-
Notifications
You must be signed in to change notification settings - Fork 4
/
Copy pathsmarteole_example.py
353 lines (310 loc) · 15.1 KB
/
smarteole_example.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
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
from __future__ import annotations
import logging
import sys
import zipfile
from functools import partial
from pathlib import Path
from typing import IO, NamedTuple
import pandas as pd
from pandas.testing import assert_frame_equal
from scipy.stats import circmean
from wind_up.caching import with_parquet_cache
from wind_up.combine_results import calc_net_uplift
from wind_up.constants import OUTPUT_DIR, PROJECTROOT_DIR, TIMESTAMP_COL, DataColumns
from wind_up.interface import AssessmentInputs
from wind_up.main_analysis import run_wind_up_analysis
from wind_up.models import Asset, PlotConfig, Toggle, Turbine, WindUpConfig
from wind_up.reanalysis_data import ReanalysisDataset
sys.path.append(str(PROJECTROOT_DIR))
from examples.helpers import download_zenodo_data, format_and_print_results_table, setup_logger
CACHE_DIR = PROJECTROOT_DIR / "cache" / "smarteole_example_data"
ANALYSIS_OUTPUT_DIR = OUTPUT_DIR / "smarteole_example"
ANALYSIS_OUTPUT_DIR.mkdir(exist_ok=True, parents=True)
ANALYSIS_TIMEBASE_S = 600
CACHE_SUBDIR = CACHE_DIR / f"timebase_{ANALYSIS_TIMEBASE_S}"
CACHE_SUBDIR.mkdir(exist_ok=True, parents=True)
CHECK_RESULTS = True
PARENT_DIR = Path(__file__).parent
ZIP_FILENAME = "SMARTEOLE-WFC-open-dataset.zip"
MINIMUM_DATA_COUNT_COVERAGE = 0.5 # 50% of the data must be present
DEFAULT_SCADA_FILE_PATH = "SMARTEOLE-WFC-open-dataset/SMARTEOLE_WakeSteering_SCADA_1minData.csv"
DEFAULT_METADATA_FILE_PATH = "SMARTEOLE-WFC-open-dataset/SMARTEOLE_WakeSteering_Coordinates_staticData.csv"
DEFAULT_TOGGLE_FILE_PATH = "SMARTEOLE-WFC-open-dataset/SMARTEOLE_WakeSteering_ControlLog_1minData.csv"
REANALYSIS_DATA_FILE_PATH = (
PROJECTROOT_DIR / "tests/test_data/smarteole/ERA5T_50.00N_2.75E_100m_1hr_20200201_20200531.parquet"
)
@with_parquet_cache(CACHE_SUBDIR / "smarteole_scada.parquet")
def unpack_smarteole_scada(
timebase_s: int, scada_data_file: Path | str | IO[bytes] = DEFAULT_SCADA_FILE_PATH
) -> pd.DataFrame:
"""Function that translates 1-minute SCADA data to x minute data in the wind-up expected format"""
def _separate_turbine_id_from_field(x: str) -> tuple[str, str]:
parts = x.split("_")
if len(parts[-1]) == 1:
wtg_id = parts[-1]
col_name = "_".join(parts[:-1])
else:
wtg_id = parts[-2]
col_name = "_".join(parts[:-2] + [parts[-1]])
return f"SMV{wtg_id}", col_name
def _make_turbine_id_a_column(df: pd.DataFrame) -> pd.DataFrame:
df.columns = pd.MultiIndex.from_tuples(
(_separate_turbine_id_from_field(i) for i in df.columns),
names=[DataColumns.turbine_name, "field"],
)
return df.stack(level=0, future_stack=True).reset_index(DataColumns.turbine_name) # noqa: PD013
def _map_and_mask_cols(df: pd.DataFrame) -> pd.DataFrame:
x_minutes_count_lower_limit = timebase_s * MINIMUM_DATA_COUNT_COVERAGE
mask_active_power = df["active_power_count"] < x_minutes_count_lower_limit
mask_wind_speed = df["wind_speed_count"] < x_minutes_count_lower_limit
mask_pitch_angle = df["blade_1_pitch_angle_count"] < x_minutes_count_lower_limit
mask_gen_rpm = df["generator_speed_count"] < x_minutes_count_lower_limit
mask_temperature = df["temperature_count"] < x_minutes_count_lower_limit
mask_nacelle_position = df["nacelle_position_count"] < x_minutes_count_lower_limit
return df.assign(
**{
DataColumns.active_power_mean: lambda d: d["active_power_avg"].mask(mask_active_power),
DataColumns.active_power_sd: lambda d: d["active_power_std"].mask(mask_active_power),
DataColumns.wind_speed_mean: lambda d: d["wind_speed_avg"].mask(mask_wind_speed),
DataColumns.wind_speed_sd: lambda d: d["wind_speed_std"].mask(mask_wind_speed),
DataColumns.yaw_angle_mean: lambda d: d["nacelle_position_avg"].mask(mask_nacelle_position),
DataColumns.yaw_angle_min: lambda d: d["nacelle_position_min"].mask(mask_nacelle_position),
DataColumns.yaw_angle_max: lambda d: d["nacelle_position_max"].mask(mask_nacelle_position),
DataColumns.pitch_angle_mean: lambda d: d["blade_1_pitch_angle_avg"].mask(mask_pitch_angle),
DataColumns.gen_rpm_mean: lambda d: d["generator_speed_avg"].mask(mask_gen_rpm),
DataColumns.ambient_temp: lambda d: d["temperature_avg"].mask(mask_temperature),
DataColumns.shutdown_duration: 0,
}
)
# unzipping the data in memory and only reading the relevant files
circular_mean = partial(circmean, low=0, high=360)
return (
pd.read_csv(scada_data_file, parse_dates=[0], index_col=0)
.pipe(_make_turbine_id_a_column)
.groupby(DataColumns.turbine_name)
.resample(f"{timebase_s}s")
.aggregate(
{
"active_power_avg": "mean",
"active_power_std": "mean",
"active_power_count": "sum",
"wind_speed_avg": "mean",
"wind_speed_std": "mean",
"wind_speed_count": "sum",
"blade_1_pitch_angle_avg": "mean", # no need for circular_mean because no wrap
"blade_1_pitch_angle_count": "sum",
"generator_speed_avg": "mean",
"generator_speed_count": "sum",
"temperature_avg": "mean",
"temperature_count": "sum",
"nacelle_position_avg": circular_mean,
"nacelle_position_max": "max",
"nacelle_position_min": "min",
"nacelle_position_count": "sum",
}
)
.reset_index(DataColumns.turbine_name)
.pipe(_map_and_mask_cols)
.loc[:, DataColumns.all()]
.rename_axis(TIMESTAMP_COL, axis=0)
.rename_axis(None, axis=1)
)
@with_parquet_cache(CACHE_DIR / "smarteole_metadata.parquet")
def unpack_smarteole_metadata(
timebase_s: int, metadata_file: Path | str | IO[bytes] = DEFAULT_METADATA_FILE_PATH
) -> pd.DataFrame:
return (
pd.read_csv(metadata_file, index_col=0)
.reset_index()
.rename(columns={"Turbine": "Name"})
.query("Name.str.startswith('SMV')") # is a turbine
.loc[:, ["Name", "Latitude", "Longitude"]]
.assign(TimeZone="UTC", TimeSpanMinutes=timebase_s / 60, TimeFormat="Start")
)
@with_parquet_cache(CACHE_SUBDIR / "smarteole_toggle.parquet")
def unpack_smarteole_toggle_data(
timebase_s: int, toggle_file: Path | str | IO[bytes] = DEFAULT_TOGGLE_FILE_PATH
) -> pd.DataFrame:
ten_minutes_count_lower_limit = timebase_s * MINIMUM_DATA_COUNT_COVERAGE
toggle_value_threshold: float = 0.95
raw_df = pd.read_csv(toggle_file, parse_dates=[0], index_col=0)
required_in_cols = [
"control_log_offset_active_avg",
"control_log_offset_active_count",
"control_log_offset_avg",
]
toggle_df = (
raw_df[required_in_cols]
.resample(f"{timebase_s}s")
.agg(
{
"control_log_offset_active_avg": "mean",
"control_log_offset_active_count": "sum",
"control_log_offset_avg": "mean",
}
)
)
toggle_df["toggle_on"] = (toggle_df["control_log_offset_active_avg"] >= toggle_value_threshold) & (
toggle_df["control_log_offset_active_count"] >= ten_minutes_count_lower_limit
)
toggle_df["toggle_off"] = (toggle_df["control_log_offset_active_avg"] <= (1 - toggle_value_threshold)) & (
toggle_df["control_log_offset_active_count"] >= ten_minutes_count_lower_limit
)
toggle_df["yaw_offset_command"] = toggle_df["control_log_offset_avg"]
toggle_df.index = toggle_df.index.tz_localize("UTC")
toggle_df.index.name = TIMESTAMP_COL
return toggle_df[["toggle_on", "toggle_off", "yaw_offset_command"]]
def define_smarteole_example_config(
analysis_timebase_s: int,
analysis_output_dir: Path,
) -> WindUpConfig:
wtg_map = {
f"SMV{i}": Turbine.model_validate(
{
"name": f"SMV{i}",
"turbine_type": {
"turbine_type": "Senvion-MM82-2050",
"rotor_diameter_m": 82.0,
"rated_power_kw": 2050.0,
"cutout_ws_mps": 25,
"normal_operation_pitch_range": (-10.0, 35.0),
"normal_operation_genrpm_range": (250.0, 2000.0),
"rpm_v_pw_margin_factor": 0.05,
"pitch_to_stall": False,
},
}
)
for i in range(1, 7 + 1)
}
northing_corrections_utc = [
("SMV1", pd.Timestamp("2020-02-17 16:30:00+0000"), 5.750994540354649),
("SMV2", pd.Timestamp("2020-02-17 16:30:00+0000"), 5.690999999999994),
("SMV3", pd.Timestamp("2020-02-17 16:30:00+0000"), 5.558000000000042),
("SMV4", pd.Timestamp("2020-02-17 16:30:00+0000"), 5.936999999999996),
("SMV5", pd.Timestamp("2020-02-17 16:30:00+0000"), 6.797253350869262),
("SMV6", pd.Timestamp("2020-02-17 16:30:00+0000"), 5.030130916842758),
("SMV7", pd.Timestamp("2020-02-17 16:30:00+0000"), 4.605999999999972),
]
wd_filter_margin = 3 + 7 * analysis_timebase_s / 600
return WindUpConfig(
assessment_name="smarteole_example",
timebase_s=analysis_timebase_s,
require_ref_wake_free=True,
detrend_min_hours=12,
ref_wd_filter=[207 - wd_filter_margin, 236 + wd_filter_margin], # steer is from 207-236
filter_all_test_wtgs_together=True,
use_lt_distribution=False,
out_dir=analysis_output_dir,
test_wtgs=[wtg_map["SMV6"], wtg_map["SMV5"]],
ref_wtgs=[wtg_map["SMV7"]],
ref_super_wtgs=[],
non_wtg_ref_names=[],
analysis_first_dt_utc_start=pd.Timestamp("2020-02-17 16:30:00+0000"),
upgrade_first_dt_utc_start=pd.Timestamp("2020-02-17 16:30:00+0000"),
analysis_last_dt_utc_start=pd.Timestamp("2020-05-25 00:00:00+0000") - pd.Timedelta(seconds=ANALYSIS_TIMEBASE_S),
lt_first_dt_utc_start=pd.Timestamp("2020-02-17 16:30:00+0000"),
lt_last_dt_utc_start=pd.Timestamp("2020-05-25 00:00:00+0000") - pd.Timedelta(seconds=ANALYSIS_TIMEBASE_S),
detrend_first_dt_utc_start=pd.Timestamp("2020-02-17 16:30:00+0000"),
detrend_last_dt_utc_start=pd.Timestamp("2020-05-25 00:00:00+0000") - pd.Timedelta(seconds=ANALYSIS_TIMEBASE_S),
years_for_lt_distribution=0,
years_for_detrend=0,
ws_bin_width=1.0,
asset=Asset.model_validate(
{
"name": "Sole du Moulin Vieux",
"wtgs": list(wtg_map.values()),
"masts_and_lidars": [],
}
),
northing_corrections_utc=northing_corrections_utc,
toggle=Toggle.model_validate(
{
"name": "wake steering",
"toggle_file_per_turbine": False,
"toggle_filename": "SMV_offset_active_toggle_df.parquet",
"detrend_data_selection": "use_toggle_off_data",
"pairing_filter_method": "any_within_timedelta",
"pairing_filter_timedelta_seconds": 3600,
"toggle_change_settling_filter_seconds": 120,
}
),
)
def print_smarteole_results(
results_per_test_ref_df: pd.DataFrame, *, print_small_table: bool = False, check_results: bool = False
) -> None:
print_df = format_and_print_results_table(results_per_test_ref_df, print_small_table=print_small_table)
if check_results:
# raise an error if results don't match expected
expected_print_df = pd.DataFrame(
{
"turbine": ["SMV6", "SMV5"],
"reference": ["SMV7", "SMV7"],
"energy uplift": ["-1.0%", "3.1%"],
"uplift uncertainty": ["0.6%", "1.2%"],
"uplift P95": ["-2.0%", "1.2%"],
"uplift P5": ["-0.1%", "5.0%"],
"valid hours toggle off": [132 + 3 / 6, 133 + 0 / 6],
"valid hours toggle on": [136 + 0 / 6, 137 + 1 / 6],
"mean power toggle on": [1148, 994],
},
index=[0, 1],
)
assert_frame_equal(print_df, expected_print_df)
class SmarteoleData(NamedTuple):
scada_df: pd.DataFrame
metadata_df: pd.DataFrame
toggle_df: pd.DataFrame
def _download_data_from_zenodo(analysis_timebase_s: int, cache_dir: Path, zip_filename: str) -> SmarteoleData:
download_zenodo_data(record_id="7342466", output_dir=cache_dir, filenames={zip_filename})
with zipfile.ZipFile(cache_dir / zip_filename) as zf:
scada_df = unpack_smarteole_scada(analysis_timebase_s, scada_data_file=zf.open(DEFAULT_SCADA_FILE_PATH))
metadata_df = unpack_smarteole_metadata(analysis_timebase_s, metadata_file=zf.open(DEFAULT_METADATA_FILE_PATH))
toggle_df = unpack_smarteole_toggle_data(analysis_timebase_s, toggle_file=zf.open(DEFAULT_TOGGLE_FILE_PATH))
return SmarteoleData(scada_df=scada_df, metadata_df=metadata_df, toggle_df=toggle_df)
def main_smarteole_analysis(
*,
smarteole_data: SmarteoleData,
analysis_timebase_s: int = ANALYSIS_TIMEBASE_S,
check_results: bool = CHECK_RESULTS,
analysis_output_dir: Path = ANALYSIS_OUTPUT_DIR,
cache_sub_dir: Path = CACHE_SUBDIR,
reanalysis_file_path: Path | str = REANALYSIS_DATA_FILE_PATH,
) -> None:
setup_logger(ANALYSIS_OUTPUT_DIR / "analysis.log")
logger = logging.getLogger(__name__)
logger.info("Merging SMV6 yaw offset command signal into SCADA data")
toggle_df_no_tz = smarteole_data.toggle_df.copy()
toggle_df_no_tz.index = toggle_df_no_tz.index.tz_localize(None)
scada_df = smarteole_data.scada_df.merge(
toggle_df_no_tz["yaw_offset_command"], left_index=True, right_index=True, how="left"
)
scada_df["yaw_offset_command"] = scada_df["yaw_offset_command"].where(scada_df["TurbineName"] == "SMV6", 0)
del toggle_df_no_tz
logger.info("Loading reference reanalysis data")
reanalysis_dataset = ReanalysisDataset(
id="ERA5T_50.00N_2.75E_100m_1hr",
data=pd.read_parquet(reanalysis_file_path),
)
logger.info("Defining Assessment Configuration")
cfg = define_smarteole_example_config(
analysis_timebase_s=analysis_timebase_s, analysis_output_dir=analysis_output_dir
)
plot_cfg = PlotConfig(show_plots=False, save_plots=True, plots_dir=cfg.out_dir / "plots")
assessment_inputs = AssessmentInputs.from_cfg(
cfg=cfg,
plot_cfg=plot_cfg,
toggle_df=smarteole_data.toggle_df,
scada_df=scada_df,
metadata_df=smarteole_data.metadata_df,
reanalysis_datasets=[reanalysis_dataset],
cache_dir=cache_sub_dir,
)
results_per_test_ref_df = run_wind_up_analysis(assessment_inputs)
net_p50, net_p95, net_p5 = calc_net_uplift(results_per_test_ref_df, confidence=0.9)
print(f"net P50: {net_p50:.1%}, net P95: {net_p95:.1%}, net P5: {net_p5:.1%}")
print_smarteole_results(results_per_test_ref_df, check_results=check_results)
if __name__ == "__main__":
smarteole_data = _download_data_from_zenodo(
analysis_timebase_s=ANALYSIS_TIMEBASE_S, cache_dir=CACHE_DIR, zip_filename=ZIP_FILENAME
)
main_smarteole_analysis(smarteole_data=smarteole_data)