Merge pull request #1102 from PyPSA/fix-grouping-logic-again-master

Fix grouping logic again

config/config.default.yaml

@@ -390,8 +390,8 @@ solar_thermal:
 
 # docs in https://pypsa-eur.readthedocs.io/en/latest/configuration.html#existing-capacities
 existing_capacities:
-  grouping_years_power: [1895, 1920, 1950, 1955, 1960, 1965, 1970, 1975, 1980, 1985, 1990, 1995, 2000, 2005, 2010, 2015, 2020, 2025, 2030]
-  grouping_years_heat: [1980, 1985, 1990, 1995, 2000, 2005, 2010, 2015, 2020] # heat grouping years >= baseyear will be ignored
+  grouping_years_power: [1920, 1950, 1955, 1960, 1965, 1970, 1975, 1980, 1985, 1990, 1995, 2000, 2005, 2010, 2015, 2020, 2025]
+  grouping_years_heat: [1980, 1985, 1990, 1995, 2000, 2005, 2010, 2015, 2019] # heat grouping years >= baseyear will be ignored
   threshold_capacity: 10
   default_heating_lifetime: 20
   conventional_carriers:

doc/release_notes.rst

@@ -19,6 +19,8 @@ Upcoming Release
   <https://www.sciencedirect.com/science/article/pii/S0306261920312551>`__.
   See configuration ``sector: enhanced_geothermal`` for details; by default switched off.
 
+* Partially revert https://github.com/PyPSA/pypsa-eur/pull/967 to return to old grouping year logic (which was mostly correct)
+
 
 PyPSA-Eur 0.11.0 (25th May 2024)
 =====================================

scripts/add_existing_baseyear.py

@@ -201,19 +201,19 @@ def add_power_capacities_installed_before_baseyear(n, grouping_years, costs, bas
     phased_out = df_agg[df_agg["DateOut"] < baseyear].index
     df_agg.drop(phased_out, inplace=True)
 
-    older_assets = (df_agg.DateIn < min(grouping_years)).sum()
-    if older_assets:
+    newer_assets = (df_agg.DateIn > max(grouping_years)).sum()
+    if newer_assets:
         logger.warning(
-            f"There are {older_assets} assets with build year "
-            f"before first power grouping year {min(grouping_years)}. "
+            f"There are {newer_assets} assets with build year "
+            f"after last power grouping year {max(grouping_years)}. "
             "These assets are dropped and not considered."
             "Consider to redefine the grouping years to keep them."
         )
-        to_drop = df_agg[df_agg.DateIn < min(grouping_years)].index
+        to_drop = df_agg[df_agg.DateIn > max(grouping_years)].index
         df_agg.drop(to_drop, inplace=True)
 
     df_agg["grouping_year"] = np.take(
-        grouping_years[::-1], np.digitize(df_agg.DateIn, grouping_years[::-1])
+        grouping_years, np.digitize(df_agg.DateIn, grouping_years, right=True)
     )
 
     # calculate (adjusted) remaining lifetime before phase-out (+1 because assuming
@@ -464,6 +464,11 @@ def add_heating_capacities_installed_before_baseyear(
         else:
             efficiency = costs.at[costs_name, "efficiency"]
 
+        too_large_grouping_years = [gy for gy in grouping_years if gy >= int(baseyear)]
+        if too_large_grouping_years:
+            logger.warning(
+                f"Grouping years >= baseyear are ignored. Dropping {too_large_grouping_years}."
+            )
         valid_grouping_years = pd.Series(
             [
                 int(grouping_year)
@@ -473,12 +478,12 @@ def add_heating_capacities_installed_before_baseyear(
             ]
         )
 
+        assert valid_grouping_years.is_monotonic_increasing
+
         # get number of years of each interval
-        _years = (
-            valid_grouping_years.diff()
-            .shift(-1)
-            .fillna(baseyear - valid_grouping_years.iloc[-1])
-        )
+        _years = valid_grouping_years.diff()
+        # Fill NA from .diff() with value for the first interval
+        _years[0] = valid_grouping_years[0] - baseyear + default_lifetime
         # Installation is assumed to be linear for the past
         ratios = _years / _years.sum()