, if no values are provided.

[7]:

df_solar['temp_air'] = data['T_amb [C]']

[8]:

# Print data of the first day
df_solar.iloc[7:17]

[8]:
ghi dhi dni temp_air
2018-01-01 07:00:00+00:00 0.000000 0.000000 -0.000000 8.7
2018-01-01 08:00:00+00:00 25.000000 22.222222 38.988546 7.8
2018-01-01 09:00:00+00:00 88.888889 58.333333 185.767900 7.7
2018-01-01 10:00:00+00:00 147.222222 80.555556 293.940748 8.0
2018-01-01 11:00:00+00:00 205.555556 72.222222 526.992007 8.4
2018-01-01 12:00:00+00:00 180.555556 97.222222 345.532381 9.1
2018-01-01 13:00:00+00:00 66.666667 61.111111 28.910901 8.6
2018-01-01 14:00:00+00:00 58.333333 44.444444 126.548503 7.9
2018-01-01 15:00:00+00:00 22.222222 19.444444 0.000000 7.3
2018-01-01 16:00:00+00:00 0.000000 0.000000 -0.000000 7.5

The next step is to create an instance of class PVSystem by selecting a type module and inverter from the pvlib’s database. The full database currently consists of more than 20,000 modules and 3,000 inverters. To see the database you can either go through the CSV-files in the “data” directory of your pvlib installation, or use the method retrieve_sam.

[9]:

import pvlib

pvlib.pvsystem.retrieve_sam(name='cecmod')  # for inverters: name='cecinverter'

[9]:
A10Green_Technology_A10J_S72_175 A10Green_Technology_A10J_S72_180 A10Green_Technology_A10J_S72_185 A10Green_Technology_A10J_M60_220 A10Green_Technology_A10J_M60_225 A10Green_Technology_A10J_M60_230 A10Green_Technology_A10J_M60_235 A10Green_Technology_A10J_M60_240 A2Peak_Power_POWER_ON_P220_6x10 Aavid_Solar_ASMS_165P ... Zytech_Solar_ZT275P Zytech_Solar_ZT280P Zytech_Solar_ZT285P Zytech_Solar_ZT290P Zytech_Solar_ZT295P Zytech_Solar_ZT300P Zytech_Solar_ZT305P Zytech_Solar_ZT310P Zytech_Solar_ZT315P Zytech_Solar_ZT320P
Technology Mono-c-Si Mono-c-Si Mono-c-Si Multi-c-Si Multi-c-Si Multi-c-Si Multi-c-Si Multi-c-Si Multi-c-Si Multi-c-Si ... Multi-c-Si Multi-c-Si Multi-c-Si Multi-c-Si Multi-c-Si Multi-c-Si Multi-c-Si Multi-c-Si Multi-c-Si Multi-c-Si
Bifacial 0 0 0 0 0 0 0 0 0 0 ... 0 0 0 0 0 0 0 0 0 0
STC 175.091 179.928 184.702 219.876 224.986 230.129 235.008 240.538 219.978 164.85 ... 275.014 280.329 285.326 290.036 295.066 300.003 305.056 310.144 315.094 320.42
PTC 151.2 155.7 160.2 189.1 193.5 204.1 208.7 213.3 195 146.3 ... 248 252.6 257.3 261.9 266.5 271.2 275.8 280.5 285.1 289.8
A_c 1.3 1.3 1.3 1.624 1.624 1.624 1.624 1.624 1.633 1.301 ... 1.931 1.931 1.931 1.931 1.931 1.931 1.931 1.931 1.931 1.931
Length 1.576 1.576 1.576 1.632 1.632 1.632 1.632 1.632 1.633 1.575 ... 1.95 1.95 1.95 1.95 1.95 1.95 1.95 1.95 1.95 1.95
Width 0.825 0.825 0.825 0.995 0.995 0.995 0.995 0.995 1 0.826 ... 0.99 0.99 0.99 0.99 0.99 0.99 0.99 0.99 0.99 0.99
N_s 72 72 72 60 60 60 60 60 60 72 ... 72 72 72 72 72 72 72 72 72 72
I_sc_ref 5.17 5.31 5.43 7.95 8.04 8.1 8.23 8.32 7.98 5.25 ... 8.31 8.4 8.48 8.55 8.64 8.71 8.87 8.9 9.01 9.12
V_oc_ref 43.99 44.06 44.14 36.06 36.24 36.42 36.72 36.84 36.72 43.5 ... 45.1 45.25 45.43 45.59 45.75 45.96 46.12 46.28 46.44 46.6
I_mp_ref 4.78 4.9 5.03 7.3 7.44 7.58 7.68 7.83 7.26 4.71 ... 7.76 7.87 7.97 8.07 8.16 8.26 8.36 8.46 8.56 8.66
V_mp_ref 36.63 36.72 36.72 30.12 30.24 30.36 30.6 30.72 30.3 35 ... 35.44 35.62 35.8 35.94 36.16 36.32 36.49 36.66 36.81 37
alpha_sc 0.002146 0.002204 0.002253 0.004357 0.004406 0.007857 0.007983 0.00807 0.00399 0.001575 ... 0.004014 0.004057 0.004096 0.00413 0.004173 0.004207 0.004284 0.004299 0.004352 0.004405
beta_oc -0.159068 -0.159321 -0.15961 -0.130681 -0.131334 -0.130748 -0.131825 -0.132256 -0.12852 -0.170955 ... -0.144275 -0.144755 -0.145331 -0.145842 -0.146354 -0.147026 -0.147538 -0.14805 -0.148562 -0.149073
T_NOCT 49.9 49.9 49.9 50.2 50.2 46.4 46.4 46.4 47.9 45 ... 46.4 46.4 46.4 46.4 46.4 46.4 46.4 46.4 46.4 46.4
a_ref 1.9817 1.98841 1.98482 1.67309 1.67178 1.68048 1.69698 1.69423 1.59703 1.96463 ... 1.81027 1.81485 1.8201 1.82278 1.83125 1.84441 1.84915 1.8574 1.86502 1.87378
I_L_ref 5.1757 5.31615 5.43568 7.95906 8.04721 8.10361 8.23464 8.32177 8.00023 5.27415 ... 8.32377 8.41015 8.4867 8.55196 8.64154 8.80531 8.87402 8.9948 9.10661 9.21845
I_o_ref 1.14916e-09 1.22524e-09 1.16164e-09 3.34415e-09 3.01424e-09 3.09549e-09 3.24284e-09 2.97878e-09 7.85133e-10 1.19571e-09 ... 1.24062e-10 1.2341e-10 1.21696e-10 1.17172e-10 1.21851e-10 1.31413e-10 1.30106e-10 1.34888e-10 1.38664e-10 1.44659e-10
R_s 0.316688 0.299919 0.311962 0.140393 0.14737 0.152058 0.151504 0.150077 0.229644 0.595855 ... 0.566493 0.552584 0.543536 0.538499 0.521134 0.515735 0.506611 0.495904 0.488376 0.475581
R_sh_ref 287.102 259.048 298.424 123.168 164.419 340.983 268.701 706.27 90.5774 129.523 ... 341.758 457.468 687.561 2348.68 2917.76 552.455 1119.07 767.958 682.292 604.221
Adjust 16.0571 16.419 15.6882 21.8752 20.6984 21.5544 21.8719 20.881 12.2172 7.16388 ... 5.42178 5.27464 5.06509 4.66051 4.9013 5.41555 5.24155 5.44634 5.57874 5.83833
gamma_r -0.5072 -0.5072 -0.5072 -0.5196 -0.5196 -0.493 -0.493 -0.493 -0.46 -0.519 ... -0.4308 -0.4308 -0.4308 -0.4308 -0.4308 -0.4308 -0.4308 -0.4308 -0.4308 -0.4308
BIPV N N N N N N N N N N ... N N N N N N N N N N
Version SAM 2018.11.11 r2 SAM 2018.11.11 r2 SAM 2018.11.11 r2 SAM 2018.11.11 r2 SAM 2018.11.11 r2 SAM 2018.11.11 r2 SAM 2018.11.11 r2 SAM 2018.11.11 r2 SAM 2018.11.11 r2 SAM 2018.11.11 r2 ... SAM 2018.11.11 r2 SAM 2018.11.11 r2 SAM 2018.11.11 r2 SAM 2018.11.11 r2 SAM 2018.11.11 r2 SAM 2018.11.11 r2 SAM 2018.11.11 r2 SAM 2018.11.11 r2 SAM 2018.11.11 r2 SAM 2018.11.11 r2
Date 1/3/2019 1/3/2019 1/3/2019 1/3/2019 1/3/2019 1/3/2019 1/3/2019 1/3/2019 1/3/2019 1/3/2019 ... 1/3/2019 1/3/2019 1/3/2019 1/3/2019 1/3/2019 1/3/2019 1/3/2019 1/3/2019 1/3/2019 1/3/2019

25 rows × 21535 columns

In this example modules and inverter from the manufacturer Canadian Solar are applied. The modules (CS6X-P) use multi-c-Si technololgy and have a nominal power of 300 Wp (STC) and an efficiency of 15.6%. 200 modules are connected to a central inverter that has a nominal AC power of 60 kW.

[10]:

pv_system = ar.PVSystem(module='Canadian_Solar_Inc__CS6X_300P',
                        inverter='Canadian_Solar_Inc___CSI_60KTL_CT__480V_')

To get the PV plant’s electrical power output, we need to provide the irradiance data determined above and the specifications for the orientation of the modules (tilt and azimuth). An azimuth value of 180 represents a surface facing south, a tilt of 0 implies a horizontal module alignment. Furthermore, the electrical output is scaled to its peak power [W/Wp].

[11]:

pv_feed_in = pv_system.get_feedin(
    weather=df_solar, location=solar.location,
    surface_tilt=30, surface_azimuth=180,  # South
    modules_per_string=20, strings_per_inverter=10, scaling='peak_power')  # [W/Wp]

[12]:

# Plot the PV feed-in (scaled to peak power)
pv_feed_in.plot()
plt.ylabel('Scaled PV Feed-in [W/Wp]')
plt.show()
_images/solar_components_22_0.png

SolarThermalCollector class

The solar-thermal collector requires data for the plane of array irradiance (POA). The POA combines the direct normal (DNI) irradiance with sky diffuse and ground-reflected irradiance components and is returned as an OrderedDict or DataFrame via method get_plane_of_array_irradiance. The method needs information about the array’s orientation (tilt and azimuth) as input arguments. Additional keyword arguments (e.g., “albedo”) can be specified via kwargs.

[13]:

poa = solar.get_plane_of_array_irradiance(
    surface_tilt=45, surface_azimuth=180)

[14]:

# Print the components of POA data
poa.iloc[7:17]

[14]:
poa_global poa_direct poa_diffuse poa_sky_diffuse poa_ground_diffuse
2018-01-01 07:00:00+00:00 0.000000 -0.000000 0.000000 0.000000 0.000000
2018-01-01 08:00:00+00:00 41.916672 22.033527 19.883144 18.967853 0.915291
2018-01-01 09:00:00+00:00 186.436000 133.391016 53.044984 49.790614 3.254369
2018-01-01 10:00:00+00:00 315.071214 240.922697 74.148516 68.758468 5.390049
2018-01-01 11:00:00+00:00 523.571366 454.400115 69.171251 61.645523 7.525729
2018-01-01 12:00:00+00:00 380.848212 291.253417 89.594795 82.984357 6.610437
2018-01-01 13:00:00+00:00 76.660323 22.057950 54.602373 52.161596 2.440777
2018-01-01 14:00:00+00:00 119.564815 79.493429 40.071386 37.935706 2.135680
2018-01-01 15:00:00+00:00 17.410464 0.000000 17.410464 16.596871 0.813592
2018-01-01 16:00:00+00:00 0.000000 -0.000000 0.000000 0.000000 0.000000

Set up a solar collector array (type: evacuated tube collectors). The internal procedure to calculate the collector thermal energy output is adopted from reference: V.Quaschning, ‘Regenerative Energiesysteme’, 10th edition, Hanser, 2019, p.131ff. Characteristic values for the parametrization of different collector types are also indicated in this reference.

[15]:

solar_collector_data = ar.SolarThermalCollector(
    optical_efficiency=0.80, thermal_loss_parameter_1=1.1,
    thermal_loss_parameter_2=0.008, irradiance_data=poa['poa_global'],
    t_ambient=data['T_amb [C]'], t_collector_in=60, t_collector_out=90)

The thermal energy output of the solar collector is calculated and returned by method get_collector_heat_output.

[16]:

solar_heat = solar_collector_data.get_collector_heat_output()

[17]:

solar_heat[7:17]

[17]:

2018-01-01 07:00:00+00:00      0.000000
2018-01-01 08:00:00+00:00      0.000000
2018-01-01 09:00:00+00:00     38.884480
2018-01-01 10:00:00+00:00    142.444971
2018-01-01 11:00:00+00:00    310.112613
2018-01-01 12:00:00+00:00    197.446090
2018-01-01 13:00:00+00:00      0.000000
2018-01-01 14:00:00+00:00      0.000000
2018-01-01 15:00:00+00:00      0.000000
2018-01-01 16:00:00+00:00      0.000000
dtype: float64

The data section above shows that despite present irradiation on the collector in the hours 1 pm to 3 pm, no useful thermal energy is released. That is explainable by the high share of heat losses at large temperature differences and a low total irradiance. The following figure shows that depending on the outside temperature, the collector efficiency can even drop to zero. Thus, no further thermal energy is extracted.

[18]:

plt.scatter(poa['poa_global'], solar_collector_data.collector_efficiency*100)
plt.xlabel('Global POA [W/m²]'), plt.ylabel('Solar collector efficiency [%]')
plt.show()
_images/solar_components_32_0.png 
[19]:

# Plot the specific solar-thermal heat output
solar_heat.plot()
plt.ylabel('Solar-thermal heat output [W/m²]')
plt.show()
_images/solar_components_33_0.png

Simple solar model

The following simple model exemplarily shows how to use the generated feed-in time series of the PVSystem and the SolarThermalCollector classes in an aristopy model. The model could easily be extended, e.g., by adding a heat pump or a (seasonal) thermal storage system.

[20]:

# Create energy system instance
es = ar.EnergySystem()

The feed-in time series of the PV system can be introduced directly to the model with an electricity providing instance of the Source class. Since the applied time series is scaled to the peak power value, we need to multiply this relative value with the overall capacity of the PV plant in a user expression.

[21]:

# Add a photovoltaic component (electricity source)
pv = ar.Source(
    ensys=es, name='pv', outlet=ar.Flow('Elec', 'elec_sink'),
    time_series_data=ar.Series('pv_feed_in', pv_feed_in),  # [MW/MWp]
    capacity=100, capex_per_capacity=700e3, opex_per_capacity=0.025*700e3,  # [MWp]
    user_expressions='Elec == CAP * pv_feed_in')

The Source class is also used to impose the solar-thermal heat output. The available time-series provides area-specific data, so the delivered thermal energy is calculated as the product of the time series values and the collector area. The area is introduced as an additional variable without time dependency via the “additional_vars” keyword. Since also the costs of the plant are area-related [EUR/m²], the Area varibale is defined as the basic variable of the component (see Component class API).

[22]:

# Add a Solar-thermal collector component (heat source)
solar_collector = ar.Source(
    ensys=es, name='solar_collector', outlet=ar.Flow('Heat', 'heat_sink'),
    basic_variable='Area', additional_vars=ar.Var('Area', has_time_set=False),
    time_series_data=ar.Series('collector_feedin', solar_heat / 1e6),  # [MW/m²]
    user_expressions='Heat == collector_feedin * Area',
    capacity=1e5, capex_per_capacity=300, opex_per_capacity=300*0.01)  # [m²], [EUR/m²]

# Add sinks for electricity and heat
elec_sink = ar.Sink(ensys=es, name='elec_sink', inlet=ar.Flow('Elec'))
heat_sink = ar.Sink(ensys=es, name='heat_sink', inlet=ar.Flow('Heat'))

# Run the optimization
es.optimize(tee=False)

[23]:

# Plot PV feed-in to electricity sink
plotter = ar.Plotter(json_file='results.json')
plotter.plot_operation('elec_sink', 'Elec', show_plot=True,
                       ylabel='Photovoltaic electricity production [MWh]')
_images/solar_components_40_0.png 
[24]:

print(f'Capacity PV system:    {pv.block.CAP.value} MW')
print(f'Electricity PV system: {sum(pv.block.Elec.get_values().values())} MWh/a')

Capacity PV system:    100.0 MW
Electricity PV system: 133830.5074627531 MWh/a

[25]:

# Plot solar collector feed-in to heat sink
plotter.plot_operation('heat_sink', 'Heat', show_plot=True,
                       ylabel='Solar-thermal heat production [MWh]')
_images/solar_components_42_0.png 
[26]:

print(f'Area solar collector:  {solar_collector.block.Area.value} m²')
print(f'Heat solar collector:  {sum(solar_collector.block.Heat.get_values().values())} MWh/a')

Area solar collector:  100000.0 m²
Heat solar collector:  78976.54196102788 MWh/a