[Scheduling] convert the units of inflexible power sensors MW

documentation/changelog.rst

@@ -14,11 +14,13 @@ New features
 
 * Add command ``flexmeasures edit transfer-ownership`` to transfer the ownership of an asset and its children from one account to another[see `PR #983 <https://github.com/FlexMeasures/flexmeasures/pull/983>`_]
 * Support defining the ``site-power-capacity``, ``site-consumption-capacity`` and ``site-production-capacity`` as a sensor in the API and CLI [see `PR #985 <https://github.com/FlexMeasures/flexmeasures/pull/985>`_]
+* Support defining inflexible power sensors with arbitrary power and energy units [see `PR #1007 <https://github.com/FlexMeasures/flexmeasures/pull/1007>`_]
 * Support saving beliefs with a ``belief_horizon`` in the ``PandasReporter``[see `PR #1013 <https://github.com/FlexMeasures/flexmeasures/pull/1013>`_]
 * Skip the check of the output event resolution in any ``Reporter`` with the field ``check_output_resolution`` [see `PR #1009 <https://github.com/FlexMeasures/flexmeasures/pull/1009>`_]
 
 Bugfixes
 -----------
+
 * Use minimum event resolution of the input (instead of the output) sensors for the belief search parameters [see `PR #1010 <https://github.com/FlexMeasures/flexmeasures/pull/1010>`_]
 
 

documentation/features/scheduling.rst

@@ -45,7 +45,7 @@ With the flexibility context, we aim to describe the system in which the flexibl
      - Description 
    * - ``inflexible-device-sensors``
      - ``[3,4]``
-     - Power sensors that are relevant, but not flexible, such as a sensor recording rooftop solar power connected behind the main meter, whose production falls under the same contract as the flexible device(s) being scheduled. * Note:* For now, the inflexible-device-sensors need to have MW units. 
+     - Power sensors that are relevant, but not flexible, such as a sensor recording rooftop solar power connected behind the main meter, whose production falls under the same contract as the flexible device(s) being scheduled.
    * - ``consumption-price-sensor``
      - ``5``
      - The sensor that defines the price of consuming energy. This sensor can be recording market prices, but also CO₂ - whatever fits your optimization problem.

flexmeasures/data/models/planning/tests/conftest.py

@@ -145,7 +145,7 @@ def inflexible_devices(db, building) -> dict[str, Sensor]:
         name="PV power sensor",
         generic_asset=building,
         event_resolution=timedelta(hours=1),
-        unit="kW",
+        unit="MW",
         attributes={"capacity_in_mw": 2},
     )
     db.session.add(pv_sensor)
@@ -186,9 +186,9 @@ def add_inflexible_device_forecasts(
     ) * (len(time_slots) // (24 * 4))
     add_as_beliefs(db, pv_sensor, pv_values, time_slots, setup_sources["Seita"])
 
-    # Residual demand (1 MW continuously)
+    # Residual demand (1 MW = 1000 kW continuously)
     residual_demand_sensor = inflexible_devices["residual demand power sensor"]
-    residual_demand_values = [-1] * len(time_slots)
+    residual_demand_values = [-1000] * len(time_slots)
     add_as_beliefs(
         db,
         residual_demand_sensor,

flexmeasures/data/models/planning/tests/test_solver.py

@@ -26,7 +26,7 @@
     integrate_time_series,
 )
 from flexmeasures.tests.utils import get_test_sensor
-
+from flexmeasures.utils.unit_utils import convert_units
 
 TOLERANCE = 0.00001
 
@@ -586,9 +586,20 @@ def test_building_solver_day_2(
     with pd.option_context("display.max_rows", None, "display.max_columns", 3):
         print(soc_schedule)
 
+    unit_factors = np.expand_dims(
+        [
+            convert_units(1, s.unit, "MW")
+            for s in add_inflexible_device_forecasts.keys()
+        ],
+        axis=1,
+    )
+    inflexible_devices_power = np.array(list(add_inflexible_device_forecasts.values()))
+
     # Check if constraints were met
     capacity = pd.DataFrame(
-        data=np.sum(np.array(list(add_inflexible_device_forecasts.values())), axis=0),
+        data=inflexible_devices_power.T.dot(
+            unit_factors
+        ),  # convert to MW and sum column-wise
         columns=["inflexible"],
     ).tail(
         -4 * 24

flexmeasures/data/models/planning/utils.py

@@ -178,7 +178,7 @@ def get_power_values(
     beliefs_before: datetime | None,
     sensor: Sensor,
 ) -> np.ndarray:
-    """Get measurements or forecasts of an inflexible device represented by a power sensor.
+    """Get measurements or forecasts of an inflexible device represented by a power or energy sensor as an array of power values in MW.
 
     If the requested schedule lies in the future, the returned data will consist of (the most recent) forecasts (if any exist).
     If the requested schedule lies in the past, the returned data will consist of (the most recent) measurements (if any exist).
@@ -208,12 +208,14 @@ def get_power_values(
         )
         df = df.fillna(0)
 
+    series = convert_units(df.values, sensor.unit, "MW")
+
     if sensor.get_attribute(
         "consumption_is_positive", False
     ):  # FlexMeasures default is to store consumption as negative power values
-        return df.values
+        return series
 
-    return -df.values
+    return -series
 
 
 def fallback_charging_policy(