Académique Documents
Professionnel Documents
Culture Documents
a r t i c l e
i n f o
Article history:
Received 7 November 2013
Received in revised form 6 December 2014
Accepted 11 January 2015
Available online 2 February 2015
Keywords:
Combinatorial optimization
Mixed binary linear programming
Network reconguration
Transmission network expansion planning
Repowering assets
a b s t r a c t
Transmission expansion planning (TEP) is a classic problem in electric power systems. In current optimization models used to approach the TEP problem, new transmission lines and two-winding transformers
are commonly used as the only candidate solutions. However, in practice, planners have resorted to nonconventional solutions such as network reconguration and/or repowering of existing network assets
(lines or transformers). These types of non-conventional solutions are currently not included in the classic mathematical models of the TEP problem. This paper presents the modeling of necessary equations,
using linear expressions, in order to include non-conventional candidate solutions in the disjunctive linear model of the TEP problem. The resulting model is a mixed integer linear programming problem,
which guarantees convergence to the optimal solution by means of available classical optimization tools.
The proposed model is implemented in the AMPL modeling language and is solved using CPLEX optimizer. The Garver test system, IEEE 24-busbar system, and a Colombian system are used to demonstrate
that the utilization of non-conventional candidate solutions can reduce investment costs of the TEP
problem.
2015 Elsevier Ltd. All rights reserved.
Introduction
Transmission expansion planning (TEP) is a classic problem in
electrical power systems; its goal is to nd the optimal plan for
expansion of lines and/or transformers to be installed in the network in order to allow a feasible operation in a pre-dened horizon
at a minimum cost. The optimal expansion plan should dene
where, how many, and when new network elements (lines or
transformers) must be installed. The necessary data for solving
the TEP problem include: current topology, candidate circuits, generation, demand for year-horizon, and investment constraints,
among others.
Corresponding author at: Calle 70 No. 52-21 Medelln, Colombia. Tel.: +57 (4)
2198557, +57 3008315893.
E-mail addresses: dtejada@xm.com.co (D. Tejada), jesusmarialopezl@yahoo.com
(J.M. Lpez-Lezama), mjrider@dee.feis.unesp.br (M.J. Rider), gevinasco@isa.com.co
(G. Vinasco).
1
Address: Calle 12 Sur No. 18-168 Bloque 2, Medelln, Colombia. Tel.: +57 (4)
3172929.
2
Address: Avenida Brasil, 56. Bairro: Centro 15385-000 Ilha Solteira, SP, Brazil.
Tel.: +55 (18) 3743 1000.
3
Address: Calle 12 Sur No. 18-168, Colombia. Tel.: +57 (4) 325 22 70.
http://dx.doi.org/10.1016/j.ijepes.2015.01.008
0142-0615/ 2015 Elsevier Ltd. All rights reserved.
214
Nomenclature
Constants
di
load in bus i
maximum generation in bus i
gi
cij
investment for the transmission in branch ij
f
maximum real power ow in branch ij
ij
n0ij
number of existing transmission lines in the branch ij
ij
maximum number of reinforcement that can be added
n
in the branch ij
xpu
reactance of transmission branch ij
ij
Sbase
base power
h
maximum phase angles in radians
fr
maximum real power ow in repowered branch ij
ij
xrij
reactance of transmission repowered branch ij
crepij
investment for the upgrade of transmission in branch ij
crij
investment for the transmission in the repowered
branch ij
Variables
wij,y
binary decision to add a new line in branch ij and circuit
y
gi
active power generation in bus i
hi
phase angle at bus i in radians
f ij
fij,y
wak ek ;1
wbk ek ;1
wak bk ;1
rij
wrij;y
0r
f ij
r
f ij;y
Sets
Xb
Xl
Xk
busbars set
branches set
set of candidate lines to be recongured
215
Minimize :
cij
ij 2 Xl
wij;y
y2Y
Subject to:
0
f ji
ji 2 Xl
X
y2Y
!
f ji;y
0
f ij
f ij;y
g i di
8i 2 Xb
y2Y
ij 2 Xl
0 hi hj
Sbase nij
8ij 2 Xl
xpu
ij
f xpu
ij;y ij
hi hj 6 2h1 wij;y
Sbase
0
f ij
jf ij j 6 n0ijf ij
8ij 2 Xl ;
8y 2 Y
8ij 2 Xl
jf ij;y j 6 wij;yf ij
8ij 2 Xl ;
4
5
8y 2 Y
8i 2 Xb
0 6 g i 6 gi
8i 2 Xb
jhi j 6 h
X
ij
wij;y 6 n
8ij 2 Xl
6
7
8
9
y2Y
10
11
hi 0
8i 2 Xb =i ref
12
wak ek ;1 wbk ek ;1
wak bk ;1 1 wak ek ;1
8k 2 Xk
8k 2 Xk
13
14
216
Table 1
Constraints for reconguration model.
Constraints
Description
wa3,1 = wb3,1
wab,1 = 1 wa3,1
0r
f ji f ji
X
f ji;y f ji;y
!
y2Y
ji 2 Xl
0r
f ij f ij
X
f ij;y f ij;y
y2Y
ij 2 Xl
8i 2 X b
g i di
19
r
f xr pu
ij;y ij
hi hj 6 2h 1 wrij;y
Sbase
r
8ij 2 Xl ; 8y 2 Y
f ij;y 6 wrij;y fr
ij
X
r
ij
wij;y 6 n
8ij 2 Xl
8ij 2 Xl ; 8y 2 Y
20
21
22
y2Y
wrij;y 6 wrij;y1
Fig. 3. Repowering of an existing line.
wrij;y 6 r ij
wij;y 6 1 rij
(f 012 > f 12 ). Also, the new lines impedance is adjusted because of
the cables change (x012 x12 ). Fig. 3 illustrates this situation.
To represent the repowering of an existing branch by using a
mathematical model, it is necessary to redene and add
constraints to the TEP model. First, the constraint described by
(3), KVL for existing lines, is redened as two new constraints
(15) and (16):
0 pu
f x
ij ij
h
h
15
16
0
f ij 6 n0ij f ij 1 r ij
0r
r
jf ij j 6 n0ij fr
ij ij
8ij 2 Xl
8ij 2 Xl
17
23
8ij 2 Xl ; 8y 2 Y
24
8ij 2 Xl ; 8y 2 Y
25
The constraint given by (20) is Ohms law for the new lines with
the same characteristics as the repowered lines. The constraint
given by (21) is the maximum power ow for the new lines with
the same characteristics as the repowered lines. Constraint (22)
limits the number of new lines, with the same characteristics as
the repowered lines that can be added to the branch ij. The constraint given by (23) avoids exploring the same solutions twice.
Constraint (24) Guaranties that new lines with the characteristics
of the repowered lines are considered only if such corridor was
effectively repowered (rij = 1). If the decision is no to repower
(rij = 0) new lines will not be built with the characteristic of repowered lines. Constraint (25) is the complement of (24). If a corridor is
repowered (rij = 1) it is not possible to have new lines with the
characteristics of the original. Conversely, if repowering is not performed (rij = 0) the new lines must have the same characteristic of
the originals.
Constraints (24) and (25) guarantee that new lines with the
same characteristics as the repowered lines are included only if
the original line between busbars i and j is repowered.
Finally, it is necessary to redene the objective function given
by (1) as the following equation:
Minimize :
X
ij 2 Xl
cij
X
y2Y
wij;y
n0ij crepij r ij
ij 2 Xl
X
ij 2 Xl
crij
wrij;y
y2Y
26
In this case the cost of repowering lines has been added to the
objective function.
18
217
Table 2
Garver system with reconguration and repowering test 1.
Busbar i
Busbar j
Existing branch
New circuits
Repowered circuits
Power ow (MW)
1
1
1
2
3
4
2
8
6
6
2
4
5
3
5
6
7
4
7
8
1
1
1
1
1
0
1
1
0
0
0
0
0
1
0
1
0
0
1
1
0
0
0
0
1
0
0
0
0
0
13.7
2.61
53.7
133.7
186.3
78.0
92.6
79.4
92.6
79.4
100
80
100
200
200
100
100
100
100
100
Table 3
Garver system with reconguration and repowering test 2.
Busbar i
Busbar j
Existing branch
New circuits
Repowered circuits
Power ow (MW)
1
1
1
2
3
4
2
7
8
2
4
5
3
5
6
7
8
4
1
1
1
1
1
0
1
0
1
0
0
0
0
1
3
0
1
0
0
0
0
0
0
0
0
0
0
40.0
40.0
66.7
100.0
173.3
300.0
100.0
100.0
100.0
100
80
100
200
200
100
100
100
100
218
To complement the previous test and validate the model, a second test is performed assuming that the cost of the reconguration
and repowering are 10 times the cost presented in Fig. 4. The
results of the second test are presented in Table 3. The total cost
of the investment is US$110 in the second test. When the reconguration cost is much higher than the cost of traditional candidates,
the optimal solution does not take into account the reconguration. For this test, the best known solution for the system was
found using only traditional candidates.
Table 4
Parameters of the IEEE 24 bus system test.
Parameter
Value
fr
ij
xrij
1.25 f ij
crepij
crij
0.5 cij
1 cij
0.95 xpu
ij
Table 5
Results of the IEEE 24 bus system.
Result
Only conventional
candidates
Including
non-conventional
candidates
164
150
Added elements
610
2 78
1012
1113
610
2 78
1113
Repowered elements
15
1012
Fig. 5. Results for the IEEE 24 bus system using only conventional candidates.
Fig. 6. Results for the IEEE 24 bus system including repowering of existing network.
219
SABAN
FUNDA
CUEST
TGUAJ
SMART
G
G
Generator
COPEY
Existing
Network
Candidates
TERMO
VALLE
Dummy
node
Repowering
candidates
Fig. 7. Alternatives for connecting Termocol generator.
Table 6
Results of the generators connection.
Busbar i
Busbar j
Existing branch
New circuits
Repowered circuits
Power ow (MW)
CUEST
CUEST
VALLE
COPEY
FUNDA
SABAN
SMART
10
SMART
11
12
TERMO
TERMO
TGUAJ
VALLE
COPEY
FUNDA
SMART
FUNDA
9
TGUAJ
11
12
TGUAJ
9
10
2
1
2
1
2
3
1
1
1
0
1
0
0
0
0
0
0
0
0
0
0
0
1
0
1
1
0
0
0
0
0
0
0
1
0
0
0
0
0
109.5
184.5
104.5
115.5
287.5
152.0
228.7
26.7
163.8
163.8
163.8
228.7
26.7
452
214
422
214
470
912
235
235
235
235
235
344
344
SABAN
FUNDA
TGUAJ
SMART
G
G
Generator
Existing
Network
Solution
COPEY
Dummy
node
Repowered
network
Fig. 8. Generator connection.
TERMO
VALLE
CUEST
220
Table 7
Summary of results.
Test case
221