-
Notifications
You must be signed in to change notification settings - Fork 4
/
Copy pathnsqMAIN.py
197 lines (187 loc) · 12.1 KB
/
nsqMAIN.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
import os
import numpy as np
import multiprocessing as mp
from multiprocessing import Pool
from pypower.ppoption import ppoption
from pypower.runopf import runopf
from scipy.sparse import lil_matrix
from pypower.case24_ieee_rts import case24_ieee_rts
import pandas as pd
import copy
## Import data format from Pypower
from pypower.idx_bus import PD, QD
def case24_failrate():
failrate = {"genmttf": [], "genmttr": [], "genweeks": [], "brlambda": [], "brdur": []} # 建立dict存放故障数据
failrate["genmttf"] = np.array([450, 450, 1960, 1960, 450,
450, 1960, 1960, 1200, 1200,
1200, 950, 950, 950, 10000,
2940, 2940, 2940, 2940, 2940,
960, 960, 1100, 1100, 1980,
1980, 1980, 1980, 1980, 1980,
960, 960, 1150]) # 33个机组的平均正常运行时间
failrate["genmttr"] = np.array([50, 50, 40, 40, 50,
50, 40, 40, 50, 50,
50, 50, 50, 50, 0.1,
60, 60, 60, 60, 60,
40, 40, 150, 150, 20,
20, 20, 20, 20, 20,
40, 40, 100]) # 33个机组的平均故障修复时间
failrate["genweeks"] = np.array([2, 2, 3, 3, 2,
2, 3, 3, 3, 3,
3, 4, 4, 4, 0.1,
2, 2, 2, 2, 2,
4, 4, 6, 6, 2,
2, 2, 2, 2, 2,
4, 4, 5]) # 33台机组的计划维护时间
failrate["brlambda"] = np.array([0.24, 0.51, 0.33, 0.39, 0.48, 0.38,
0.02, 0.36, 0.34, 0.33, 0.30, 0.44,
0.44, 0.02, 0.02, 0.02, 0.02, 0.40,
0.39, 0.40, 0.52, 0.49, 0.38, 0.33,
0.41, 0.41, 0.41, 0.35, 0.34, 0.32,
0.54, 0.35, 0.35, 0.38, 0.38, 0.34,
0.34, 0.45]) # 38条线路平均一年故障次数
failrate["brdur"] = np.array([16, 10, 10, 10, 10, 768, 10, 10, 35, 10, 10, 10,
10, 768, 768, 768, 768, 11, 11, 11, 11, 11, 11, 11,
11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11,
11, 11]) # 38条线路故障平均持续时间
return failrate
def failprob():
failrate = case24_failrate()
probgen = failrate["genmttr"] / (failrate["genmttf"]+ failrate["genmttr"]) # 每台机组的事故停机率
brmiu = 8760 / failrate["brdur"]
probbr = failrate["brlambda"] / (failrate["brlambda"] + brmiu) # 计算线路的事故停机率=年停机频率*停机时间/一年
totalprob = np.hstack((probgen, probbr)) # 33个机组故障率和38条线路的故障率共71个故障率数据
return totalprob
def mc_sampling(totalprob, SIMUNIT, Ng, Nl):
eqstatus = np.random.rand(SIMUNIT, Ng + Nl) < np.tile(totalprob, (SIMUNIT,1))
# 生成随机数与故障率进行比较,一次mc_sampling抽样100次，eqstatus为1表示原件故障，为0表示原件正常
eqstatus[:, 14] = 0 # 忽略节点14的机组的故障率
"""""
# eqstatus = sparse(doubl(eqstatus))
此处未转化稀疏矩阵
"""
return eqstatus
def mc_simulation(para):
ppopt = para[2]
Ng = para[3]
Nl = para[4]
statusgen = para[0][0: Ng]
Testsys = para[1]
Testsys["gen"][0: Ng, 7] = 1-statusgen # 此处已改
statusbranch = para[0][Ng:Ng+Nl]
Testsys["branch"][0:Nl, 10] = 1-statusbranch
Result =runopf(casedata=Testsys, ppopt=ppopt)
dns = Result["f"] + Testsys["load"]
if dns < 0.1:
dns = 0
return dns
class NonSequentialMonteCarlo():
def __init__(self):
self.pwd = os.getcwd()
self.n_processors = os.cpu_count() # 表明程序开始
def run( self, Testsys = case24_ieee_rts(), BETAlimit = 0.0017, ITER_max = 10000, SIMUNIT = 1000):
Nb = Testsys["bus"].shape[0] # Load test system，Nb为节点数，Ng为发电机组数，Nl为馈线数
Ng = Testsys["gen"].shape[0]
Nl = Testsys["branch"].shape[0]
# Set initial value
iter = 0
betavalue = float('inf') # The stopping criteria停止迭代的标准
row_index = 0
# Build matrices that have fix dimension to avoid changing size in each loop
eqstatus_total = np.zeros((ITER_max, Ng + Nl + 3))# 建一个100000*（33+38+3）的矩阵
beta_table = np.zeros((1, ITER_max // SIMUNIT)) # "//"除法得到的才是整数
edns_table = np.zeros((1, ITER_max // SIMUNIT)) # 存放评价指标，大小为1*1000
lole_table = np.zeros((1, ITER_max // SIMUNIT))
plc_table = np.zeros((1, ITER_max // SIMUNIT))
genbus = np.nonzero((Testsys["bus"][:, PD]))[0] # 第三列（python中坐标是2）是节点的有功功率，表示节点有有功负荷,此处返回该列非零元素的索引，共有17个元素非零
sizegenbus = genbus.shape[0] # 有负荷的节点数量赋值给sizegenbus
Testsys["load"] = sum(Testsys["bus"][:, PD]) # 系统需要的总有功功率
Testsys["gencost"] = np.tile([2, 0, 0, 3, 0, 0, 0], (Ng, 1)) # np.tile建立重复矩阵块（设置机组费用）
# treat all load as negtive generator and set their parameters, then add these vitual generators to real gens
# 将所有载荷视为负发电机，并设置其参数，然后将这些发电机加到实际的发电机中
loadcost = np.tile([2, 0, 0, 3, 0, 1, 0], (sizegenbus, 1)) # np.tile建立重复矩阵块（负荷的“机组费用”）
Testsys["gencost"] = np.append(Testsys["gencost"], loadcost, axis=0)
Index = copy.deepcopy(Testsys["gen"][0:sizegenbus, :]) # 将前17台机组的数据取出
Index[:, 0:10] = np.hstack((Testsys["bus"][genbus, 0].reshape(-1, 1), -Testsys["bus"][genbus, 2].reshape(-1, 1),
-Testsys["bus"][genbus, 3].reshape(-1, 1), np.zeros((sizegenbus, 1)),
-Testsys["bus"][genbus, 3].reshape(-1, 1), np.zeros((sizegenbus, 1)),
Testsys["baseMVA"] * np.ones((sizegenbus, 1)), np.ones((sizegenbus, 1)),
np.zeros((sizegenbus, 1)), -Testsys["bus"][genbus, 2].reshape(-1, 1)))
# 负荷参数代替取出的机组数据，将负荷套入机组模型,上面矩阵取数注意与matlab相比坐标要减一
Testsys["gen"] = np.append(Testsys["gen"], Index, axis=0)
del Index
Testsys["bus"][genbus, 2:4] = 0 # 将原来节点中的第3、4列（有功、无功）负荷设为零
totalprob = failprob() # 引用前面定义的函数
ppopt = ppoption(PF_DC=1, VERBOSE=0, OUT_ALL=0, OPF_ALG_DC=200, OPF_FLOW_LIM=1) # 可以通过ppoption()采用默认变量来看里面需要什么样的输入，这个按照matlab来输入没问题吧？
result = runopf(casedata=Testsys, ppopt=ppopt)
while (betavalue > BETAlimit) & (iter < ITER_max):
eqstatus_indi = mc_sampling(totalprob, SIMUNIT, Ng, Nl) # eqstatus为元件的状态矩阵，为1表示元件故障，为0表示原件正常
eqstatus_indi = np.hstack((eqstatus_indi, np.ones((eqstatus_indi.shape[0], 1)), np.zeros((eqstatus_indi.shape[0], 2))))
# 在eqstatus_indi矩阵中加入三列，第一列代表状态重复次数，第二列记载切负荷量大小（没有切负荷则为零），第三列记载是否为容量不足
eqstatus_indi, ia1 = np.unique(eqstatus_indi, axis=0, return_inverse=True) # 找出抽样中的相同结果
for i in range(eqstatus_indi.shape[0]):
eqstatus_indi[i, Ng + Nl] = sum(ia1 == i) # 将重复记录次数在第Ng + Nl + 1
if iter:
x = 0
y = eqstatus_indi.shape[0]
for i in range(y):
indi_x = eqstatus_indi[x, 0:Ng + Nl]
for j in range(row_index):
if (indi_x == eqstatus_total[j, 0:Ng + Nl]).all():
eqstatus_total[j, Ng + Nl] = eqstatus_total[j, Ng + Nl] + eqstatus_indi[x, Ng + Nl] # 遇见相同的，就在eqstatus_total的计数中累加次数
eqstatus_indi = np.delete(eqstatus_indi, x, axis=0)
x = x - 1
break
x = x + 1
parfortemp = np.zeros((eqstatus_indi.shape[0], 2))
para = [0] * eqstatus_indi.shape[0]
n_sample = [0] * eqstatus_indi.shape[0]
for i in range(eqstatus_indi.shape[0]):
para[i] = [0] * 5
para[i][0] = eqstatus_indi[i, 0: Ng + Nl]
para[i][1] = Testsys
para[i][2] = ppopt
para[i][3] = Ng
para[i][4] = Nl
with Pool(self.n_processors) as p:
load_shedding = list(p.map(mc_simulation, para))
parfortemp[:, 0] = np.asarray(load_shedding)
parfortemp[:, 1] = (parfortemp[:, 0]) != 0
eqstatus_indi[:, Ng + Nl + 1: Ng + Nl + 3] = parfortemp
eqstatus_total[row_index: row_index + eqstatus_indi.shape[0], :] = eqstatus_indi
row_index = row_index + eqstatus_indi.shape[0]
else:
parfortemp = np.zeros((eqstatus_indi.shape[0], 2))
para = [0] * eqstatus_indi.shape[0]
c = [0] * eqstatus_indi.shape[0]
for i in range(eqstatus_indi.shape[0]):
para[i] = [0] * 5
para[i][0] = eqstatus_indi[i, 0: Ng + Nl]
para[i][1] = Testsys
para[i][2] = ppopt
para[i][3] = Ng
para[i][4] = Nl
with Pool(self.n_processors) as p:
load_shedding = list(p.map(mc_simulation, para)) # 计算负荷短缺值
parfortemp[:, 0] = np.asarray(load_shedding)
parfortemp[:, 1] = (parfortemp[:, 0]) != 0 # 记录是否负荷短缺
eqstatus_indi[:, Ng + Nl + 1: Ng + Nl + 3] = parfortemp
eqstatus_total[row_index: row_index + eqstatus_indi.shape[0], :] = eqstatus_indi
row_index = row_index + eqstatus_indi.shape[0]
## Update index
edns = sum(eqstatus_total[0:row_index, Ng + Nl] * eqstatus_total[0: row_index, Ng + Nl + 1]) / (iter + SIMUNIT)
lole = sum(eqstatus_total[0:row_index, Ng + Nl] * eqstatus_total[0: row_index, Ng + Nl + 2]) / (
iter + SIMUNIT) * 8760
plc = sum(eqstatus_total[0:row_index, Ng + Nl] * eqstatus_total[0: row_index, Ng + Nl + 2]) / (iter + SIMUNIT)
betavalue = (sum(eqstatus_total[0:row_index, Ng + Nl]*(
eqstatus_total[0:row_index, Ng+Nl+1] - edns)**2)) ** 0.5/(iter + SIMUNIT) / edns
beta_table[0, ((iter + SIMUNIT) // SIMUNIT)-1] = betavalue
edns_table[0, ((iter + SIMUNIT) // SIMUNIT)-1] = edns
lole_table[0, ((iter + SIMUNIT) // SIMUNIT)-1] = lole
plc_table[0, ((iter + SIMUNIT) // SIMUNIT)-1] = plc
iter = iter + SIMUNIT
return edns
if __name__=="__main__":
## Start the function
non_sequential_monte_carlo = NonSequentialMonteCarlo()
non_sequential_monte_carlo.run()