workflows

Functions:

• CSP_PTR_ERA5 –

  This function is the overall workflow for the csp simulation and calls all subfunctions.

• CSP_PTR_ERA5_specific_dataset –

  Calculates the heat output from the solar field based on parabolic trough technology (PTC).

CSP_PTR_ERA5

CSP_PTR_ERA5(
    placements,
    era5_path,
    global_solar_atlas_dni_path,
    global_solar_atlas_tamb_path=None,
    datasets=None,
    cost_year=2050,
    HTF_sel=["Heliosol", "SolarSalt", "Therminol"],
    elev_path=None,
    output_netcdf_path=None,
    output_variables=None,
    return_self=True,
    JITaccelerate=False,
    verbose=False,
    debug_vars=False,
    onlynightuse=True,
    fullvariation=False,
    _validation=False,
    time_slice=None,
)

This function is the overall workflow for the csp simulation and calls all subfunctions. It is a wrapper around the function "CSP_PTR_ERA5_specific_dataset" below, to include the case of multiple datasets. Multiple datasets are refering to multiple heat transfer fluids and therefore also different power plants. In that case, for each placement the optimal HTF is selected.

Parameters:

• placements

  –

  Locations where to perform simulations at. Required columns: longitude, latitude, and one area column ("area_m2", "area", "aperture_area_m2", or "land_area_m2").

• era5_path

  –

  Path to ERA5 weather data or an NCSource object. Required columns: "direct_horizontal_irradiance", "surface_wind_speed", "surface_air_temperature".

• global_solar_atlas_dni_path

  –

  Long-term average DNI data for placements. Could be provided as: - str: Path to raster file with DNI data from Global Solar Atlas. - float: Constant DNI value for all placements. - np.ndarray: Array of DNI values, shape must match (placements,) or (time, placements).

• global_solar_atlas_tamb_path

  –

  path to temperature data from global solar atlas. Used for selecting at which placement which HTF is optimal.

• datasetname

  –

  Name of the CSP technology dataset (e.g., "Heliosol", "SolarSalt"). Defaults to "Validation 10". See csp/data/CSP_database.xlsx for options.

• cost_year

  –

  Year for cost calculation of the CSP plant. Defaults to 2050.

• HTF_sel

  –

  List of heat transfer fluids to consider for the simulation. Each heat transfer fluid corresponds to a different CSP technology dataset. Options: ["Heliosol", "SolarSalt", "Therminol"].

• elev_path

  –

  Elevation data for placements. Could be provided as: - str: Path to raster file with elevation data. - list: Elevation values for each placement.

• output_netcdf_path

  –

  Path to a file that you want to save your output NETCDF file at. Defaults to None.

• output_variables

  –

  Output variables of the simulation that you want to save. If None, includes all suitable variables from placements, workflow parameters, simulation data, and time index.

• return_self

  –

  If True, returns the workflow manager object. If False, returns the output as an XArray dataset. Defaults to True.

• JITaccelerate

  –

  If True, enables JIT acceleration for some calculations. Defaults to False.

• verbose

  –

  If True, prints progress information. Defaults to False.

• debug_vars

  –

  If True, retains intermediate variables for debugging. Defaults to False.

• onlynightuse

  –

  If True, optimizes plant size for night use only. Defaults to True.

• fullvariation

  –

  If True, applies full variation in calculations. Defaults to False.

• _validation

  –

  If True, runs in validation mode (some input data may be replaced). Defaults to False.

• time_slice

  (slice, default: None ) –

  Limit the time span loaded from the ERA5 source. Only supported for Zarr-backed ERA5 sources, where it is strongly recommended to avoid loading whole multi-year cloud stores. Raises for netCDF4-backed ERA5 sources; support for those is planned.

Returns:

• if return_self is True, the workflow manager object (PTRWorkflowManager) containing all simulation results and data. –
• if return_self is False, a xarray dataset containing the final simulation results. This dataset is optionally written to disk if output_netcdf_path was specified. –

  rest: see CSP_PTR_ERA5_specific_dataset

Source code in reskit/csp/workflows/workflows.py

def CSP_PTR_ERA5(
    placements,
    era5_path,
    global_solar_atlas_dni_path,
    global_solar_atlas_tamb_path=None,
    datasets=None,
    cost_year=2050,
    HTF_sel=["Heliosol", "SolarSalt", "Therminol"],
    elev_path=None,
    output_netcdf_path=None,
    output_variables=None,
    return_self=True,
    JITaccelerate=False,
    verbose=False,
    debug_vars=False,
    onlynightuse=True,
    fullvariation=False,
    _validation=False,
    time_slice=None,
):
    """
    This function is the overall workflow for the csp simulation and calls all subfunctions.
    It is a wrapper around the function "CSP_PTR_ERA5_specific_dataset" below, to include the case of multiple datasets.
    Multiple datasets are refering to multiple heat transfer fluids and therefore also different power plants. In that case, for each placement the optimal HTF is selected.

    Parameters
    ----------
    placements: Pandas DataFrame
        Locations where to perform simulations at.
        Required columns: longitude, latitude, and one area column ("area_m2", "area", "aperture_area_m2", or "land_area_m2").

    era5_path: str or rk_weather.NCSource
        Path to ERA5 weather data or an NCSource object.
        Required columns: "direct_horizontal_irradiance", "surface_wind_speed", "surface_air_temperature".

    global_solar_atlas_dni_path: str or float or np.ndarray
        Long-term average DNI data for placements. Could be provided as:
            - str: Path to raster file with DNI data from Global Solar Atlas.
            - float: Constant DNI value for all placements.
            - np.ndarray: Array of DNI values, shape must match (placements,) or (time, placements).

    global_solar_atlas_tamb_path: str, optional
        path to temperature data from global solar atlas. Used for selecting at which placement which HTF is optimal.

    datasetname: str, optional
        Name of the CSP technology dataset (e.g., "Heliosol", "SolarSalt").
        Defaults to "Validation 10". See csp/data/CSP_database.xlsx for options.

    cost_year: int, optional
        Year for cost calculation of the CSP plant.
        Defaults to 2050.

    HTF_sel: list of str, optional
        List of heat transfer fluids to consider for the simulation.
        Each heat transfer fluid corresponds to a different CSP technology dataset.
        Options: ["Heliosol", "SolarSalt", "Therminol"].

    elev_path: str or list, optional
        Elevation data for placements. Could be provided as:
            - str: Path to raster file with elevation data.
            - list: Elevation values for each placement.

    output_netcdf_path: str, optional
        Path to a file that you want to save your output NETCDF file at.
        Defaults to None.

    output_variables: list of {str, number}, optional
        Output variables of the simulation that you want to save.
        If None, includes all suitable variables from placements, workflow parameters, simulation data, and time index.

    return_self: bool, optional
        If True, returns the workflow manager object.
        If False, returns the output as an XArray dataset.
        Defaults to True.

    JITaccelerate: bool, optional
        If True, enables JIT acceleration for some calculations.
        Defaults to False.

    verbose: bool, optional
        If True, prints progress information.
        Defaults to False.

    debug_vars: bool, optional
        If True, retains intermediate variables for debugging.
        Defaults to False.

    onlynightuse: bool, optional
        If True, optimizes plant size for night use only.
        Defaults to True.

    fullvariation: bool, optional
        If True, applies full variation in calculations.
        Defaults to False.

    _validation: bool, optional
        If True, runs in validation mode (some input data may be replaced).
        Defaults to False.

    time_slice : slice, optional
        Limit the time span loaded from the ERA5 source. Only supported for
        Zarr-backed ERA5 sources, where it is strongly recommended to avoid
        loading whole multi-year cloud stores. Raises for netCDF4-backed ERA5
        sources; support for those is planned.

    Returns
    -------
    if return_self is True, the workflow manager object (PTRWorkflowManager) containing all simulation results and data.
    if return_self is False, a xarray dataset containing the final simulation results. This dataset is optionally written to disk if output_netcdf_path was specified.

        rest: see CSP_PTR_ERA5_specific_dataset
    """
    # handle inputs for datasets
    single_dataset = False
    if datasets == None:
        # get datasets from HTF_sel and cost_year
        datasets = ["Dataset_" + htf + "_" + str(cost_year) for htf in HTF_sel]
        if len(datasets) == 1:
            single_dataset = True
            datasets = datasets[0]
    elif isinstance(datasets, str):
        # datasets=datasets
        single_dataset = True
    elif isinstance(datasets, list):
        # datasets=datasets
        if len(datasets) == 1:
            single_dataset = True
            datasets = datasets[0]
    else:
        raise TypeError(f"datasets got unknown datatype")

    if not single_dataset:
        assert isinstance(global_solar_atlas_tamb_path, str)

    if single_dataset:  # only one dataset given
        output = CSP_PTR_ERA5_specific_dataset(
            placements=placements,
            era5_path=era5_path,
            global_solar_atlas_dni_path=global_solar_atlas_dni_path,
            datasetname=datasets,
            elev_path=elev_path,
            output_netcdf_path=output_netcdf_path,
            output_variables=output_variables,
            return_self=return_self,
            JITaccelerate=JITaccelerate,
            verbose=verbose,
            debug_vars=debug_vars,
            onlynightuse=onlynightuse,
            fullvariation=fullvariation,
            _validation=_validation,
            time_slice=time_slice,
        )
        return output

    else:  # multiple datasets found
        # 1) split up placements for each htf (e.g. solar salt or heliosol)
        d = dataset_handler(datasets)
        placements = d.split_placements(
            placements=placements,
            gsa_dni_path=global_solar_atlas_dni_path,
            gsa_tamb_path=global_solar_atlas_tamb_path,
        )
        del d

        # 2) run each simulation
        outputs = []
        for dataset in datasets:
            # select placements for current dataset
            placements_dataset = placements[placements["Dataset_opt"] == dataset]

            # skip empty batches
            if len(placements_dataset) == 0:
                continue

            # starting core simulation
            output_dataset = CSP_PTR_ERA5_specific_dataset(
                placements=placements_dataset,
                era5_path=era5_path,
                global_solar_atlas_dni_path=global_solar_atlas_dni_path,
                datasetname=dataset,
                elev_path=elev_path,
                output_netcdf_path=output_netcdf_path,
                output_variables=output_variables,
                return_self=False,
                JITaccelerate=JITaccelerate,
                verbose=verbose,
                debug_vars=debug_vars,
                onlynightuse=onlynightuse,
                fullvariation=fullvariation,
                _validation=_validation,
                time_slice=time_slice,
            )

            # add dataset to column (force output this!)
            if not "datasetname" in output_dataset.variables:
                output_dataset["datasetname"] = (output_dataset["lon"] * 0).astype(str)
                output_dataset["datasetname"][:] = dataset
            if "Dataset_opt" in output_dataset.variables:
                output_dataset = output_dataset.drop("Dataset_opt")

            # set index from placements
            output_dataset["location"] = placements_dataset.index

            # remember outputs for each dataset
            outputs.append(output_dataset)
            del output_dataset

        # 3) merge data
        # TODO: merge together
        output = xr.concat(outputs, dim="location").sortby("location")
        return output

CSP_PTR_ERA5_specific_dataset

CSP_PTR_ERA5_specific_dataset(
    placements,
    era5_path,
    global_solar_atlas_dni_path,
    datasetname="Validation 10",
    elev_path=None,
    output_netcdf_path=None,
    output_variables=None,
    return_self=True,
    JITaccelerate=False,
    verbose=False,
    debug_vars=False,
    onlynightuse=True,
    fullvariation=False,
    _validation=False,
    time_slice=None,
)

Calculates the heat output from the solar field based on parabolic trough technology (PTC). This workflow simulates the performance of a CSP plant for given placements and technology datasets. Not yet fully validated, but results are plausible.

Date: 27.07.2021 Author: David Franzmann, IEK-3

Parameters:

• placements

  –

  Locations where to perform simulations at. Required columns: longitude, latitude, and one area column ("area_m2", "area", "aperture_area_m2", or "land_area_m2").

• era5_path

  –

  Path to ERA5 weather data or an NCSource object. Required columns: "direct_horizontal_irradiance", "surface_wind_speed", "surface_air_temperature".

• global_solar_atlas_dni_path

  –

  Long-term average DNI data for placements. Could be provided as: - str: Path to raster file with DNI data from Global Solar Atlas. - float: Constant DNI value for all placements. - np.ndarray: Array of DNI values, shape must match (placements,) or (time, placements).

• datasetname

  –

  Name of the CSP technology dataset (e.g., "Heliosol", "SolarSalt"). Defaults to "Validation 10". See csp/data/CSP_database.xlsx for options.

• elev_path

  –

  Elevation data for placements. Could be provided as: - str: Path to raster file with elevation data. - list: Elevation values for each placement.

• output_netcdf_path

  –

  Path to a file that you want to save your output NETCDF file at. Defaults to None.

• output_variables

  –

  Output variables of the simulation that you want to save. If None, includes all suitable variables from placements, workflow parameters, simulation data, and time index.

• return_self

  –

  If True, returns the workflow manager object. If False, returns the output as an XArray dataset. Defaults to True.

• JITaccelerate

  –

  If True, enables JIT acceleration for some calculations. Defaults to False.

• verbose

  –

  If True, prints progress information. Defaults to False.

• debug_vars

  –

  If True, retains intermediate variables for debugging. Defaults to False.

• onlynightuse

  –

  If True, optimizes plant size for night use only. Defaults to True.

• fullvariation

  –

  If True, applies full variation in calculations. Defaults to False.

• _validation

  –

  If True, runs in validation mode (some input data may be replaced). Defaults to False.

• time_slice

  (slice, default: None ) –

  Limit the time span loaded from the ERA5 source. Only supported for Zarr-backed ERA5 sources, where it is strongly recommended to avoid loading whole multi-year cloud stores. Raises for netCDF4-backed ERA5 sources; support for those is planned.

Returns:

• if return_self is True, the workflow manager object (PTRWorkflowManager) containing all simulation results and data. –
• if return_self is False, a xarray dataset containing the final simulation results. This dataset is optionally written to disk if output_netcdf_path was specified. –

Source code in reskit/csp/workflows/workflows.py

def CSP_PTR_ERA5_specific_dataset(
    placements,
    era5_path,
    global_solar_atlas_dni_path,
    datasetname="Validation 10",
    elev_path=None,
    output_netcdf_path=None,
    output_variables=None,
    return_self=True,
    JITaccelerate=False,
    verbose=False,
    debug_vars=False,
    onlynightuse=True,
    fullvariation=False,
    _validation=False,
    time_slice=None,
):
    """
    Calculates the heat output from the solar field based on parabolic trough technology (PTC).
    This workflow simulates the performance of a CSP plant for given placements and technology datasets.
    Not yet fully validated, but results are plausible.

    Date: 27.07.2021
    Author: David Franzmann, IEK-3

    Parameters
    ----------
    placements: Pandas DataFrame
        Locations where to perform simulations at.
        Required columns: longitude, latitude, and one area column ("area_m2", "area", "aperture_area_m2", or "land_area_m2").

    era5_path: str or rk_weather.NCSource
        Path to ERA5 weather data or an NCSource object.
        Required columns: "direct_horizontal_irradiance", "surface_wind_speed", "surface_air_temperature".

    global_solar_atlas_dni_path: str or float or np.ndarray
        Long-term average DNI data for placements. Could be provided as:
            - str: Path to raster file with DNI data from Global Solar Atlas.
            - float: Constant DNI value for all placements.
            - np.ndarray: Array of DNI values, shape must match (placements,) or (time, placements).

    datasetname: str, optional
        Name of the CSP technology dataset (e.g., "Heliosol", "SolarSalt").
        Defaults to "Validation 10". See csp/data/CSP_database.xlsx for options.

    elev_path: str or list, optional
        Elevation data for placements. Could be provided as:
            - str: Path to raster file with elevation data.
            - list: Elevation values for each placement.

    output_netcdf_path: str, optional
        Path to a file that you want to save your output NETCDF file at.
        Defaults to None.

    output_variables: list of {str, number}, optional
        Output variables of the simulation that you want to save.
        If None, includes all suitable variables from placements, workflow parameters, simulation data, and time index.

    return_self: bool, optional
        If True, returns the workflow manager object.
        If False, returns the output as an XArray dataset.
        Defaults to True.

    JITaccelerate: bool, optional
        If True, enables JIT acceleration for some calculations.
        Defaults to False.

    verbose: bool, optional
        If True, prints progress information.
        Defaults to False.

    debug_vars: bool, optional
        If True, retains intermediate variables for debugging.
        Defaults to False.

    onlynightuse: bool, optional
        If True, optimizes plant size for night use only.
        Defaults to True.

    fullvariation: bool, optional
        If True, applies full variation in calculations.
        Defaults to False.

    _validation: bool, optional
        If True, runs in validation mode (some input data may be replaced).
        Defaults to False.

    time_slice : slice, optional
        Limit the time span loaded from the ERA5 source. Only supported for
        Zarr-backed ERA5 sources, where it is strongly recommended to avoid
        loading whole multi-year cloud stores. Raises for netCDF4-backed ERA5
        sources; support for those is planned.

    Returns
    -------
    if return_self is True, the workflow manager object (PTRWorkflowManager) containing all simulation results and data.
    if return_self is False, a xarray dataset containing the final simulation results. This dataset is optionally written to disk if output_netcdf_path was specified.

    """
    # 1) Load input data
    wf = PTRWorkflowManager(placements)

    ptr_data = wf.loadPTRdata(
        datasetname=datasetname
    )  # PTRdata referes to the different csp models, i.e. helisol or solar salt: csp/data/CSP_database.xlsx
    wf.determine_area()  # either determines aperture_area from land_area, or the other way around

    # 3) read in Input data
    if verbose:
        tic_start = time.time()
        print(
            "Simulation started for {n_placements} placements.".format(n_placements=len(wf.placements)),
            flush=True,
        )
        print("Reading in Weather data.", flush=True)
    wf.read(
        variables=[  # "global_horizontal_irradiance",
            "direct_horizontal_irradiance",
            "surface_wind_speed",
            # "surface_pressure",
            "surface_air_temperature",
        ],
        source_type="ERA5",
        source=era5_path,
        set_time_index=True,
        time_index_from="direct_horizontal_irradiance",
        time_slice=time_slice,
        verbose=False,
    )

    wf.check_ERA5_input()

    if verbose:
        tic_read = time.time()
        print(
            "Data read in within {dt}s.".format(dt=str(tic_read - tic_start)),
            flush=True,
        )
        print("Starting preanalysis.", flush=True)
    # 4) get length of timesteps for later numpy sizing

    wf.get_timesteps()

    # apply elevation
    wf.apply_elevation(elev_path)
    wf.apply_azimuth()
    # 5) calculate the solar position based on pvlib

    wf.calculateSolarPosition()

    # calculate DNI from ERA5 to DNi convention
    # ERA5 DIN: Heat flux per horizontal plane
    # DNI convention: Heat flux per normal (to zenith) plane
    wf.direct_normal_irradiance_from_trigonometry()

    if _validation:
        # when doing the valitadion, the dni atlas was not finally processed,
        # so it was replaced with 1. As only the plant and not the weather simulation
        # was validated, this does not have an impact, as the same weather input
        # was used
        wf.placements["LRA_factor_direct_normal_irradiance"] = 1
    else:
        wf.adjust_variable_to_long_run_average(
            variable="direct_normal_irradiance",
            source_long_run_average=rk_weather.Era5Source.LONG_RUN_AVERAGE_DNI_2020_03,
            real_long_run_average=global_solar_atlas_dni_path,
            real_lra_scaling=1000 / 24,  # cast to hourly average kWh
            nodata_fallback="source",
        )

    # manipulationof input values for variation calculation
    wf._applyVariation()  # only for developers, can be ignored otherwise

    # 6) doing selfmade calulations until Heat to HTF (Heat transfer fluid)
    wf.calculateIAM(a1=ptr_data["a1"], a2=ptr_data["a2"], a3=ptr_data["a3"])
    wf.calculateShadowLosses(method="wagner2011", SF_density=ptr_data["SF_density_direct"])
    wf.calculateWindspeedLosses(max_windspeed_threshold=ptr_data["maxWindspeed"])
    wf.calculateDegradationLosses(
        efficiencyDropPerYear=ptr_data["efficiencyDropPerYear"],
        lifetime=ptr_data["lifetime"],
    )
    wf.calculateHeattoHTF(
        eta_ptr_max=ptr_data["eta_ptr_max"],
        eta_cleaness=ptr_data["eta_cleaness"],
        eta_other=ptr_data["eta_other"],
    )

    wf.apply_capacity_sf()
    if not debug_vars:
        del (
            wf.sim_data["surface_wind_speed"],
            wf.sim_data["tracking_angle"],
            wf.sim_data["solar_altitude_angle_degree"],
            wf.sim_data["direct_horizontal_irradiance"],
        )
        del (
            wf.sim_data["eta_wind"],
            wf.sim_data["eta_degradation"],
            wf.sim_data["eta_shdw"],
        )

    if verbose:
        tic_pre = time.time()
        print("Preanalysis within {dt}s.".format(dt=str(tic_pre - tic_read)), flush=True)
        print("Starting core simulation of the solar field.", flush=True)
    # 7) calculation heat to plant with loss model
    wf.applyHTFHeatLossModel(
        calculationmethod="dersch2018",
        params={
            "b": ptr_data["b"],
            "relTMplant": ptr_data["relTMplant"],
            "maxHTFTemperature": ptr_data["maxHTFTemperature"],
            "JITaccelerate": JITaccelerate,  # TODO: from ptr manager
            "minHTFTemperature": ptr_data["minHTFTemperature"],
            "inletHTFTemperature": ptr_data["inletHTFTemperature"],
            "add_losses_coefficient": ptr_data["add_losses_coefficient"],
            "discretizationmethod": ptr_data["discretizationmethod"],
        },
    )

    if not debug_vars:
        del (
            wf.sim_data["surface_air_temperature"],
            wf.sim_data["HTF_mean_temperature_C"],
            wf.sim_data["Heat_Losses_W"],
            wf.sim_data["theta"],
            wf.sim_data["IAM"],
        )

    # 8) calculate Parasitic Losses of the plant
    wf.calculateParasitics(
        calculationmethod="dersch2018",  # 'gafurov2013',
        params={
            "PL_sf_fixed_W_per_m^2_ap": 1.486,
            "PL_sf_pumping_W_per_m^2_ap": 8.3,
            "PL_plant_fix": ptr_data["PL_plant_fix"],
            # 'PL_sf_track': ptr_data['PL_sf_track'], #gafurov
            # 'PL_sf_pumping': ptr_data['PL_sf_pumping'], #gafurov
            "PL_plant_pumping": ptr_data["PL_plant_pumping"],
            "PL_plant_other": ptr_data["PL_plant_other"],
        },
    )

    # 9) calculate economics
    wf.calculateEconomics_SolarField(
        WACC=ptr_data["WACC"],
        lifetime=ptr_data["lifetime"],
        calculationmethod="franzmann2021",
        params={
            "CAPEX_solar_field_EUR_per_m^2_aperture": ptr_data["CAPEX_solar_field_EUR_per_m^2_aperture"],
            "CAPEX_land_EUR_per_m^2_land": ptr_data["CAPEX_land_EUR_per_m^2_land"],
            "CAPEX_indirect_cost_perc_CAPEX": ptr_data["CAPEX_indirect_cost_perc_CAPEX"],
            "electricity_price_EUR_per_kWh": ptr_data["electricity_price_EUR_per_kWh"],
            "OPEX_perc_CAPEX": ptr_data["OPEX_perc_CAPEX"],
        },
    )

    if verbose:
        tic_sf_sim = time.time()
        print(
            "Solar field simulation done in {dt}s.".format(dt=str(tic_sf_sim - tic_pre)),
            flush=True,
        )
        print("Starting optimizing plant electric output.", flush=True)

    wf.optimize_plant_size(onlynightuse=onlynightuse, fullvariation=fullvariation, debug_vars=debug_vars)

    # wf.optimize_heat_output_4D()
    # wf.calculateEconomics_Plant_Storage_4D()
    # wf.optimal_Plant_Configuration_4D()

    if verbose:
        tic_opt_plant = time.time()
        print(
            "Optimal Sizing done in  {dt}s.".format(dt=str(tic_opt_plant - tic_sf_sim)),
            flush=True,
        )

    wf.calculate_electrical_output(onlynightuse=onlynightuse, debug_vars=debug_vars)
    wf.calculate_LCOE()
    wf.calculateCapacityFactors()

    if verbose:
        tic_final = time.time()
        print(
            "Total simulation done in {dt}s.".format(dt=str(tic_final - tic_start)),
            flush=True,
        )

    if return_self == True:
        return wf
    else:
        return wf.to_xarray(output_netcdf_path=output_netcdf_path, output_variables=output_variables)