fixed recycle stuff and test

pandapower/test/timeseries/test_timeseries.py

@@ -3,8 +3,6 @@
 # Copyright (c) 2016-2019 by University of Kassel and Fraunhofer Institute for Energy Economics
 # and Energy System Technology (IEE), Kassel. All rights reserved.
 
-import copy
-import os
 import tempfile
 
 import numpy as np
@@ -13,11 +11,9 @@
 
 import pandapower.control.util.diagnostic
 import pandapower as pp
-import pandapower.toolbox as tb
 import logging
-from pandapower.control import DiscreteTapControl, ContinuousTapControl, ConstControl
+from pandapower.control import ContinuousTapControl, ConstControl
 
-from pandapower import networks as nw
 from pandapower.timeseries import DFData
 from pandapower.timeseries import OutputWriter
 from pandapower.timeseries.run_time_series import run_timeseries, control_diagnostic
@@ -148,11 +144,11 @@ def test_timeseries_results():
 
     n_timesteps = 5
     profiles, ds = create_data_source(n_timesteps)
-    recycle = False
+
     # 1load
     ConstControl(net, element='load', variable='p_mw', element_index=[0, 1, 2],
                  data_source=ds, profile_name=["load1", "load2_mv_p", "load3_hv_p"],
-                 scale_factor=0.5, recycle=recycle)
+                 scale_factor=0.5)
 
     time_steps = range(0, n_timesteps)
     ow = OutputWriter(net, time_steps, output_path=tempfile.gettempdir(), output_file_type=".json")
@@ -161,18 +157,17 @@ def test_timeseries_results():
 
     ow.log_variable('res_line', 'loading_percent')
     ow.log_variable('res_line', 'i_ka')
-    run_timeseries(net, time_steps, output_writer=ow, recycle=recycle)
+    run_timeseries(net, time_steps, output_writer=ow)
     assert np.allclose(ow.output['res_load.p_mw'].sum().values * 2,
                        profiles[["load1", "load2_mv_p", "load3_hv_p"]].sum().values)
 
     # 3load - @Rieke What is this test for compared to the first one?
     # @Flo in / out of service testen ...
     ow.log_variable('res_load', 'p_mw')
     net.controller.in_service = False  # set the first controller out of service
-    ConstControl(net, 'load', 'p_mw', element_index=0, data_source=ds, profile_name='load1',
-                 recycle=recycle)
+    ConstControl(net, 'load', 'p_mw', element_index=0, data_source=ds, profile_name='load1')
 
-    run_timeseries(net, time_steps, output_writer=ow, recycle=recycle)
+    run_timeseries(net, time_steps, output_writer=ow)
     assert np.allclose(ow.output['res_load.p_mw'][0].sum(), profiles["load1"].sum())
 
 
@@ -184,19 +179,19 @@ def test_timeseries_var_func():
 
     n_timesteps = 5
     profiles, ds = create_data_source(n_timesteps)
-    recycle = False
+
     # 1load
     ConstControl(net, element='load', variable='p_mw', element_index=[0, 1, 2],
                  data_source=ds, profile_name=["load1", "load2_mv_p", "load3_hv_p"],
-                 scale_factor=0.5, recycle=recycle)
+                 scale_factor=0.5)
 
     time_steps = range(0, n_timesteps)
     ow = OutputWriter(net, time_steps, output_path=tempfile.gettempdir(), output_file_type=".json")
     ow.log_variable('res_load', 'p_mw', eval_function=np.max)
     ow.log_variable('res_bus', 'vm_pu', eval_function=np.min)
     ow.log_variable('res_bus', 'q_mvar', eval_function=np.sum)
 
-    run_timeseries(net, time_steps, output_writer=ow, recycle=recycle)
+    run_timeseries(net, time_steps, output_writer=ow)
     # asserts if last value of output_writers output is the minimum value
     assert net["res_load"]["p_mw"].max() == ow.output["res_load.p_mw"].iloc[-1].values
     assert net["res_bus"]["vm_pu"].min() == ow.output["res_bus.vm_pu"].iloc[-1].values
@@ -211,72 +206,13 @@ def test_timeseries_var_func():
                     eval_name="hv_bus_min")
     ow.log_variable('res_bus', 'vm_pu', index=mv_busses_index, eval_function=np.min,
                     eval_name="mv_bus_min")
-    run_timeseries(net, time_steps, output_writer=ow, recycle=recycle)
+    run_timeseries(net, time_steps, output_writer=ow)
     assert net["res_bus"].loc[hv_busses_index, "vm_pu"].min() == ow.output["res_bus.vm_pu"].loc[
         time_steps[-1], "hv_bus_min"]
     assert net["res_bus"].loc[mv_busses_index, "vm_pu"].min() == ow.output["res_bus.vm_pu"].loc[
         time_steps[-1], "mv_bus_min"]
 
 
-@pytest.mark.xfail(
-    reason="@author: please rework this test in cooporation with steffen, after publication of simbench within pandapower")
-def test_timeseries_recycle_simbench():
-    # Todo: Get simbench net and test trafo controllers with recycle functions
-    # test net
-    #    net = sb.get_simbench_net("1-HV-mixed--0-sw")
-    net = nw.create_cigre_network_hv()  # fixing the import error bug
-    # random datasource --> will be memmaps
-    tb.drop_inactive_elements(net)
-    n_timesteps = 3
-    profiles = create_rand_data_source(net, n_timesteps)
-
-    # 1load
-    ds = DFData(profiles["load"])
-    ConstControl(net, element='load', variable='p_mw', element_index=net.load.index,
-                 data_source=ds, profile_name=profiles["load"].keys(),
-                 scale_factor=0.5, recycle=True)
-    # sgen
-    ds = DFData(profiles["sgen"])
-    ConstControl(net, element='sgen', variable='p_mw', element_index=net.sgen.index,
-                 data_source=ds, profile_name=profiles["sgen"].keys(),
-                 scale_factor=0.5, recycle=True)
-
-    # tc = ContinuousTapControl(net, tid=6, u_set=1.04, side='lv', tol=1e-6,  trafotype="3W", recycle=True)
-    # tc = ContinuousTapControl(net, tid=5, u_set=.98, side='lv', tol=1e-6, trafotype="3W", recycle=True)
-    tc = DiscreteTapControl(net, tid=0, u_lower=1.04, u_upper=1.044, side='lv', recycle=True)
-    tc = DiscreteTapControl(net, tid=1, u_lower=.98, u_upper=.984, side='lv', recycle=True)
-    # ds = DFData(profiles["trafo3w"])
-    # ConstControl(net, element='trafo3w', variable='tap_pos', element_index=net.trafo3w.index,
-    #              data_source=ds, profile_name=profiles["trafo3w"].keys(), scale_factor=1,  recycle=True)
-    net.trafo.loc[0, "tap_pos"] = -10
-    net.trafo.loc[1, "tap_pos"] = 2
-    time_steps = range(0, n_timesteps)
-    ow = OutputWriter(net, time_steps, output_path=tempfile.gettempdir())
-    ow.log_variable('trafo', 'tap_pos')
-    ow.log_variable('res_load', 'p_mw')
-    ow.log_variable('res_bus', 'vm_pu')
-    ow.log_variable('res_line', 'loading_percent')
-    ow.log_variable('res_trafo', 'loading_percent')
-    ow.log_variable('res_line', 'i_ka')
-
-    run_timeseries(net, time_steps, output_writer=ow, recycle=False)
-    res_normal = copy.deepcopy(ow.output)
-
-    net.trafo.loc[0, "tap_pos"] = -10
-    net.trafo.loc[1, "tap_pos"] = 2
-    run_timeseries(net, time_steps, output_writer=ow, recycle=True)
-    res_recycle = copy.deepcopy(ow.output)
-
-    for key, val in res_normal.items():
-        val_recycle = res_recycle[key]
-        if key == "Parameters":
-            # Parameters is complex datatype compared to others
-            np.testing.assert_array_equal(val_recycle.values, val.values)
-        else:
-            mask = ~(np.isnan(val)).values
-            assert np.allclose(val_recycle.values[mask], val.values[mask])
-
-
 def test_time_steps():
     net = simple_test_net()
     n_timesteps = 11
@@ -300,11 +236,10 @@ def test_output_dump_after_time():
 
     n_timesteps = 100
     profiles, ds = create_data_source(n_timesteps)
-    recycle = False
+
     # 1load
     ConstControl(net, element='load', variable='p_mw', element_index=[0, 1, 2],
-                 data_source=ds, profile_name=["load1", "load2_mv_p", "load3_hv_p"],
-                 recycle=recycle)
+                 data_source=ds, profile_name=["load1", "load2_mv_p", "load3_hv_p"])
 
     time_steps = range(0, n_timesteps)
     # write output after 0.1 minutes to disk
@@ -315,7 +250,7 @@ def test_output_dump_after_time():
 
     ow.log_variable('res_line', 'loading_percent')
     ow.log_variable('res_line', 'i_ka')
-    run_timeseries(net, time_steps, output_writer=ow, recycle=recycle)
+    run_timeseries(net, time_steps, output_writer=ow)
     # ToDo: read partially dumped results and compare with all stored results
 
 

pandapower/timeseries/run_time_series.py

@@ -140,22 +140,34 @@ def run_time_step(net, time_step, ts_variables, **kwargs):
 
 
 def all_controllers_recycleable(net):
+    # checks if controller are recycleable
     recycleable = np.alltrue(net["controller"]["recycle"].values)
     if not recycleable:
         logger.warning("recycle feature not supported by some controllers in net. I have to deactive recycle")
     return recycleable
 
 
-def get_run_function(net, output_writer, recycle, **kwargs):
+def get_run_function(net, output_writer, **kwargs):
     """
-    checks if faster runpp function for time series is possible and activated or if another function is specified
-    in **kwargs
-    @param kwargs:
-    @return:
+    checks if "run" is specified in kwargs and calls this function in time series loop.
+    if "recycle" is in kwargs we use the TimeSeriesRunpp class
+
+    INPUT:
+        **net** - The pandapower format network
+        **output_writer** - The outputwriter class which stores results
+
+    RETURN:
+        **run** - the run function to be called (default is pp.runpp())
+        **recycle_class** - class to recycle implementation
     """
 
-    recycle = True if recycle and all_controllers_recycleable(net) else False
+    recycle = False
     recycle_class = None
+
+    if "recycle" in kwargs:
+        recycle = kwargs.pop("recycle")
+        recycle = True if recycle and all_controllers_recycleable(net) else False
+
     if recycle:
         from timeseries.ts_runpp import TimeSeriesRunpp
         recycle_class = TimeSeriesRunpp(net, output_writer)
@@ -168,6 +180,7 @@ def get_run_function(net, output_writer, recycle, **kwargs):
 
 
 def init_time_steps(net, time_steps, **kwargs):
+    # initializes time steps if as a range
     if not (isinstance(time_steps, list) or isinstance(time_steps, range)):
         if time_steps is None and ("start_step" in kwargs and "stop_step" in kwargs):
             logger.warning("start_step and stop_step are depricated. "
@@ -183,7 +196,7 @@ def init_time_steps(net, time_steps, **kwargs):
     return time_steps
 
 
-def init_time_series(net, time_steps, output_writer=None, recycle=False, continue_on_divergence=False, verbose=True,
+def init_time_series(net, time_steps, output_writer=None, continue_on_divergence=False, verbose=True,
                      **kwargs):
     """
     inits the time series calculation
@@ -196,7 +209,6 @@ def init_time_series(net, time_steps, output_writer=None, recycle=False, continu
     OPTIONAL:
         **output_writer** - A predefined output writer. If None the a default one is created with
                             get_default_output_writer()
-        **recycle** (bool, False) - If True faster runpp is used (rather experimental, use with caution)
         **continue_on_divergence** (bool, False) - If True time series calculation continues in case of errors.
         **verbose** (bool, True) - prints progess bar or logger debug messages
     """
@@ -208,7 +220,7 @@ def init_time_series(net, time_steps, output_writer=None, recycle=False, continu
     output_writer = init_outputwriter(net, time_steps, output_writer)
     level, order = get_controller_order(net)
     # use faster runpp if timeseries possible
-    run, recycle_class = get_run_function(net, output_writer, recycle, **kwargs)
+    run, recycle_class = get_run_function(net, output_writer, **kwargs)
 
     # True at default. Initial power flow is calculated before each control step (some controllers need inits)
     ts_variables["initial_powerflow"] = check_for_initial_powerflow(order)
@@ -246,7 +258,7 @@ def print_progress(i, time_step, time_steps, verbose, **kwargs):
     if logger.level == pplog.DEBUG and verbose:
         logger.debug("run time step %i" % time_step)
 
-    # print luigi progress
+    # print luigi pipeline progress
     if "luigi_progress" in kwargs:
         if i % 365 == 0:
             # print only every 365 time steps
@@ -258,12 +270,11 @@ def print_progress(i, time_step, time_steps, verbose, **kwargs):
             progress(progress_percentage)
 
 
-def run_timeseries(net, time_steps=None, output_writer=None, recycle=False,
-                   continue_on_divergence=False, verbose=True, **kwargs):
+def run_timeseries(net, time_steps=None, output_writer=None, continue_on_divergence=False, verbose=True, **kwargs):
     """
     Time Series main function
     Runs multiple PANDAPOWER AC power flows based on time series in controllers
-    Optionally other functions than the pp power flow can be called by setting the run function
+    Optionally other functions than the pp power flow can be called by setting the run function in kwargs
 
     INPUT:
         **net** - The pandapower format network
@@ -273,16 +284,14 @@ def run_timeseries(net, time_steps=None, output_writer=None, recycle=False,
                                                 if None, all time steps from provided data source are simulated
         **output_writer** - A predefined output writer. If None the a default one is created with
                             get_default_output_writer()
-        **recycle** (bool, False) - If True faster runpp is used (rather experimental, use with caution)
         **continue_on_divergence** (bool, False) - If True time series calculation continues in case of errors.
         **verbose** (bool, True) - prints progress bar or if logger.level == Debug it prints debug messages
         **kwargs** - Keyword arguments for run_control and runpp
     """
 
-    ts_variables = init_time_series(net, time_steps, output_writer, recycle, continue_on_divergence, verbose,
+    ts_variables = init_time_series(net, time_steps, output_writer, continue_on_divergence, verbose,
                                     **kwargs)
     control_diagnostic(net)
-
     for i, time_step in enumerate(ts_variables["time_steps"]):
         print_progress(i, time_step, ts_variables["time_steps"], verbose, **kwargs)
         run_time_step(net, time_step, ts_variables, **kwargs)