-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathDataManagement.py
199 lines (157 loc) · 9 KB
/
DataManagement.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
import pandas as pd
import numpy as np
import itertools as it
import matplotlib.pyplot as plt
def get_data(deep_copy = True):
return data.copy(deep_copy)
def get_features(deep_copy = True):
# data have to be stored in a pandas DataFrame
# build feature Normalization
#GHI_kt = GHI.div(ENI.mul(np.sin(np.deg2rad(El))))
BNI_kt = BNI_train.div(ENI_train)
B_BNI_kt_1 = BNI_kt
B_BNI_kt = BNI_kt
B_GHI_kt_1 = GHI_KT_train
B_GHI_kt = GHI_KT_train
featuresB_GHI = pd.DataFrame()
featuresB_BNI = pd.DataFrame()
featuresL_GHI = pd.DataFrame()
featuresL_BNI = pd.DataFrame()
featuresV_GHI = pd.DataFrame()
featuresV_BNI = pd.DataFrame()
featuresL_GHI.insert(0, column='L_GHI_kt_0', value=B_GHI_kt)
featuresL_BNI.insert(0, column='L_BNI_kt_0', value=B_BNI_kt)
Pdc_shift = pd.DataFrame()
for col,shift in zip(range(0, window), range(-1, -(window+1), -1)):
featuresB_GHI.insert(col, column='B_GHI_kt_%i' % col, value=B_GHI_kt)
featuresB_BNI.insert(col, column='B_BNI_kt_%i' % col, value=B_BNI_kt)
BGHI_shift = B_GHI_kt_1.shift(periods=col+1)
BBNI_shift = B_BNI_kt_1.shift(periods=col+1)
clmn = col + 1
featuresL_GHI.insert(clmn, column='L_GHI_kt_%i' % clmn, value=BGHI_shift)
featuresL_BNI.insert(clmn, column='L_BNI_kt_%i' % clmn, value=BBNI_shift)
B_GHI_kt = featuresL_GHI.mean(axis=1)
B_BNI_kt = featuresL_BNI.mean(axis=1)
featuresL_GHI = featuresL_GHI.drop('L_GHI_kt_6', axis=1)
featuresL_BNI = featuresL_BNI.drop('L_BNI_kt_6', axis=1)
GHI_kt_mean = featuresB_GHI["B_GHI_kt_5"].to_frame()
BNI_kt_mean = featuresB_BNI["B_BNI_kt_5"].to_frame()
for col in range(0, window):
delta_kt_GHI = featuresL_GHI.iloc[:, 0:col+1].sub(GHI_kt_mean.values.reshape(len(GHI_kt_mean), col+1)).pow(2)
delta_kt_BNI = featuresL_BNI.iloc[:, 0:col+1].sub(BNI_kt_mean.values.reshape(len(BNI_kt_mean), col+1)).pow(2)
GHI_kt_mean.insert(col+1, column="times_%i" % col, value=GHI_kt_mean.B_GHI_kt_5.values)
BNI_kt_mean.insert(col+1, column="times_%i" % col, value=BNI_kt_mean.B_BNI_kt_5.values)
V_GHI_kt = pd.Series(np.sqrt(np.divide(delta_kt_GHI.sum(axis=1), col+1)))
V_BNI_kt = pd.Series(np.sqrt(np.divide(delta_kt_BNI.sum(axis=1), col+1)))
featuresV_GHI.insert(col, column='V_GHI_kt_%i' % col, value=V_GHI_kt)
featuresV_BNI.insert(col, column='V_BNI_kt_%i' % col, value=V_BNI_kt)
#Pdc_norm = Pdc.div(np.max(Pdc))
#Pdc_norm = Pdc.div(ENI.mul(Capacity))
Pdc_norm = Pdc_train.div(ENI_train)
for col in range(0, window):
Pdc_shift.insert(col, column='Pdc_{}min'.format(delta*(col+1)), value=Pdc_norm)
Pdc_norm = Pdc_norm.shift(periods=1) # check whether periods 1 or -1
Pdc_shift = pd.concat([Pdc_shift, ENI_train, El_train], axis=1)
features = pd.concat([time_train, featuresB_GHI, featuresB_BNI, featuresV_GHI, featuresV_BNI,
featuresL_GHI, featuresL_BNI, Pdc_shift, Pdc_train], axis=1)
return features.copy(deep_copy)
def get_target(deep_copy = True):
BNI_kt = BNI_test.div(ENI_test)
B_BNI_kt_1 = BNI_kt
B_BNI_kt = BNI_kt
B_GHI_kt_1 = GHI_KT_test
B_GHI_kt = GHI_KT_test
featuresB_GHI = pd.DataFrame()
featuresB_BNI = pd.DataFrame()
featuresL_GHI = pd.DataFrame()
featuresL_BNI = pd.DataFrame()
featuresV_GHI = pd.DataFrame()
featuresV_BNI = pd.DataFrame()
featuresL_GHI.insert(0, column='L_GHI_kt_0', value=B_GHI_kt)
featuresL_BNI.insert(0, column='L_BNI_kt_0', value=B_BNI_kt)
Pdc_shift = pd.DataFrame()
for col, shift in zip(range(0, window), range(-1, -(window + 1), -1)):
featuresB_GHI.insert(col, column='B_GHI_kt_%i' % col, value=B_GHI_kt)
featuresB_BNI.insert(col, column='B_BNI_kt_%i' % col, value=B_BNI_kt)
BGHI_shift = B_GHI_kt_1.shift(periods=col + 1)
BBNI_shift = B_BNI_kt_1.shift(periods=col + 1)
clmn = col + 1
featuresL_GHI.insert(clmn, column='L_GHI_kt_%i' % clmn, value=BGHI_shift)
featuresL_BNI.insert(clmn, column='L_BNI_kt_%i' % clmn, value=BBNI_shift)
B_GHI_kt = featuresL_GHI.mean(axis=1)
B_BNI_kt = featuresL_BNI.mean(axis=1)
featuresL_GHI = featuresL_GHI.drop('L_GHI_kt_6', axis=1)
featuresL_BNI = featuresL_BNI.drop('L_BNI_kt_6', axis=1)
GHI_kt_mean = featuresB_GHI["B_GHI_kt_5"].to_frame()
BNI_kt_mean = featuresB_BNI["B_BNI_kt_5"].to_frame()
for col in range(0, window):
delta_kt_GHI = featuresL_GHI.iloc[:, 0:col + 1].sub(GHI_kt_mean.values.reshape(len(GHI_kt_mean), col + 1)).pow(
2)
delta_kt_BNI = featuresL_BNI.iloc[:, 0:col + 1].sub(BNI_kt_mean.values.reshape(len(BNI_kt_mean), col + 1)).pow(
2)
GHI_kt_mean.insert(col + 1, column="times_%i" % col, value=GHI_kt_mean.B_GHI_kt_5.values)
BNI_kt_mean.insert(col + 1, column="times_%i" % col, value=BNI_kt_mean.B_BNI_kt_5.values)
V_GHI_kt = pd.Series(np.sqrt(np.divide(delta_kt_GHI.sum(axis=1), col + 1)))
V_BNI_kt = pd.Series(np.sqrt(np.divide(delta_kt_BNI.sum(axis=1), col + 1)))
featuresV_GHI.insert(col, column='V_GHI_kt_%i' % col, value=V_GHI_kt)
featuresV_BNI.insert(col, column='V_BNI_kt_%i' % col, value=V_BNI_kt)
# Pdc_norm = Pdc.div(np.max(Pdc))
# Pdc_norm = Pdc.div(ENI.mul(Capacity))
Pdc_norm = Pdc_test.div(ENI_test)
for col in range(0, window):
Pdc_shift.insert(col, column='Pdc_{}min'.format(delta * (col + 1)), value=Pdc_norm)
Pdc_norm = Pdc_norm.shift(periods=1) # check whether periods 1 or -1
Pdc_shift = pd.concat([Pdc_shift, ENI_test, El_test], axis=1)
target = pd.concat([time_test, featuresB_GHI, featuresB_BNI, featuresV_GHI, featuresV_BNI,
featuresL_GHI, featuresL_BNI, Pdc_shift, Pdc_test], axis=1)
return target.copy(deep_copy)
global Capacity
filename = 'Daten/PVAMM_201911-202011_PT5M_merged.csv'
data = pd.read_csv(filename)
data_min = data
Capacity = 20808.66
window = 6 # window size * delta[min] = forecasting period
delta = 5 # step size [min]
for Irr in ['GHI', 'DHI', 'gti30t187a', 'ENI']: # 'Pdc_33'
data_min = data_min.drop(data_min.index[data_min[Irr] == 0])
data_min = data_min.dropna(subset=['GHI', 'BNI', 'DHI', 'gti30t187a', 'ENI', 'El'])
first = ["Sep", "Dec", "Mar", "Jun"]
second = ["Oct", "Jan", "Apr", "Jul"]
third = ["Nov", "Feb", "May", "Aug"]
for t in [first, second, third]:
herbst = data_min[data_min.t.str.contains(t[0])]
winter = data_min[data_min.t.str.contains(t[1])]
spring = data_min[data_min.t.str.contains(t[2])]
summer = data_min[data_min.t.str.contains(t[3])]
"""train = pd.concat([herbst[0:int(len(herbst) * 0.8)], winter[0:int(len(winter) * 0.8)],
spring[0:int(len(spring) * 0.8)], summer[0:int(len(summer) * 0.8)]], axis=0)
test = pd.concat([herbst[int(len(herbst) * 0.8):len(herbst)], winter[int(len(winter) * 0.8):len(winter)],
spring[int(len(spring) * 0.8):len(spring)], summer[int(len(summer) * 0.8):len(summer)]], axis=0)"""
train, test = data_min[0:round(len(data_min)*0.8)], data_min[round(len(data_min)*0.8):len(data_min)]
time_train, time_test = train.t, test.t
GHI_train, GHI_test = train.GHI, test.GHI
BNI_train, BNI_test = train.BNI, test.BNI
ENI_train, ENI_test = train.ENI, test.ENI
El_train, El_test = train.El, test.El
Pdc_train, Pdc_test = train.Pdc_33, test.Pdc_33
GHI_KT_train, GHI_KT_test = train.kt, test.kt
""" first = ["Sep", "Dec", "Mar", "Jun"]
second = ["Oct", "Jan", "Apr", "Jul"]
third = ["Nov","Feb", "May", "Aug"]
for t in [first, second, third]:
herbst = features[features.t.str.contains(t[0])]
winter = features[features.t.str.contains(t[1])]
spring = features[features.t.str.contains(t[2])]
summer = features[features.t.str.contains(t[3])]
tarherbst = tar[tar.t.str.contains(t[0])]
tarwinter = tar[tar.t.str.contains(t[1])]
tarspring = tar[tar.t.str.contains(t[2])]
tarsummer = tar[tar.t.str.contains(t[3])]
train = pd.concat([herbst[0:int(len(herbst)*0.8)], winter[0:int(len(winter)*0.8)],
spring[0:int(len(spring)*0.8)], summer[0:int(len(summer)*0.8)]], axis=0)
test = pd.concat([herbst[int(len(herbst)*0.8):len(herbst)], winter[int(len(winter)*0.8):len(winter)],
spring[int(len(spring)*0.8):len(spring)], summer[int(len(summer)*0.8):len(summer)]], axis=0)
train_y = pd.concat([tarherbst[0:int(len(tarherbst)*0.8)], tarwinter[0:int(len(tarwinter)*0.8)],
tarspring[0:int(len(tarspring)*0.8)], tarsummer[0:int(len(tarsummer)*0.8)]], axis=0)
test_y = pd.concat([tarherbst[int(len(tarherbst)*0.8):len(tarherbst)], tarwinter[int(len(tarwinter)*0.8):len(tarwinter)],
tarspring[int(len(tarspring)*0.8):len(tarspring)], tarsummer[int(len(tarsummer)*0.8):len(tarsummer)]], axis=0)"""