Features/#684 distribute pv rooftop buildings (#934)

* added pv rooftop distribution and sanity checks
* changed peak load buildings querry to match #826
* updated pre-commit hooks
* made sure that only grid districts with any buildings are assigned pv rooftop capacity
* ensure that no more pv rooftop capacity is allocated to any mv grid district than there is rooftop potential
* match with new table 'boundaries.egon_map_zensus_mvgd_buildings'
* change dataset version

Co-authored-by: Julian.Endres <julian.endres@rl-institut.de>
Co-authored-by: nesnoj <jonathan.amme@rl-institut.de>

CHANGELOG.rst

@@ -201,6 +201,10 @@ Added
   `#382 <https://github.com/openego/eGon-data/issues/382>`_
 * Add motorized individual travel
   `#553 <https://github.com/openego/eGon-data/issues/553>`_
+* Allocating MaStR PV rooftop power plants to OSM and synthetic buildings.
+  Desaggregating PV rooftop scenarios to mv grid districts and OSM and synthetic
+  buildings.
+  `#684 <https://github.com/openego/eGon-data/issues/684>`_
 * Add mapping zensus - weather cells
   `#845 <https://github.com/openego/eGon-data/issues/845>`_
 * Add pv rooftop plants per mv grid for eGon100RE
@@ -609,7 +613,7 @@ Bug Fixes
 * Fix model load timeseries in motorized individual travel
   `#830 <https://github.com/openego/eGon-data/issues/830>`_
 * Fix gas costs
-  `#830 <https://github.com/openego/eGon-data/issues/847>`_
+  `#847 <https://github.com/openego/eGon-data/issues/847>`_
 * Add imports that have been wrongly deleted
   `#849 <https://github.com/openego/eGon-data/issues/849>`_
 * Fix final demand of heat demand timeseries

setup.py

@@ -106,8 +106,9 @@ def read(*names, **kwargs):
         "rioxarray",
         "rtree",
         "saio",
+        "seaborn",
         "shapely",
-        "snakemake",
+        "snakemake<7",
         "sqlalchemy<1.4",  # Airflow<2.0 is not compatible with SQLAlchemy>=1.4
         "wtforms<3",  # WTForms>=3.0 breaks Airflow<2.0
         "xarray",

src/egon/data/datasets.yml

@@ -346,6 +346,7 @@ hotmaps_current_policy_scenario_heat_demands_buildings:
 
 power_plants:
   sources:
+      mastr_pv: "bnetza_mastr_solar_cleaned.csv"
       mastr_biomass: "bnetza_mastr_biomass_cleaned.csv"
       mastr_hydro: "bnetza_mastr_hydro_cleaned.csv"
       mastr_location: "location_elec_generation_raw.csv"

src/egon/data/datasets/power_plants/__init__.py

@@ -1,15 +1,7 @@
 """The central module containing all code dealing with power plant data.
 """
 from geoalchemy2 import Geometry
-from sqlalchemy import (
-    BigInteger,
-    Boolean,
-    Column,
-    Float,
-    Integer,
-    Sequence,
-    String,
-)
+from sqlalchemy import BigInteger, Column, Float, Integer, Sequence, String
 from sqlalchemy.dialects.postgresql import JSONB
 from sqlalchemy.ext.declarative import declarative_base
 from sqlalchemy.orm import sessionmaker
@@ -25,6 +17,10 @@
     select_no_chp_combustion_mastr,
 )
 from egon.data.datasets.power_plants.pv_rooftop import pv_rooftop_per_mv_grid
+from egon.data.datasets.power_plants.pv_rooftop_buildings import (
+    geocode_mastr_data,
+    pv_rooftop_to_buildings,
+)
 import egon.data.config
 import egon.data.datasets.power_plants.assign_weather_data as assign_weather_data
 import egon.data.datasets.power_plants.pv_ground_mounted as pv_ground_mounted
@@ -53,7 +49,7 @@ class PowerPlants(Dataset):
     def __init__(self, dependencies):
         super().__init__(
             name="PowerPlants",
-            version="0.0.10",
+            version="0.0.14",
             dependencies=dependencies,
             tasks=(
                 create_tables,
@@ -63,7 +59,11 @@ def __init__(self, dependencies):
                 {
                     wind_onshore.insert,
                     pv_ground_mounted.insert,
-                    pv_rooftop_per_mv_grid,
+                    (
+                        pv_rooftop_per_mv_grid,
+                        geocode_mastr_data,
+                        pv_rooftop_to_buildings,
+                    ),
                 },
                 wind_offshore.insert,
                 assign_weather_data.weatherId_and_busId,
@@ -799,7 +799,8 @@ def allocate_other_power_plants():
         """EinheitBetriebsstatus=='InBetrieb'and Energietraeger=='Klaerschlamm'"""
     )
     mastr_geothermal = pd.read_csv(cfg["sources"]["mastr_gsgk"]).query(
-        """EinheitBetriebsstatus=='InBetrieb' and Energietraeger=='Geothermie' and Technologie == 'ORCOrganicRankineCycleAnlage'"""
+        "EinheitBetriebsstatus=='InBetrieb' and Energietraeger=='Geothermie' "
+        "and Technologie == 'ORCOrganicRankineCycleAnlage'"
     )
 
     mastr_sg = mastr_sludge.append(mastr_geothermal)

src/egon/data/datasets/power_plants/pv_rooftop.py

@@ -1,11 +1,23 @@
 """The module containing all code dealing with pv rooftop distribution.
 """
+from loguru import logger
+from numpy import isclose
 import geopandas as gpd
 import pandas as pd
 
 from egon.data import config, db
+from egon.data.datasets.power_plants.pv_rooftop_buildings import (
+    PV_CAP_PER_SQ_M,
+    ROOF_FACTOR,
+    load_building_data,
+)
 from egon.data.datasets.scenario_parameters import get_sector_parameters
 
+# assumption on theoretical maximum occupancy of rooftops within an mv grid
+# district
+# TODO: this is a wild guess
+MAX_THEORETICAL_PV_OCCUPANCY = 0.5
+
 
 def pv_rooftop_per_mv_grid():
     """Execute pv rooftop distribution method per scenario
@@ -88,6 +100,20 @@ def pv_rooftop_per_mv_grid_and_scenario(scenario, level):
          """
     )
 
+    # make sure only grid districts with any buildings are used
+    valid_buildings_gdf = load_building_data()
+
+    valid_buildings_gdf = valid_buildings_gdf.assign(
+        bus_id=valid_buildings_gdf.bus_id.astype(int),
+        overlay_id=valid_buildings_gdf.overlay_id.astype(int),
+        max_cap=valid_buildings_gdf.building_area.multiply(
+            ROOF_FACTOR * PV_CAP_PER_SQ_M
+        ),
+    )
+
+    bus_ids = valid_buildings_gdf.bus_id.unique()
+    demand = demand.loc[demand.bus_id.isin(bus_ids)]
+
     # Distribute to mv grids per federal state or Germany
     if level == "federal_state":
         targets_per_federal_state = db.select_dataframe(
@@ -129,6 +155,18 @@ def pv_rooftop_per_mv_grid_and_scenario(scenario, level):
             """
         ).capacity[0]
 
+        dataset = config.settings()["egon-data"]["--dataset-boundary"]
+
+        if dataset == "Schleswig-Holstein":
+            # since the required data is missing for a SH run, it is implemented
+            # manually here
+            total_2035 = 84070
+            sh_2035 = 2700
+
+            share = sh_2035 / total_2035
+
+            target *= share
+
         demand["share_country"] = demand.demand / demand.demand.sum()
 
         demand.set_index("bus_id", inplace=True)
@@ -149,6 +187,48 @@ def pv_rooftop_per_mv_grid_and_scenario(scenario, level):
         }
     )
 
+    # ensure that no more pv rooftop capacity is allocated to any mv grid district than
+    # there is rooftop potential
+    max_cap_per_bus_df = (
+        valid_buildings_gdf[["max_cap", "bus_id"]].groupby("bus_id").sum()
+        / 1000
+        * MAX_THEORETICAL_PV_OCCUPANCY
+    )
+
+    pv_rooftop = pv_rooftop.merge(
+        max_cap_per_bus_df, how="left", left_on="bus", right_index=True
+    )
+
+    assert ~pv_rooftop.max_cap.isna().any(), (
+        "There are bus IDs within 'pv_rooftop' which are not included within "
+        " 'max_cap_per_bus_df'."
+    )
+
+    pv_rooftop = pv_rooftop.assign(delta=pv_rooftop.max_cap - pv_rooftop.p_nom)
+    loss = pv_rooftop.delta.clip(upper=0).sum()
+    total = pv_rooftop.p_nom.sum()
+
+    pv_rooftop = pv_rooftop.assign(
+        p_nom=pv_rooftop[["p_nom", "max_cap"]].min(axis=1)
+    )
+
+    pos_delta = pv_rooftop.loc[pv_rooftop.delta > 0].delta
+    rel_delta = pos_delta / pos_delta.sum()
+    add_pv_cap = rel_delta * abs(loss)
+
+    pv_rooftop.loc[add_pv_cap.index, "p_nom"] += add_pv_cap
+    pv_rooftop = pv_rooftop.drop(columns=["max_cap", "delta"])
+
+    assert isclose(
+        total, pv_rooftop.p_nom.sum()
+    ), f"{total} != {pv_rooftop.p_nom.sum()}"
+
+    if loss < 0:
+        logger.debug(
+            f"{loss:g} MW got redistributed from MV grids with too little rooftop "
+            f"potential towards other MV grids."
+        )
+
     # Select feedin timeseries
     weather_cells = db.select_geodataframe(
         f"""