def static_mreg_svr(windpark):
target = windpark.get_target()
measurements = repair_nrel(target.get_measurements()[:10000])
neighbors = windpark.get_turbines()[:-1]
nseries = [repair_nrel(t.get_measurements()[:10000]) for t in neighbors]
reg = linear_model.LinearRegression()
gamma_range = [0.0001, 0.000001]
C_range = [2 ** i for i in range(1, 4, 1)]
regargs = {
"epsilon" : 0.1,
"cv_method" : KFold,
"cv_args" : {"n_folds" : 10},
"kernel" : 'rbf',
"tuned_parameters" : [{
'kernel': [kernel],
'C': C_range,
'gamma': gamma_range}]}
iargs = {'method':'mreg',
'timestep': timestep,
'neighbor_series': nseries,
'reg' : reg,
'regargs' : regargs}
return iargs
def static_forwardcopy(windpark):
target = windpark.get_target()
measurements = repair_nrel(target.get_measurements()[:10000])
neighbors = windpark.get_turbines()[:-1]
nseries = [repair_nrel(t.get_measurements()[:10000]) for t in neighbors]
return {'method': 'forwardcopy', 'timestep': timestep}, nseries
def static_linear(windpark):
target = windpark.get_target()
measurements = repair_nrel(target.get_measurements()[:10000])
neighbors = windpark.get_turbines()[:-1]
nseries = [repair_nrel(t.get_measurements()[:10000]) for t in neighbors]
return {"method": "linear", "timestep": timestep}, nseries
def static_forwardcopy(windpark):
target = windpark.get_target()
measurements = repair_nrel(target.get_measurements()[:10000])
neighbors = windpark.get_turbines()[:-1]
nseries = [repair_nrel(t.get_measurements()[:10000]) for t in neighbors]
return {'method':'forwardcopy','timestep': timestep}, nseries
def static_mreg_lin(windpark):
target = windpark.get_target()
measurements = repair_nrel(target.get_measurements()[:10000])
neighbors = windpark.get_turbines()[:-1]
nseries = [repair_nrel(t.get_measurements()[:10000]) for t in neighbors]
iargs = {"method": "mreg", "timestep": timestep, "neighbor_series": nseries, "reg": "linear_model"}
return iargs, nseries
def static_mreg_lin(windpark):
target = windpark.get_target()
measurements = repair_nrel(target.get_measurements()[:10000])
neighbors = windpark.get_turbines()[:-1]
nseries = [repair_nrel(t.get_measurements()[:10000]) for t in neighbors]
iargs = {'method':'mreg',
'timestep': timestep,
'neighbor_series': nseries,
'reg' : 'linear_model'}
return iargs
def static_mreg_knn(windpark):
target = windpark.get_target()
measurements = repair_nrel(target.get_measurements()[:10000])
neighbors = windpark.get_turbines()[:-1]
nseries = [repair_nrel(t.get_measurements()[:10000]) for t in neighbors]
iargs = {'method':'mreg',
'timestep': timestep,
'neighbor_series': nseries,
'reg' : 'knn',
'regargs' : {'kfold': 5, 'n' : [5,10,20,50,100], 'variant' : 'uniform'}}
return iargs
def experiment(method, windpark, windpark_test, damaged, rate):
args, nseries = argfuncs[method](windpark)
reconstructed = interpolate(damaged, **args)
target = windpark.get_target()
measurements = repair_nrel(target.get_measurements()[:10000])
turbines = windpark.get_turbines()
for t in range(len(turbines)):
turbines[t].add_measurements(\
repair_nrel(turbines[t].get_measurements()[:10000]))
# this is the target turbine, use the reconstructed here.
turbines[-1].add_measurements(reconstructed)
feature_window, horizon = 3, 3
mapping = PowerMapping()
# with damaged
X = mapping.get_features_park(windpark, feature_window, horizon)
Y = mapping.get_labels_turbine(target, feature_window, horizon)
train_to = int(math.floor(len(X)))
train_step, test_step = 1, 1
reg = linear_model.LinearRegression()
reg = reg.fit(X[0:train_to:train_step], Y[0:train_to:train_step])
# USE THE 2005 YEAR FOR TESTING, WITHOUT DAMAGE
# predict on second year without damage
turbines = windpark_test.get_turbines()
for t in turbines:
t.add_measurements(repair_nrel(t.get_measurements()[:10000]))
target_test = windpark_test.get_target()
XT = mapping.get_features_park(windpark_test, feature_window, horizon)
test_to = int(math.floor(len(XT)))
YT = mapping.get_labels_turbine(target_test, feature_window,
horizon)[:test_to]
y_hat = reg.predict(XT[:test_to])
mse_y_hat = 0
for i in range(0, len(y_hat)):
y = YT[i]
mse_y_hat += (y_hat[i] - y)**2
mse_y_hat /= float(len(y_hat))
return mse_y_hat
def static_topologic(windpark):
target = windpark.get_target()
measurements = repair_nrel(target.get_measurements()[:10000])
tloc = (target.longitude, target.latitude)
neighbors = windpark.get_turbines()[:-1]
nseries = [repair_nrel(t.get_measurements()[:10000]) for t in neighbors]
nlocs = [(t.longitude, t.latitude) for t in neighbors]
iargs = {'method' : 'topologic',
'timestep': timestep,
'location': tloc,
'neighbor_series': nseries,
'neighbor_locations': nlocs}
return iargs
def experiment(method, windpark, windpark_test, damaged, rate):
args, nseries = argfuncs[method](windpark)
reconstructed = interpolate(damaged, **args)
target = windpark.get_target()
measurements = repair_nrel(target.get_measurements()[:10000])
turbines = windpark.get_turbines()
for t in range(len(turbines)):
turbines[t].add_measurements(\
repair_nrel(turbines[t].get_measurements()[:10000]))
# this is the target turbine, use the reconstructed here.
turbines[-1].add_measurements(reconstructed)
feature_window, horizon = 3,3
mapping = PowerMapping()
# with damaged
X = mapping.get_features_park(windpark, feature_window, horizon)
Y = mapping.get_labels_turbine(target, feature_window, horizon)
train_to = int(math.floor(len(X)))
train_step, test_step = 1, 1
reg = linear_model.LinearRegression()
reg = reg.fit(X[0:train_to:train_step], Y[0:train_to:train_step])
# USE THE 2005 YEAR FOR TESTING, WITHOUT DAMAGE
# predict on second year without damage
turbines = windpark_test.get_turbines()
for t in turbines:
t.add_measurements(repair_nrel(t.get_measurements()[:10000]))
target_test = windpark_test.get_target()
XT = mapping.get_features_park(windpark_test, feature_window, horizon)
test_to = int(math.floor(len(XT)))
YT = mapping.get_labels_turbine(target_test, feature_window, horizon)[:test_to]
y_hat = reg.predict(XT[:test_to])
mse_y_hat = 0
for i in range(0, len(y_hat)):
y = YT[i]
mse_y_hat += (y_hat[i] - y) ** 2
mse_y_hat /= float(len(y_hat))
return mse_y_hat
def static_mreg_knn(windpark):
target = windpark.get_target()
measurements = repair_nrel(target.get_measurements()[:10000])
neighbors = windpark.get_turbines()[:-1]
nseries = [repair_nrel(t.get_measurements()[:10000]) for t in neighbors]
iargs = {
"method": "mreg",
"timestep": timestep,
"neighbor_series": nseries,
"reg": "knn",
"regargs": {"kfold": 5, "n": [5, 10, 20, 50, 100], "variant": "uniform"},
}
return iargs, nseries
def static_topologic(windpark):
target = windpark.get_target()
measurements = repair_nrel(target.get_measurements()[:10000])
tloc = (target.longitude, target.latitude)
neighbors = windpark.get_turbines()[:-1]
nseries = [repair_nrel(t.get_measurements()[:10000]) for t in neighbors]
nlocs = [(t.longitude, t.latitude) for t in neighbors]
iargs = {
"method": "topologic",
"timestep": timestep,
"location": tloc,
"neighbor_series": nseries,
"neighbor_locations": nlocs,
}
return iargs
def static_mreg_svr(windpark):
target = windpark.get_target()
measurements = repair_nrel(target.get_measurements()[:10000])
neighbors = windpark.get_turbines()[:-1]
nseries = [repair_nrel(t.get_measurements()[:10000]) for t in neighbors]
reg = linear_model.LinearRegression()
gamma_range = [0.0001, 0.000001]
C_range = [2 ** i for i in range(1, 4, 1)]
regargs = {
"epsilon": 0.1,
"cv_method": KFold,
"cv_args": {"n_folds": 10},
"kernel": "rbf",
"tuned_parameters": [{"kernel": [kernel], "C": C_range, "gamma": gamma_range}],
}
iargs = {"method": "mreg", "timestep": timestep, "neighbor_series": nseries, "reg": reg, "regargs": regargs}
return iargs
mse_y_hat /= float(len(y_hat))
return mse_y_hat
# generating destroyed measurements which are constant over all
# methods
data = []
for park in parks.keys():
windpark = NREL().get_windpark_nearest(parks[park], 5, 2004)
windpark_test = NREL().get_windpark_nearest(parks[park], 5, 2005)
target = windpark.get_target()
measurements = repair_nrel(target.get_measurements()[:10000])
for i in range(2):
damaged_series = {}
de = lambda rate: (rate, (destroy(
measurements, method=destroy_method, percentage=rate)[0]))
dseries = map(de, rates)
for rate, series in dseries:
damaged_series[rate] = series
# with reconstruction
def run(pars):
method, rate = pars
def experiment(method, windpark, damaged, rate):
args = argfuncs[method](windpark)
reconstructed = interpolate(damaged, **args)
error, var, std = scores(measurements, reconstructed)
return error, var, std
# generating destroyed measurements which are constant over all
# methods
data = []
park = 'reno'
windpark = NREL().get_windpark_nearest(parks[park], 5, 2004)
target = windpark.get_target()
measurements = repair_nrel(target.get_measurements()[:10000])
for i in range(2):
damaged_series = {}
de = lambda rate : (rate, (destroy(measurements, method=destroy_method, percentage=rate)[0]))
dseries = map(de, rates)
for rate, series in dseries:
damaged_series[rate] = series
def run(pars):
method, rate = pars
error, var, std = experiment(method, windpark, damaged_series[rate], rate)
return method, rate, error, var, std
results = map(run, list(chain(product(methods, rates))))
