This article has been accepted for inclusion in a future issue of this journal.

Content is final as presented, with the exception of pagination.

IEEE TRANSACTIONS ON POWER SYSTEMS 1

Optimal Power Flow With the Consideration of Flexible

Transmission Line Impedance
Tao Ding, Student Member, IEEE, Rui Bo, Senior Member, IEEE, Fangxing Li, Senior Member, IEEE, and
Hongbin Sun, Senior Member, IEEE

Abstract—This letter presents an optimal power flow model II. MATHEMATICAL FORMULATION
with the consideration of flexible transmission line impedance.
By the use of big-M based complementary constraints, the orig- The general DC optimal power flow can be mathematically
inal non-convex model can be transformed into a mixed integer expressed as a quadratic program with linear constraints and a
quadratic programming that can be solved by branch and bound quadratic cost function in the following form:
method. Numerical results from several test systems show that
flexible transmission lines can provide better economic dispatch (1a)
and reduce the total generation cost. Moreover, some infeasible
optimal power flow scenarios can become solvable with optimal
adjustment of the flexible transmission line impedance.
Index Terms—Big-M, complementary constraint, flexible trans-
mission line impedance, optimal power flow. (1b)

(1c)
I. INTRODUCTION
(1d)

O PTIMAL power flow (OPF) is widely used in economic

dispatch to minimize the total generation cost, while
guaranteeing the energy balance within generator physical where is the generation of the th generator and is the load
(1e)

limits and transmission line capacity limits. Traditionally, the demand at bus is the triplet coefficients of quadratic
transmission elements are treated as fixed assets in the network, cost function of the th generator; denotes the set of indices
except during times of forced outages or maintenance [1]. of generators connected to bus ; the subscript -ij denotes the th
However, with installing more flexible elements, such as line with its “from” bus and “to” bus ; the set of “from” buses
FACTs [2], power system operation becomes more flexible. and “to” buses are and is the voltage angle of the th bus and
“ref” refers to reference bus; is the reactance of the th trans-
The transmission line impedances can be changed by TCSC to
mission line; and are the minimum and maximum
increase transfer capacity nowadays. It is expected that, with
generation limit of the th generator; is the transmission ca-
the breakthroughs in the area of materials, sensors, and controls
pacity of the th line; and are the lower and upper bound
in the future, economically viable flexible devices such as
of voltage angle at bus and denote the total number
piezoresistive and thermistor impedance may be developed and
of buses, generators, and transmission lines, respectively.
installed in power systems on a large scale that can enable the
If the reactance of each line is taken as a variable, in the
adjustment of transmission line impedance. range of for , it yields a new model:
By reducing impedance, transfer capacity can be improved.
By increasing impedance on congested transmission lines,
(2a)
power will be shifted to other facilities to mitigate the conges-
tion. In this letter, we propose an optimal power flow model
with the consideration of flexible transmission line impedances (2b)
(FTLIs) to achieve better operational economics. Certainly,
possible power system protection problems should be taken Note that (2) is a non-convex model, because the term
into account in practical application [5]. becomes a bilinear function, instead of a linear
function, when is taken as a variable. Here, we assume
Manuscript received June 30, 2014; revised October 23, 2014; accepted the and are positive.
January 27, 2015. This work was supported in part by NSFC (51428701, Fortunately, model (2) has a special structure that the bi-
51321005), the 973 Program of China (2013CB228206), and US NSF linear terms always show up together and the
EEC-1041877. Paper no. PESL-00094-2014.
T. Ding is with the Department of Electrical Engineering, Tsinghua Univer- variables only appear in (2b). Therefore, we introduce
sity, Beijing, China, and also with The University of Tennessee, Knoxville, USA a new dummy vector for each line,
(e-mail: [email protected]). whose physical meaning is the power flowing on each line. After
R. Bo is with the Mid-Continent Independent Transmission System Operator obtaining the optimal solution of (2) with variables ,
(Midwest ISO), St Paul, MN 55108 USA.
F. Li is with the Department of Electrical Engineering and Computer Science the optimal reactance can be uniquely determined by
(EECS), The University of Tennessee, Knoxville, TN 37996 USA. . Therefore, replace the
H. Sun is with Department of Electrical Engineering, Tsinghua University, variable by and the constraint (2b) becomes
Beijing, China.
Digital Object Identifier 10.1109/TPWRS.2015.2412682 (3)

0885-8950 © 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

2 IEEE TRANSACTIONS ON POWER SYSTEMS

However, the sign of cannot be determined beforehand, TABLE I

so it is difficult to transform (3) into a linear constraint by multi- TRANSMISSION LINE LIMIT AND COMPUTATIONAL TIME
plying for both sides of the inequalities. Moreover, if the
denominator is zero, the numerator must be zero.
According to the different sign of , (3) can be simplified as

(4)

These “if” constraints can be simplified by the use of binary

TABLE II
variables and big-M complementary constraints [3], such that GEN COST ($) OF OPTIMAL POWER FLOW WITH/WITHOUT FTLIS

(5)
where is a big number and is a dummy binary variable.
1) When , we have
. Therefore,
whichever equals, 1 or 0, there must be one equation
that leads to , and the other one is a
redundant constraint that is always satisfied, such tion, where the second scenario (“S2”) has tighter transmission
that . line limits than the first scenario (“S1”). The computation time
2) When , the constraints is presented in Table I, where it needs more time for large sys-
tems. Note, if there are limited FTLIs in the system, the discrete
and are
variables in the model (6) will be greatly reduced, which will
complementary as only one constraint would be active, and
alleviate the computational time.
the other one is a redundant constraint, leading to
Besides, Table II shows the generation cost and it can be
.
observed that the cost with FTLIs is reduced comparing to that
As discussed above, (4) is equivalent to (3), so that the orig-
without FTLIs. As well, with the increase of (i.e., increase
inal model (2) can be transformed into a mixed integer quadratic
the available range of impedances), the total cost will be further
programming (MIQP), which can be easily solved by branch
reduced. Moreover, none of the traditional optimal power flow
and bound method or cutting plane method. But it has been re-
without FTLIs can be converged due to tight constraints for “S2”,
ported in [3] that should be sufficiently large but not overly
but with the help of FTLIs, they may become feasible. For the
big, because an extremely large number may cause numerical
first three systems, when %, the optimal power flow is
instability. Moreover, when solving MIQP, large will result
still infeasible, but when %, it become feasible. For the
in large feasible region of the relaxed model, so that more it-
last three systems, when %, the optimal power flow can
erations are needed to find the optimal solution. In our model,
be converged and the cost from % is better than %.
it can be found that the angle difference of line must be
in with respect to the power system stability re-
quirement, so is chosen to be . IV. CONCLUSION
Finally, the proposed economic dispatch model with flexible This letter proposes an optimal power flow model with the
transmission line impedances can be cast as follows: consideration of flexible transmission line impedance, which
can be formulated as a mixed integer quadratic programming.
(6a) The results and comparisons on six test systems show that
a more economic dispatch is achieved as transmission line
(6b) impedance becomes flexible. Most importantly, the flexible
transmission line impedance can help traditionally unsolvable
(6c) optimal power flow find an optimal solution.
(6d)
(6e) REFERENCES
[1] K. W. Hedman, R. P. O'Neill, E. B. Fisher, and S. S. Oren, “Optimal
(6f) transmission switching with contingency analysis,” IEEE Trans.
Power Syst., vol. 24, no. 3, pp. 1577–1586, Aug. 2009.
[2] S. N. Singh and A. K. David, “A new approach for placement of
III. NUMERICAL EXAMPLE FACTS devices in open power markets,” IEEE Power Eng. Rev., vol.
21, no. 9, pp. 58–60, 2001.
In this section, the proposed model has been studied on six [3] T. Ding, R. Bo, W. Gu, and H. Sun, “Big-M based MIQP method for
test systems available from MATPOWER [4] and implemented economic dispatch with disjoint prohibited zones,” IEEE Trans. Power
Syst., vol. 29, no. 2, pp. 976–977, Mar. 2014.
in MATLAB with the CPLEX 12.5. The range of impedance [4] R. D. Zimmerman, C. E. Murillo-Sánchez, and R. J. Thomas, “MAT-
is assumed to be and is chosen as 10% and POWER: Steady-state operations, planning, and analysis tools for
20%, respectively. Note that many transmission line limits are power systems research and education,” IEEE Trans. Power Syst., vol.
not available in [4]. In order to study the impact of FTLIs on op- 26, no. 1, pp. 12–19, Feb. 2011.
[5] P. K. Dash, A. K. Pradhan, G. Panda, and A. C. Liew, “Adaptive relay
timal power flow, uniform transmission line limits are assumed setting for flexible AC transmission systems (FACTS),” IEEE Trans.
in Table I. In addition, two scenarios are taken into considera- Power Del., vol. 15, no. 1, pp. 38–43, Jan. 2000.