Académique Documents
Professionnel Documents
Culture Documents
Energy
journal homepage: www.elsevier.com/locate/energy
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 19 May 2016
Received in revised form
14 October 2016
Accepted 19 October 2016
In this paper, a stochastic model is proposed for coordinated scheduling of combined heat and power
units in micro grid considering wind turbine and photovoltaic units. Uncertainties of electrical market
price; the speed of wind and solar radiation are considered using a scenario-based method. In the
method, scenarios are generated using roulette wheel mechanism based on probability distribution
functions of input random variables. Using this method, the probabilistic specics of the problem are
distributed and the problem is converted to a deterministic one. The type of the objective function,
coordinated scheduling of combined heat and power, wind turbine, and photovoltaic units change this
problem to a mixed integer nonlinear one. Therefore to solve this problem modied particle swarm
optimization algorithm is employed. The mentioned uncertainties lead to an increase in prot. Moreover,
the optimal coordinated scheduling of renewable energy resources and thermal units in micro grids
increase the total prot. In order to evaluate the performance of the proposed method, its performance is
executed on modied 33 bus distributed system as a micro grid.
2016 Elsevier Ltd. All rights reserved.
Keywords:
Micro grid
Optimal scheduling
Deregulated electricity market
Wind turbine
Photovoltaic unit
Combined heat and power
1. Introduction
Nowadays, the majority of the electrical energy is supplied by
nuclear or large fossil fuel power plants around the world and they
have high reliability and capacity. However, regarding the fact that
the investment cost of these power plants is very high, in order to
draw the attention of small capitals and reducing environmental
emissions, Renewable Energy Sources (RESs) at small scales attract
a higher interest [1]. On the other hand, increasing the use of
renewable energy and Distributed Generation (DG) sources of high
efciency in power systems has provided new points of view. This
new viewpoint, i.e. generation units generating power in a smaller
scale, is close to consumers in the Micro Grid (MG). MGs comprise
distribution networks with distributed energy resources: i.e. Wind
List of abbreviations: CHP, Combined Heat and Power; MG, Micro Grid; WT,
Wind Turbine; PV, Photovoltaic; MPSO, Modied Particle Swarm Optimization algorithm; DG, Distributed Generation; TH, Thermal unit; FC, Fuel Cell; RES,
Renewable Energy Source.
* Corresponding author.
E-mail addresses: mbornapour@eng.ui.ac.ir (M. Bornapour), hooshmand_r@eng.
ui.ac.ir (R.-A. Hooshmand), aminkh@eng.ui.ac.ir (A. Khodabakhshian), parastegari@
eng.ui.ac.ir (M. Parastegari).
http://dx.doi.org/10.1016/j.energy.2016.10.072
0360-5442/ 2016 Elsevier Ltd. All rights reserved.
Turbine (WT), Fuel Cells (FCs) based Combined Heat and Power
(CHP), Photovoltaic (PV), etc. [2,3]. MGs can be managed in a nonautonomous mode, if interconnected to the main grid, or in an
autonomous mode, if disconnected from the main grid [4]. The
operation of DGs in the network can provide different benets for
the overall network performance, if optimally scheduled and coordinated [4]. One of the most important issues in MGs is optimal
operation in deregulated markets by determining the best value
generation of each unit to achieve the highest prot. In this regard,
distributed generators that are capable of CHP production are
increasingly used in MGs as well as in RESs such as solar and wind
energy sources which have stochastic nature [5]. Since, the generation of RESs is uncertain, their generation should be forecasted; on
the other hand, the uncertainties of load and RESs are considered in
Ref. [6].
Many studies have been carried out on the performance of MG
and prot maximization objective function [7e18]. The pool electricity market of MG for PV and FC units has also been investigated
in Ref. [7]. In this study, the prot maximization objective function
is considered and genetic optimization algorithm is used to solve
the problem. In Ref. [8], the optimal operation and participation of
MGs in electricity market are proposed by considering energy
sources such as WT, PV, and FC, while adopting prot-maximizing
Nomenclature:
CBuy,t,s
cost of buying electrical energy in sth scenario
cvf
coefcient of variation
0
DTs;s0
the distance between scenarios s and s
fTH,h,fFCCHP,l,fWind,j, fPV,q cost functions related to thermal, CHP,
WT, PV units respectively
HBoiler,l,t,s the heat generated by the boiler can be parallel CHP
units in sth scenario
HCHP,l,t,s heat generated by CHP units in sth scenario
HDemand,t,s thermal load demand in sth scenario
Loadt,s
electrical load in time t in sth scenario
Nbus
total number of buses
NFC
total number of FCs
Ns
number of scenarios of each batch
PBuy,t,s
amount of electrical power bought from market in sth
scenario
PenaltyMarket, PenaltyDemand penalty factor for market and
demand
t
PFCj;s
active power generated by FCj during time t in sth
scenario
t
PHj;s
equivalent electric power for hydrogen production
during time t in sth scenario
PLRj
part load ratio of FCj (equal to electrical generated
power/maximum power)
PMax,FC maximum power of FC
min
max
PUnit;i;t;s
, PUnit;i;t;s
minimum and maximum of active power
produced by Uniti during time t in sth scenario
Pnot-supplied,Market,t,s not-supplied power on the market at the
time t in sth scenario
PSell,t,s,PH,Sell,t,s electrical power and heat energy sold to market in
sth scenario
177
t
Pthi;s
178
Costs
24
X
CBuy;t;s PBuy;t;s
t1
N
CHP
X
NTH
24 X
X
24
X
fTH;h PTH;h;t;s
t1 h1
t1
fFCCHP;l PFCCHP;l;t;s ; HFCCHP;l;t;s ; HBoiler;l;t;s
l1
24 NX
Wind
X
NPV
24 X
X
fWT;j PWT;j;t;s
fPV;q PPV;q;t;s
t1 j1
t1 q1
PenaltyMarket s PenaltyDemand s
(5)
where each term of cost function is calculated as follows:
fTH;h PTH;h;t;s CTH;s $PTH;h;t;s
(6)
fFCCHP;l PFCCHP;l;t;s ; HFCCHP;l;t;s ; HBoiler;l;t;s
2. Problem formulation
fWT;j PWT;j;t;s CWT;s $PWT;j;t;s
(8)
fPV;q PPV;q;t;s CPV;s $PPV;q;t;s
(9)
:::
X PFC;1;t;s
PFC;NFC ;t;s
PTH;1;t;s
:::
PTH;NTH ;t;s
PenaltyMarket s
24
X
(7)
Pnotsupplied;Market;t;s ppenalti;Market;t;s
t1
(10)
(
ppenalti;Market;t;s
1 b$pSell;t;s
1 b$pSell;t;s
~
PSell;t;s > P
Sell;t;s
~
PSell;t;s < P
Sell;t;s
(11)
NTH NFC T
(1)
2.3. Constraints
2.2. Objective function of market prot
Maximizing prots (expected prot-EPF) for the participation of
an MG in the electrical market is considered as the objective
function of this problem. EPF is dened as follows:
MaxEPF
rs $PFs
(2)
s2S
PBuy;t;s
NX
Thermal
PTH;h;t;s
h1
(3)
Revenues
24
X
pTariff ;s Loadt;s
t1
24
X
24
X
psell;t;s PSell;t;s
NPV
X
N
CHP
X
PCHP;l;t;s
l1
NX
Wind
PWind;j;t;s
j1
PPV;q;t;s
q1
Loadt;s Losst;s
ct 1 : 24
(12)
HCHP;l;t;s HBoiler;l;t;s HDemand;t;s
t1
(13)
l1
pH;sell;t;s PH;Sell;t;s
(4)
t1
The proposed cost function includes seven terms which are the
costs of buying electrical energy from market; costs of generated
electrical energy by TH, FC-CHP, WT, and PV units; and penalties for
failure to supply the market and customers. The cost term of prot
function is as follows:
VMin Vit VMax
Output active power of DG units
(14)
min
max
PUnit;i;t;s
PUnit;i;t;s It PUnit;i;t;s
ci2SUnit
t 1 : 24
(15)
Sij;t;s Smax
ij
Et;s Smax
exch
ci j2Sb
t 1 : 24
ct 1 : 24
(16)
(17)
179
PLRj
3. Steady state model for RES in micro grid
For
t
PFCj
PMax;FC
(20)
For
(21)
t
t
t
Pthi
rTEj PFCj
PHj
(22)
t
t
PMax;FC PFCj
PHj
(23)
htoverall
(18)
Without considering FCs as CHP, their efciency is approximately in the range of 30e40%. Employing FCs as CHP noticeably
improves their efciency. The overall efciency of FCs could be
calculated as follows [22]:
t h P t min P t ; Lt
PFC
st
H
th th
(24)
t
PFC
PHt
htFCj
(19)
The electrical energy generated by variable speed wind turbine
is done in two steps: rst, a synchronous or a double-fed generator
applied to their structure converts the kinetic and mechanical
Pr
Vc i
Fig. 1. Performance cure of the FC.
Vr
Vc o
180
8
< 0
j wt
: max
Pw
(25)
PtWT
8
r wt vci ; vci w v1
>
>
> 1
>
r v vci r2 wt v1 ; v1 w v2
>
>
< 1 1
r1 v
1t vci r2 v2 v1
Pmax
w
>
w v2 ; v2 w vr
r
3
>
>
>
>
1 vr w vco
>
:
0 otherwise
(26)
In Eq. (26) vci4 m/s, vr14 m/s, vco25 m/s, v17 m/s, v212 m/
s, r10.2/(v1vci), r2(0.960.2)/(v2v1), r3(10.96)/(vrv2).
Also, ri and vi refer to the slope and the break-point of the ith part of
the wind power generation curve, respectively [24]. Alsovci, vco and
vr are the cut-in, cut-off, and rated speed of the wind turbine,
respectively. For characterizing the proper position of the wind
turbines, it is necessary to predict the wind speed hour by hour in
the next day. By using Eq. (26), the wind power can be obtained
from the estimated information. These results can be used as the
upper bond of the associated decision variable.
3.3. Photovoltaic systems (PV)
PV is arrays of cells containing a material that converts solar
radiation into direct current electricity. PV electric power generation is a promising clean technology with enormous potential [25].
Orbiting satellites were the rst practical application of PVs, but
nowadays it gains much attention to be used as a source of power in
micro grids. There are three available factors inuencing the generation of electric power by PV module including, i.e. solar radiance,
ambient temperature, and module characteristics. The following
relation expresses the conversion function of solar radiance to
electric power [26]:
PPV
t
8
G
>
>
P
>
sn
<
Gstd Rc
t
>
>
>
: Psn G
Gstd
0 < Gt < Rc
(27)
Gt > Rc
Probability
Density
Level 1
4, 5
A thermal generation unit converts heat energy to electric power. In this paper, an important producer of electricity in the MG is
considered alongside other DG units. Micro gas turbines that can be
synchronized with network are used as thermal units in this paper.
Relations related to its cost and revenue (prot objective function)
are brought up in Section 2.
6, 7
Level 2
Level 3
2, 3
Leve 7
Level 4
Level 5
-3
Leve 6
-2
Forecast Error
Fig. 3. Typical discretization of the probability distribution functions of the load level
forecast error.
S S fdg; DS DS fdg; rr rr rd
Accumulated
181
(32)
Probability
Step 5: By cycling steps 2 to 4, the ideal number of mitigated scenarios can be visible.
This scenario-reduction method is performed on basic scenarios; now as a result, the reduced scenarios can be operated for
solving uncertain problems.
Error level
Fig. 4. Accumulated normalized probabilities of the forecast error levels.
cvf
sf
p
mf Ns
(33)
DT
s;s0
r
Xd
s xs0
x
DTxs ; xs0
i
i
i1
s; s0 2S
(28)
(29)
(30)
(31)
James Kennedy was the rst who discovered a global method for
optimization by analyzing the social life of animals that lived
collectively and achieved their purposes like feeding and traveling
by communication [30]. This optimization method that is based on
population is called particle swarm optimization algorithm. In this
algorithm rst of all a set of particles should be characterized. In
fact, these particles are the considered occurrence solutions. Then,
this algorithm produces the modied particles in their new situations by operating the equations of velocity update (34) and position update (35) until the best solution is reached.
k
Velk1
u Velkj C1 rand$ PBest;j
Xjk C2 rand
j
GkBest Xjk
(34)
Xjk1 Xjk Velk1
j
(35)
182
follow: three different particles are selected from the initial population and a mutation operator will be applied to them. Furthermore, other particles are muted by the selected ones to constitute
the population. Thus, in each iteration, an Xj that is unequal to the
three selected particles is selected so as to be proper for all
searching domains. This operation is cycled for all particles of the
population. The following equation shows how to produce a
k
mutant vector ( Xmut;j
):
k
k
k
k
Xmut;j
Xm1
rand$ Xm2
Xm3
k
xmut1 ; xmut2 ; :::; xmutD
Xmut;j
i
h
k
Xnew;j
xknew;1 ; xknew;2 ; :::; xknew;D
(
xkmut;z if Cr rand$
k
xnew;z
xkswarm;z otherwise
No
(37)
start
i=i+1
(36)
No
183
Table 1
Primary data related to prot objective function and network [22,33].
Parameter
Value ($/kWh)
0.005
0.005
0.039
0.039
0.045
0.045
0.05
8
< Xk
new;j if
: Xk
swarm;j
k
k
f Xswarm;j
f Xnew;j
otherwise
(38)
Fig. 7. a) Daily load prole; b) daily market price; c) daily wind speed; d) daily solar radiation during the next day.
184
Step 12: If this algorithm is run, less than the criteria is the determined value, so go to step 6; otherwise, criteria are
stopped. While the iteration counter value is lower than
the considered maximum number refer to step 6; otherwise, the algorithm is ended.
Table 2
Comparison of cost between the results of the proposed method and other ones.
Method
3664
369.76
552.3
358.55
5.1695
3.8323
4.1881
3.1349
104
104
104
104
Fig. 5 illustrates the owchart of the MPSO to solve the coordinated scheduling problem.
Table 3
The values of the prot objective function in different cases.
7. Simulation results
Problem
Case
Prot ($)
Stochastic
Considering CHP
Without considering CHP
Considering CHP
Without considering CHP
871.7179
631.1588
819.6278
534.4823
Deterministic
Step 2: The initial population (Xj) and an initial velocity (Velj) need
to be generated randomly.
Step 3: In this step, the value of the objective function (Eq. (2))
should be determined for each particle by using the distribution load ow.
Step 4: The positions of each particle represent real power output
k )
for all FCs and THs and the best position of particles (PBest
should be determined.
Step 5: The best position among the selected positions in the
previous step, i.e. (GkBest ) should be recognized.
Step 6: Set i 1.
Step 7: Updating the velocity and position based on Eqs. (34) and
(35).
Step 8: Executing of mutation (presented in Section 5.2).
Step 9: If all the individuals are selected, go to step 10; otherwise,
set up i i 1and return to step 7.
k
Step 10: Update PBest
using the method presented in step 4.
Step 11: Update GkBest using the method presented in step 5.
Table 4
Result of deterministic problem without considering CHP.
Time (h)
DG Source (kW)
WT
PV
FC
TH
Substation
Heat of FC-CHP
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
160
160
160
160
160
160
160
462.1638
462.1638
462.1638
462.1638
583.7517
583.7517
583.7517
583.7517
583.7517
705.3517
705.3517
705.3517
399.5155
399.5155
399.5155
399.5155
160
0
0
0
0
0
0
0
317.592
317.592
317.592
317.592
528
528
528
528
528
212.8896
212.8896
212.8896
0
0
0
0
0
224.5823
136.8596
104.9418
60.3242
133.7998
168.9676
96.1852
148.0003
133.6091
74.8991
211.0335
157.5988
139.6033
182.3453
171.1952
117.6049
122.1727
159.2811
10.2055
160.7015
195.4638
160.157
219.1467
219.9871
1196.1055
976.9995
973.112
1042.2979
946.2623
937.9398
1173.9553
872.4261
856.7697
799.0612
1179.1892
944.2178
686.7447
783.1635
990.1472
779.2042
1022.0498
627.4338
897.4739
1150.1086
1147.4562
1197.2123
1229.1108
1249.233
1448.5284
1763.461
1800.1889
1776.0632
1798.9991
1770.2354
1603.528
2009.9074
2040.4523
2164.4311
1629.0833
2772.5082
3056.387
2908.824
2706.2602
2987.3543
2537.9819
2912.0096
2787.0759
3700.9256
3664.902
3650.5944
3552.0702
1806.8473
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1488
1488
1488
1488
1488
1488
1488
1488
1488
1488
1488
1488
1488
1488
1488
1488
1488
1488
1488
1488
1488
1488
1488
1488
1000
PV Generation (kW)
WT Generation (kW)
500
10 12
14
16 18
20
22 24
185
800
600
400
200
0
300
200
100
0
10 12
14
10 12
14
16 18
20
22 24
16 18
20
22 24
Time (h)
TH Generation (kW)
FC Generation (kW)
Time (h)
16 18
20
22 24
1500
1000
500
0
Time (h)
10 12
14
Time (h)
Fig. 8. Generated power by different RESs of the MG in deterministic case with neglecting the CHP effect.
in Section 5.2.
Simulation results are divided into two parts of deterministic
and stochastic. In each part, the simulation is done with and
without considering the CHP effect, separately. In case of with and
without considering CHP effect for supplying the thermal loads, the
shortage of thermal energy is supplied by burning natural gas in
parallel boilers. It should be noted that all DGs generate only active
power and FCs have been considered as CHP.
7.2. Evaluation of proposed method
1000
PV Generation (kW)
WT Generation (kW)
500
10 12
14
16 18
20
22 24
800
600
400
200
0
300
200
100
0
10 12
14
Time (h)
10 12
14
16 18
20
22 24
16 18
20
22 24
Time (h)
TH Generation (kW)
FC Generation (kW)
Time (h)
16 18
20
22 24
1500
1000
500
0
10 12
14
Time (h)
Fig. 9. Generated power by different RESs of the MG in deterministic case with considering the CHP effect.
186
Table 5
Result of deterministic problem considering CHP.
Time (h)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
DG Source (kW)
WT
PV
FC
TH
Substation
Heat of FC-CHP
160
160
160
160
160
160
160
462.16384
462.16384
462.16384
462.16384
583.7517
583.7517
583.7517
583.7517
583.7517
705.3517
705.3517
705.3517
399.5155
399.5155
399.5155
399.5155
160
0
0
0
0
0
0
0
317.592
317.592
317.592
317.592
528
528
528
528
528
212.8896
212.8896
212.8896
0
0
0
0
0
234.967
146.2088
127.3397
131.2292
119.4482
203.6735
205.9898
207.0843
116.0736
159.0779
175.3288
141.5329
126.7371
76.7161
166.6388
123.8297
144.4376
38.5084
125.0347
186.258
207.9963
192.9157
212.9669
224.0838
1253.4201
1003.678
927.011
980.5864
910.8367
997.1158
1069.5104
793.7987
891.9162
718.6969
1219.6333
920.2638
683.6919
828.3859
993.201
773.7523
1015.6939
812.8775
783.8224
1130.6957
1022.7328
1140.8748
1247.5935
1261.9408
1379.0614
1726.2518
1823.7408
1764.8834
1848.8138
1675.2433
1596.3423
2029.8966
2022.8616
2158.7501
1624.3452
2811.171
3075.6087
2980.1301
2710.1312
2983.5348
2520.2596
2842.1882
2781.3019
3694.2167
3783.7317
3674.4298
3539.3743
1789.3543
236.6221
236.6221
236.6221
236.6221
236.6221
236.6221
236.6221
236.6221
236.6221
236.6221
236.6221
236.6221
236.6221
236.6221
236.6221
236.6221
236.6221
236.6221
236.6221
236.6221
236.6221
236.6221
236.6221
236.6221
1251.3779
1251.3779
1251.3779
1251.3779
1251.3779
1251.3779
1251.3779
1251.3779
1251.3779
1251.3779
1251.3779
1251.3779
1251.3779
1251.3779
1251.3779
1251.3779
1251.3779
1251.3779
1251.3779
1251.3779
1251.3779
1251.3779
1251.3779
1251.3779
Table 6
Result of stochastic problem without considering CHP.
Time (h)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
DG Source (kW)
WT
PV
FC
TH
Substation
Heat of FC-CHP
265.8027
651.2537
600.7695
119.6954
185.526
393.3251
174.3731
796.7623
479.4441
413.4723
800
782.5996
800
774.4829
489.2117
465.7514
672.0219
776.1898
800
441.9131
504.5326
800
465.2652
448.3933
0
0
0
0
0
0
0
264.66
396.99
264.66
317.592
660
528
660
396
484
230.6304
266.112
230.6304
0
0
0
0
0
138.7152
189.3716
70.6446
64.172
64.9105
225.9985
160.5914
66.9819
50.3208
75.3329
208.835
167.9726
106.042
155.18758
115.2461
113.921
163.9259
26.3899
110.5426
223.3801
213.13969
126.0244
133.5643
185.302
706.2041
812.7886
805.9647
872.1938
388.4285
437.7965
622.706
699.0134
717.4546
773.2855
698.2022
660.0107
704.6313
614.9319
644.2205
869.3716
726.1966
820.7325
673.419
965.4507
903.823
979.804
804.4591
989.7741
1929.285
1373.1104
1559.198
1987.6467
2423.2714
1983.009
2087.267
1979.4948
2173.1504
2292.5944
1777.4957
2712.9126
2843.6717
2783.2103
3367.8779
3065.6583
2821.1473
2724.849
2793.6015
3785.4644
3788.7085
3486.6386
4024.7494
1813.2615
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1488
1488
1488
1488
1488
1488
1488
1488
1488
1488
1488
1488
1488
1488
1488
1488
1488
1488
1488
1488
1488
1488
1488
1488
187
Fig. 10. Generated power by different RESs of the MG in stochastic case with neglecting the CHP effect.
Table 7
Result of stochastic problem considering CHP.
Time (h)
DG Source (kW)
WT
PV
FC
TH
Substation
Heat of FC-CHP
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
265.8027
651.2537
600.7695
119.6954
185.526
393.3251
174.3731
796.7623
479.4441
413.4723
800
782.5996
800
774.4829
489.2117
465.7514
672.0219
776.1898
800
441.9131
504.5326
800
465.2652
448.3933
0
0
0
0
0
0
0
264.66
396.99
264.66
317.592
660
528
660
396
484
230.6304
266.112
230.6304
0
0
0
0
0
235
235
126
111
235
2
126
2
235
111
2
235
2
126
116.7159
2
111
111
235
235
2
2
85.6404
235
1021
921.4
433.4
921.4
921.4
537.811
523
413.4
623.0372
772
523
921.4
889.2654
762
386.6184
672.4
862
1021
433.4
593
583.52615
1021
782
862
1507.4203
1219.6692
1879.8401
1897.0193
1697.9495
2120.453
2224.6563
2345.9424
2077.7702
2258.137
2168.5025
2371.9012
2762.6941
2648.9582
3647.2867
3398.8138
2739.171
2419.3692
2919.2302
4167.4399
4385.5277
3571.7038
4101.9749
1894.7997
236.692
236.692
236.692
236.692
236.692
236.692
236.692
236.692
236.692
236.692
236.692
236.692
236.692
236.692
236.692
236.692
236.692
236.692
236.692
236.692
236.692
236.692
236.692
236.692
1251.308
1251.308
1251.308
1251.308
1251.308
1251.308
1251.308
1251.308
1251.308
1251.308
1251.308
1251.308
1251.308
1251.308
1251.308
1251.308
1251.308
1251.308
1251.308
1251.308
1251.308
1251.308
1251.308
1251.308
188
Fig. 11. Generated power by different RESs of the MG in stochastic case with considering the CHP effect.
Table 8
Comparison of the results obtained by different algorithms considering CHP in case of probabilistic problem for 100 trials.
Objective function
Algorithm
Best solution
Mean solution
Worst solution
SD
Prot ($)
Proposed MPSO
PSO
GA
871.7179
857.8112
846.5346
871.5927
856.9606
843.1765
869.6314
851.7352
834.9548
0.4955
2.1083
5.2545
94
86
71
245.6
272.5
283.7
8. Conclusion
In this paper, the stochastic problem of coordinated scheduling
of CHP units in MG is solved with regard to WT and PV units using
MPSO algorithm. Maximizing the total prots for the participation
of an MG in the electrical market is the objective function considered in this study. The uncertainties of electrical market price, solar
radiation, and the speed of wind are considered using a scenariobased method which enhances the operational reliability. In the
189