data.datasets.power_plants.wind_farms.generate_wind_farms()

Complexity

Conditions

12

Size

Total Lines	134
Code Lines	72

Duplication

Lines	0
Ratio	0 %

Importance

Changes

0

from matplotlib import pyplot as plt
def generate_wind_farms():
    """Generate wind farms based on existing wind farms.

    Parameters
    ----------
    *No parameters required

    """
    # get config
    cfg = egon.data.config.datasets()["power_plants"]

    # Due to typos in some inputs, some areas of existing wind farms
    # should be discarded using perimeter and area filters
    def filter_current_wf(wf_geometry):
        if wf_geometry.geom_type == "Point":
            return True
        if wf_geometry.geom_type == "Polygon":
            area = wf_geometry.area
            length = wf_geometry.length
            # Filter based on the biggest (# of WT) wind farm
            return (area < 40000000) & (length < 40000)
        if wf_geometry.geom_type == "LineString":
            length = wf_geometry.length
            return length < 1008  # 8 * rotor diameter (8*126m)

    # The function 'wind_farm' returns the connection point of a wind turbine
    def wind_farm(x):
        try:
            return map_ap_wea_farm[x]
        except:
            return np.nan

    # The function 'voltage' returns the voltage level a wind turbine operates
    def voltage(x):
        try:
            return map_ap_wea_voltage[x]
        except:
            return np.nan

    # Connect to the data base
    con = db.engine()
    sql = "SELECT geom FROM supply.egon_re_potential_area_wind"
    # wf_areas has all the potential areas geometries for wind farms
    wf_areas = gpd.GeoDataFrame.from_postgis(sql, con)
    # bus has the connection points of the wind farms
    bus = pd.read_csv(
        WORKING_DIR_MASTR_NEW / cfg["sources"]["mastr_location"],
        index_col="MaStRNummer",
    )
    # Drop all the rows without connection point
    bus.dropna(subset=["NetzanschlusspunktMastrNummer"], inplace=True)
    # wea has info of each wind turbine in Germany.
    wea = pd.read_csv(WORKING_DIR_MASTR_NEW / cfg["sources"]["mastr_wind"])

    # Delete all the rows without information about geographical location
    wea = wea[(pd.notna(wea["Laengengrad"])) & (pd.notna(wea["Breitengrad"]))]
    # Delete all the offshore wind turbines
    wea = wea.loc[wea["Lage"] == "Windkraft an Land"]
    # the variable map_ap_wea_farm have the connection point of all the
    # available wt in the dataframe bus.
    map_ap_wea_farm = {}
    map_ap_wea_voltage = {}

    wea["connection point"] = wea["LokationMastrNummer"].map(
        bus["NetzanschlusspunktMastrNummer"]
    )
    wea["voltage"] = wea["LokationMastrNummer"].map(
        bus["NetzanschlusspunktMastrNummer"]
    )

    # Create the columns 'geometry' which will have location of each WT in a
    # point type
    wea = gpd.GeoDataFrame(
        wea,
        geometry=gpd.points_from_xy(
            wea["Laengengrad"], wea["Breitengrad"], crs=4326
        ),
    )

    # wf_size storages the number of WT connected to each connection point
    wf_size = wea["connection point"].value_counts()
    # Delete all the connection points with less than 3 WT
    wf_size = wf_size[wf_size >= 3]
    # Filter all the WT which are not part of a wind farm of at least 3 WT
    wea = wea[wea["connection point"].isin(wf_size.index)]
    # current_wfs has all the geometries that represent the existing wind
    # farms
    current_wfs = gpd.GeoDataFrame(
        index=wf_size.index, crs=4326, columns=["geometry", "voltage"]
    )
    for conn_point, wt_location in wea.groupby("connection point"):
        current_wfs.at[conn_point, "geometry"] = MultiPoint(
            wt_location["geometry"].values
        ).convex_hull
        current_wfs.at[conn_point, "voltage"] = wt_location["voltage"].iat[0]

    current_wfs["geometry2"] = current_wfs["geometry"].to_crs(3035)
    current_wfs["area"] = current_wfs["geometry2"].apply(lambda x: x.area)
    current_wfs["length"] = current_wfs["geometry2"].apply(lambda x: x.length)
    # The 'filter_wts' is used to discard atypical values for the current wind
    # farms
    current_wfs["filter2"] = current_wfs["geometry2"].apply(
        lambda x: filter_current_wf(x)
    )

    # Apply the filter based on area and perimeter
    current_wfs = current_wfs[current_wfs["filter2"]]
    current_wfs = current_wfs.drop(columns=["geometry2", "filter2"])

    wf_areas["area [km²]"] = wf_areas.area / 1000000

    # Exclude areas smaller than X m². X was calculated as the area of
    # 3 WT in the corners of an equilateral triangle with l = 4*rotor_diameter
    min_area = 4 * (0.126**2) * np.sqrt(3)
    wf_areas = wf_areas[wf_areas["area [km²]"] > min_area]

    # Find the centroid of all the suitable potential areas
    wf_areas["centroid"] = wf_areas.centroid

    # find the potential areas that intersects the convex hulls of current
    # wind farms and assign voltage levels
    wf_areas = wf_areas.to_crs(4326)
    for i in wf_areas.index:
        intersection = current_wfs.intersects(wf_areas.at[i, "geom"])
        if intersection.any() == False:
            wf_areas.at[i, "voltage"] = "No Intersection"
        else:
            wf_areas.at[i, "voltage"] = current_wfs[intersection].voltage[0]

    # wf_areas_ni has the potential areas which don't intersect any current
    # wind farm
    wf_areas_ni = wf_areas[wf_areas["voltage"] == "No Intersection"]
    wf_areas = wf_areas[wf_areas["voltage"] != "No Intersection"]
    return wf_areas, wf_areas_ni