Add tests for yaw opt with turbines disabled, as well as basic serial…

… refine operation.

tests/floris_model_integration_test.py

@@ -10,7 +10,7 @@
     TimeSeries,
     WindRose,
 )
-from floris.core.turbine.operation_models import POWER_SETPOINT_DEFAULT
+from floris.core.turbine.operation_models import POWER_SETPOINT_DEFAULT, POWER_SETPOINT_DISABLED
 
 
 TEST_DATA = Path(__file__).resolve().parent / "data"
@@ -44,7 +44,24 @@ def test_assign_setpoints():
 
     # power_setpoints and disable_turbines (disable_turbines overrides power_setpoints)
     fmodel.set(power_setpoints=[[1e6, 2e6]], disable_turbines=[[True, False]])
-    assert np.allclose(fmodel.core.farm.power_setpoints, np.array([[0.001, 2e6]]))
+    assert np.allclose(fmodel.core.farm.power_setpoints, np.array([[POWER_SETPOINT_DISABLED, 2e6]]))
+
+    # Setting sequentially is equivalent to setting together
+    fmodel.reset_operation()
+    fmodel.set(disable_turbines=[[True, False]])
+    fmodel.set(yaw_angles=[[0, 30]])
+    assert np.allclose(
+        fmodel.core.farm.power_setpoints,
+        np.array([[POWER_SETPOINT_DISABLED, POWER_SETPOINT_DEFAULT]])
+    )
+    assert np.allclose(fmodel.core.farm.yaw_angles, np.array([[0, 30]]))
+
+    fmodel.set(disable_turbines=[[True, False]], yaw_angles=[[0, 30]])
+    assert np.allclose(
+        fmodel.core.farm.power_setpoints,
+        np.array([[POWER_SETPOINT_DISABLED, POWER_SETPOINT_DEFAULT]])
+    )
+    assert np.allclose(fmodel.core.farm.yaw_angles, np.array([[0, 30]]))
 
 def test_set_run():
     """

tests/serial_refine_unit_test.py

@@ -0,0 +1,107 @@
+
+import numpy as np
+import pandas as pd
+
+from floris import FlorisModel
+from floris.optimization.yaw_optimization.yaw_optimizer_sr import YawOptimizationSR
+
+
+DEBUG = False
+VELOCITY_MODEL = "gauss"
+DEFLECTION_MODEL = "gauss"
+
+# Inputs for basic yaw optimizations
+WIND_DIRECTIONS = [0.0, 90.0, 180.0, 270.0]
+WIND_SPEEDS = [8.0] * 4
+TURBULENCE_INTENSITIES = [0.06] * 4
+LAYOUT_X = [0.0, 600.0, 1200.0]
+LAYOUT_Y = [0.0, 0.0, 0.0]
+MAXIMUM_YAW_ANGLE = 25.0
+
+def test_basic_optimization(sample_inputs_fixture):
+    """
+    The Serial Refine (SR) method optimizes yaw angles based on a sequential, iterative yaw
+    optimization scheme. This test checks basic properties of the optimization result.
+    """
+    sample_inputs_fixture.core["wake"]["model_strings"]["velocity_model"] = VELOCITY_MODEL
+    sample_inputs_fixture.core["wake"]["model_strings"]["deflection_model"] = DEFLECTION_MODEL
+
+    fmodel = FlorisModel(sample_inputs_fixture.core)
+
+    fmodel.set(
+        layout_x=LAYOUT_X,
+        layout_y=LAYOUT_Y,
+        wind_directions=WIND_DIRECTIONS,
+        wind_speeds=WIND_SPEEDS,
+        turbulence_intensities=TURBULENCE_INTENSITIES
+    )
+    fmodel.set_operation_model("cosine-loss")
+
+    yaw_opt = YawOptimizationSR(fmodel, minimum_yaw_angle=0.0, maximum_yaw_angle=MAXIMUM_YAW_ANGLE)
+    df_opt = yaw_opt.optimize()
+
+    # Unaligned conditions
+    assert np.allclose(df_opt.loc[0, "yaw_angles_opt"], 0.0)
+    assert np.allclose(df_opt.loc[2, "yaw_angles_opt"], 0.0)
+
+    # Check aligned conditions
+    # Check maximum and minimum are respected
+    assert (df_opt.loc[1, "yaw_angles_opt"] <= MAXIMUM_YAW_ANGLE).all()
+    assert (df_opt.loc[3, "yaw_angles_opt"] <= MAXIMUM_YAW_ANGLE).all()
+    assert (df_opt.loc[1, "yaw_angles_opt"] >= 0.0).all()
+    assert (df_opt.loc[3, "yaw_angles_opt"] >= 0.0).all()
+
+    # Check 90.0 and 270.0 are symmetric
+    assert np.allclose(df_opt.loc[1, "yaw_angles_opt"], np.flip(df_opt.loc[3, "yaw_angles_opt"]))
+
+    # Check last turbine's angles are zero at 270.0
+    assert np.allclose(df_opt.loc[3, "yaw_angles_opt"][-1], 0.0)
+
+    # Check that optimizer reports a power improvement
+    assert (df_opt["farm_power_opt"] >= df_opt["farm_power_baseline"]).all()
+
+def test_disabled_turbines(sample_inputs_fixture):
+    """
+    Tests SR when some turbines are disabled and checks that the results are equivalent to removing
+    those turbines from the wind farm. Need a tight layout to ensure that the front-to-back distance
+    is not too large.
+    """
+
+    sample_inputs_fixture.core["wake"]["model_strings"]["velocity_model"] = VELOCITY_MODEL
+    sample_inputs_fixture.core["wake"]["model_strings"]["deflection_model"] = DEFLECTION_MODEL
+
+    fmodel = FlorisModel(sample_inputs_fixture.core)
+
+    fmodel.set(
+        layout_x=LAYOUT_X,
+        layout_y=LAYOUT_Y,
+        wind_directions=WIND_DIRECTIONS,
+        wind_speeds=WIND_SPEEDS,
+        turbulence_intensities=TURBULENCE_INTENSITIES
+    )
+    fmodel.set_operation_model("mixed")
+
+    # Disable the middle turbine in all wind conditions, run optimization, and extract results
+    fmodel.set(disable_turbines=[[False, True, False]]*4)
+    yaw_opt = YawOptimizationSR(fmodel, minimum_yaw_angle=0.0, maximum_yaw_angle=MAXIMUM_YAW_ANGLE)
+    df_opt = yaw_opt.optimize()
+    yaw_angles_opt_disabled = df_opt.loc[3, "yaw_angles_opt"]
+    farm_power_opt_disabled = df_opt.loc[3, "farm_power_opt"]
+
+    # Set up a new wind farm with the middle turbine removed
+    fmodel = FlorisModel(sample_inputs_fixture.core)
+    fmodel.set(
+        layout_x=np.array(LAYOUT_X)[[0, 2]],
+        layout_y=np.array(LAYOUT_Y)[[0, 2]],
+        wind_directions=WIND_DIRECTIONS,
+        wind_speeds=WIND_SPEEDS,
+        turbulence_intensities=TURBULENCE_INTENSITIES
+    )
+    fmodel.set_operation_model("cosine-loss")
+    yaw_opt = YawOptimizationSR(fmodel, minimum_yaw_angle=0.0, maximum_yaw_angle=MAXIMUM_YAW_ANGLE)
+    df_opt = yaw_opt.optimize()
+    yaw_angles_opt_removed = df_opt.loc[3, "yaw_angles_opt"]
+    farm_power_opt_removed = df_opt.loc[3, "farm_power_opt"]
+
+    assert np.allclose(yaw_angles_opt_disabled[[0, 2]], yaw_angles_opt_removed)
+    assert np.allclose(farm_power_opt_disabled, farm_power_opt_removed)