/
test_multi_steps_env.py
274 lines (227 loc) · 12 KB
/
test_multi_steps_env.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
# Copyright (c) 2019-2023, RTE (https://www.rte-france.com)
# See AUTHORS.txt
# This Source Code Form is subject to the terms of the Mozilla Public License, version 2.0.
# If a copy of the Mozilla Public License, version 2.0 was not distributed with this file,
# you can obtain one at http://mozilla.org/MPL/2.0/.
# SPDX-License-Identifier: MPL-2.0
# This file is part of Grid2Op, Grid2Op a testbed platform to model sequential decision making in power systems.
import grid2op
import unittest
import warnings
import numpy as np
import copy
import pdb
import os
from grid2op.tests.helper_path_test import *
import grid2op
import numpy as np
class ForecastEnvTester(unittest.TestCase):
def setUp(self) -> None:
with warnings.catch_warnings():
warnings.filterwarnings("ignore")
# this needs to be tested with pandapower backend
self.env = grid2op.make(os.path.join(PATH_DATA_TEST, "5bus_example_forecasts"), test=True)
self.env.seed(0)
self.env.set_id(0)
self.dn = self.env.action_space()
def _check_ok(self, f_obs, obs, h):
sim_obs, *_ = obs.simulate(self.dn, h)
self._check_obs(sim_obs, f_obs, h)
def _check_obs(self, sim_obs, f_obs, h):
assert np.all(sim_obs.load_p == f_obs.load_p), f"error for h={h}"
assert np.all(sim_obs.load_q == f_obs.load_q), f"error for h={h}"
assert np.all(sim_obs.gen_p == f_obs.gen_p), f"error for h={h}"
assert np.all(sim_obs.gen_q == f_obs.gen_q), f"error for h={h}"
assert np.all(sim_obs.rho == f_obs.rho), f"error for h={h}"
def test_when_do_nothing(self):
obs = self.env.reset()
forecast_env = obs.get_forecast_env()
f_obs = forecast_env.reset()
self._check_ok(f_obs, obs, 0)
for h in range(12):
f_obs, *_ = forecast_env.step(self.dn)
self._check_ok(f_obs, obs, h + 1)
def test_soft_overflow(self):
# the forecasted env should start with the same "values" for cooldowns, soft overflows etc.
# get ready for the soft overflow
a_or_first = np.array([442.308, 198.55365, 116.50534, 93.63006,
442.2703 , 110.96754, 110.96754, 92.05039])
th_lim = a_or_first * 2.
th_lim[5] /= 2.5
self.env.set_thermal_limit(th_lim)
param = self.env.parameters
param.NO_OVERFLOW_DISCONNECTION = False
param.NB_TIMESTEP_RECONNECTION = 4
self.env.change_parameters(param)
self.env.change_forecast_parameters(param)
obs = self.env.reset()
obs, *_ = self.env.step(self.dn)
assert obs.timestep_overflow[5] == 1
forecast_env = obs.get_forecast_env()
f_obs = forecast_env.reset()
assert f_obs.timestep_overflow[5] == 1
f_obs2, *_ = forecast_env.step(self.dn)
assert f_obs2.timestep_overflow[5] == 2
f_obs3, *_ = forecast_env.step(self.dn)
assert f_obs3.timestep_overflow[5] == 0
assert not f_obs3.line_status[5]
assert f_obs3.time_before_cooldown_line[5] == 4
f_obs4, *_ = forecast_env.step(self.dn)
assert not f_obs4.line_status[5]
assert f_obs4.time_before_cooldown_line[5] == 3
def test_cooldown(self):
act = self.env.action_space({"set_line_status": [(5, -1)]})
un_act = self.env.action_space({"set_line_status": [(5, +1)]})
obs = self.env.reset()
obs, *_ = self.env.step(act)
assert obs.time_before_cooldown_line[5] == 3
forecast_env = obs.get_forecast_env()
f_obs = forecast_env.reset()
assert f_obs.time_before_cooldown_line[5] == 3
f_obs2, *_ = forecast_env.step(self.dn)
assert f_obs2.time_before_cooldown_line[5] == 2
f_obs3, *_ = forecast_env.step(self.dn)
assert f_obs3.time_before_cooldown_line[5] == 1
f_obs4, r, done, info = forecast_env.step(un_act)
assert f_obs4.time_before_cooldown_line[5] == 0
assert info["is_illegal"] # because cooldown is 1 when i took the action
# now I can reco
f_obs5, r, done, info = forecast_env.step(un_act)
assert f_obs5.time_before_cooldown_line[5] == 3
assert not info["is_illegal"] # because cooldown is 1 when i took the action
assert f_obs5.line_status[5]
def test_maintenance(self):
reco = self.env.action_space({"set_line_status": [(5, +1)]})
obs = self.env.reset() # no maintenance
obs = self.env.reset() # maintenance
assert obs.time_next_maintenance[5] == 6
assert obs.duration_next_maintenance[5] == 4
forecast_env = obs.get_forecast_env()
f_obs = forecast_env.reset()
assert f_obs.time_next_maintenance[5] == 6
assert f_obs.duration_next_maintenance[5] == 4
f_obs1, *_ = forecast_env.step(self.dn)
f_obs2, *_ = forecast_env.step(self.dn)
f_obs3, *_ = forecast_env.step(self.dn)
f_obs4, *_ = forecast_env.step(self.dn)
f_obs5, *_ = forecast_env.step(self.dn)
f_obs6, *_ = forecast_env.step(self.dn)
assert f_obs6.time_next_maintenance[5] == 0
assert f_obs6.duration_next_maintenance[5] == 4
assert f_obs6.time_before_cooldown_line[5] == 4
assert not f_obs6.line_status[5]
f_obs7, *_ = forecast_env.step(self.dn)
assert f_obs7.time_next_maintenance[5] == 0
assert f_obs7.duration_next_maintenance[5] == 3
assert f_obs7.time_before_cooldown_line[5] == 3
assert not f_obs7.line_status[5]
f_obs8, *_ = forecast_env.step(self.dn)
f_obs9, *_ = forecast_env.step(self.dn)
# I cannot reco yet
f_obs10, r, d, info = forecast_env.step(reco)
assert f_obs10.time_next_maintenance[5] == -1
assert f_obs10.duration_next_maintenance[5] == 0
assert f_obs10.time_before_cooldown_line[5] == 0
assert not f_obs10.line_status[5]
assert info["is_illegal"]
f_obs11, r, d, info = forecast_env.step(reco)
assert f_obs11.time_next_maintenance[5] == -1
assert f_obs11.duration_next_maintenance[5] == 0
assert f_obs11.time_before_cooldown_line[5] == 3 # because i could act
assert f_obs11.line_status[5]
assert not info["is_illegal"]
def test_with_actions(self):
disco = self.env.action_space({"set_line_status": [(5, -1)]})
reco = self.env.action_space({"set_line_status": [(5, 1)]})
change_bus = self.env.action_space({"set_bus": {"substations_id": [(0, [1, 2, 1, 2, 1, 2])]}})
obs = self.env.reset() # no maintenance
forecast_env = obs.get_forecast_env()
f_obs1, *_ = forecast_env.step(disco)
sim_obs1, *_ = obs.simulate(disco)
self._check_obs(f_obs1, sim_obs1, 1)
assert np.all(f_obs1.time_before_cooldown_line == sim_obs1.time_before_cooldown_line)
f_obs2, *_ = forecast_env.step(self.dn)
sim_obs2, *_ = sim_obs1.simulate(self.dn)
self._check_obs(f_obs2, sim_obs2, 2)
f_obs3, f_r, f_d, f_info = forecast_env.step(change_bus)
sim_obs3, s_r, s_d, s_info = sim_obs2.simulate(change_bus)
assert not f_d
assert not s_d
self._check_obs(f_obs3, sim_obs3, 3)
assert np.all(f_obs3.time_before_cooldown_line == sim_obs3.time_before_cooldown_line)
assert np.all(f_obs3.time_before_cooldown_sub == sim_obs3.time_before_cooldown_sub)
f_obs4, *_ = forecast_env.step(self.dn)
sim_obs4, *_ = sim_obs3.simulate(self.dn)
self._check_obs(f_obs4, sim_obs4, 4)
assert np.all(f_obs4.time_before_cooldown_line == sim_obs4.time_before_cooldown_line)
assert np.all(f_obs4.time_before_cooldown_sub == sim_obs4.time_before_cooldown_sub)
f_obs5, *_ = forecast_env.step(reco)
sim_obs5, *_ = sim_obs4.simulate(reco)
self._check_obs(f_obs5, sim_obs5, 1)
assert np.all(f_obs5.time_before_cooldown_line == sim_obs5.time_before_cooldown_line)
assert np.all(f_obs5.time_before_cooldown_sub == sim_obs5.time_before_cooldown_sub)
def _aux_equal_tuple(self, ref, other):
assert len(ref) == len(other)
for i, (el_ref, el_other) in enumerate(zip(ref, other)):
assert np.all(el_ref == el_other), f"error for arrays {i}"
def test_simulate_does_not_impact_reality(self):
disco = self.env.action_space({"set_line_status": [(5, -1)]})
obs = self.env.reset() # no maintenance
forecast_env = obs.get_forecast_env()
forecast_env_cpy = forecast_env.copy()
forecast_env2 = obs.get_forecast_env()
loads_init = copy.deepcopy(self.env.backend.loads_info())
gens_init = copy.deepcopy(self.env.backend.generators_info())
lines_or_init = copy.deepcopy(self.env.backend.lines_or_info())
lines_ex_init = copy.deepcopy(self.env.backend.lines_ex_info())
# backend is properly copied
assert self.env.backend is not forecast_env.backend
assert self.env.backend is not forecast_env2.backend
assert forecast_env.backend is not forecast_env2.backend
assert forecast_env.backend is not forecast_env_cpy.backend
# I do an action in one of the simulate, it has only an impact there
f_obs, *_ = forecast_env.step(self.dn)
self._aux_equal_tuple(loads_init, self.env.backend.loads_info())
self._aux_equal_tuple(gens_init, self.env.backend.generators_info())
self._aux_equal_tuple(lines_or_init, self.env.backend.lines_or_info())
self._aux_equal_tuple(lines_ex_init, self.env.backend.lines_ex_info())
assert np.all(f_obs.load_p != forecast_env_cpy.get_obs().load_p)
assert np.all(f_obs.load_p != forecast_env2.get_obs().load_p)
# now try to do a real action
f_obs2, *_ = forecast_env.step(disco)
self._aux_equal_tuple(loads_init, self.env.backend.loads_info())
self._aux_equal_tuple(gens_init, self.env.backend.generators_info())
self._aux_equal_tuple(lines_or_init, self.env.backend.lines_or_info())
self._aux_equal_tuple(lines_ex_init, self.env.backend.lines_ex_info())
assert np.all(f_obs2.load_p != forecast_env_cpy.get_obs().load_p)
assert np.all(f_obs2.load_p != forecast_env2.get_obs().load_p)
# now try to do a real action
f2_obs1, *_ = forecast_env2.step(disco)
self._aux_equal_tuple(loads_init, self.env.backend.loads_info())
self._aux_equal_tuple(gens_init, self.env.backend.generators_info())
self._aux_equal_tuple(lines_or_init, self.env.backend.lines_or_info())
self._aux_equal_tuple(lines_ex_init, self.env.backend.lines_ex_info())
assert np.all(f2_obs1.load_p != forecast_env_cpy.get_obs().load_p)
assert np.all(f2_obs1.load_p != forecast_env.get_obs().load_p)
def test_right_parameters(self):
# the forecasted env should start with the same "values" for cooldowns, soft overflows etc.
assert self.env.parameters.NB_TIMESTEP_RECONNECTION == 10
param = self.env.parameters
param.NO_OVERFLOW_DISCONNECTION = False
param.NB_TIMESTEP_RECONNECTION = 4
self.env.change_parameters(param)
obs = self.env.reset()
assert self.env.parameters.NB_TIMESTEP_RECONNECTION == 4
forecast_env = obs.get_forecast_env()
f_obs = forecast_env.reset()
assert forecast_env.parameters.NB_TIMESTEP_RECONNECTION == 10
self.env.change_forecast_parameters(param)
obs2 = self.env.reset()
forecast_env2 = obs2.get_forecast_env()
assert forecast_env2.parameters.NB_TIMESTEP_RECONNECTION == 4
param.NB_TIMESTEP_RECONNECTION = 17
forecast_env2.change_parameters(param)
obs3 = forecast_env2.reset()
assert forecast_env2.parameters.NB_TIMESTEP_RECONNECTION == 17
if __name__ == "__main__":
unittest.main()