See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/280841020

Optimal PMU Placement for Complete Observability Using Heuristic Methods

Conference Paper · December 2012

CITATIONS READS
3 450

3 authors, including:

Almoataz Y. Abdelaziz Riham Salem

Future University in Egypt Ain Shams University
429 PUBLICATIONS   2,687 CITATIONS    7 PUBLICATIONS   18 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Power System Control View project

Impact des D-FACTS sur les Performances des Systèmes de Protection dans les Réseaux Électriques MT View project

All content following this page was uploaded by Almoataz Y. Abdelaziz on 10 August 2015.

The user has requested enhancement of the downloaded file.

Proceedings of the 15th International Middle East Power Systems Conference (MEPCON’12), Alexandria University,
Egypt, December 23-25, 2012, Paper ID 127.

Optimal PMU Placement for Complete Observability

Using Heuristic Methods
Almoataz Y. Abdelaziz Amr M. Ibrahim Reham H. Salem
[email protected] [email protected] [email protected]
Department of Electrical Power & Machines, Faculty of Engineering, Ain Shams University, Cairo, Egypt

Abstract- An essential tool for power system monitoring is state estimation problem, the power system is said to be observable.
estimation. Phasor Measurement Units (PMUs) can greatly Because of the cost of the measurement devices, it is not
improve the state estimation process. However, for state economically feasible to place them at every network node.
estimation, the PMUs should be placed appropriately in the This explains why many measurement placement methods
network. The problem of optimal PMU placement for full
have been proposed in the literature.
observability is analyzed in this paper. The objective of the paper
is to minimize the size of the PMU configuration while allowing The complete observability of the power system, while using
full observability of the network. At first, an optimal phasor measurements, implies that each bus of the network
measurement set is determined to achieve full power system must have one voltage phasor measurement or a voltage
observability with ignoring zero injection buses. Then, the phasor pseudo-measurement. These phasor measurements are
derived schemes are modified to maintain the observability after obtained from the phasor measurement units (PMUs) directly
considering zero injection buses. Greedy algorithm and Single at the locations, where these have been installed. Then by
Vertex Algorithm are used as an optimization tool to obtain the applying Kirchhoff‟s and Ohm‟s laws, the remaining variables
minimal number of PMUs and their corresponding locations can easily be calculated as pseudo measurements [3].
while satisfying associated constraint. Applications of the
The problem of network observability has been studied by
proposed optimization algorithms are applied to the IEEE 14-bus
test system as a standard system and to two real systems various researchers in the past. Two different approaches used
customized from the England and Egyptian networks to for solving this problem [4] are based on numerical
demonstrate the effectiveness of the proposed approaches. observability and topological observability, which have their
own advantages and disadvantages. Numerical Observability
Index Terms — Phasor Measurement Unit (PMU); Observability; based approach utilizes the information (or gain) matrix or the
Greedy algorithm; Single Vertex Algorithm; Optimal PMU measurement Jacobian, when the measurement Jacobian is of
Placement (OPP). full rank, the network is said to be numerically observable.
I. INTRODUCTION Many OPP techniques, based on this concept, have been
devised. Simulated Annealing [5], Tabu Search [6], and
The Phasor Measurement Unit (PMU) is a power system Genetic Algorithm [7] have been used to find the optimal
device capable of measuring the synchronized voltage and PMU locations in the system. However, these methods are
current phasor in a power system. iterative in nature and involve extensive matrix manipulations
PMU is considered to be one of the most important and are, therefore, computationally extensive. On the other
measuring devices in the future of power systems. The hand, topological observability based approaches focus on the
distinction comes from its unique ability to provide placement of measurements to obtain an observable system
synchronized phasor measurements of voltages and currents utilizing the graph concept.
from widely dispersed locations in an electric power grid. The A few methods, based on this concept, Depth First Search [8],
commercialization of the global positioning satellite (GPS) Spanning Tree based method [9].
with accuracy of timing pulses in the order of 1 microsecond The PMU placement problem should be seen as a
made possible the commercial production of phasor combinatorial optimization problem. This can be solved using
measurement units [1, 2]. One of the applications of phasor Heuristic methods or metaheuristic methods. Unfortunately,
measurements in power systems is state estimation which is metaheuristics cannot prove optimality as in exact methods
performed in a control center to provide a platform for such as integer programming [19]. Most metaheuristics
monitoring, and security applications such as contingency algorithms are only approximation algorithms, because they
analysis and optimal power flow. The first step in state cannot always find the global optimal solution.
estimation is to gather measured data from different Difficulties faced by exact, conventional optimization methods
substations in a power network. These measurements must be often end up by preventing these methods to determine a
sufficient to make the system observable. solution to the optimization problem within a reasonable
A fundamental question is whether it is possible to amount of time. To avoid such cases, alternative methods have
determine all bus voltage angles and magnitudes of the studied been proposed, which are able to determine not perfectly
power system using the measurement set. When the accurate, but good quality approximations to exact solutions.
measurement set allows a unique solution of the state These methods, called heuristics, were initially based
essentially on experts‟ knowledge and experience and aimed bus voltage phasor and all current phasors along the lines
to explore the search space in a particularly convenient way connected to that bus will be available. This also implies that
[10]. the bus voltage, along with all adjacent bus voltages will also
Heuristics were first introduced by G. Polya in 1945 and be available. A solution i.e. xi a set of minimum is to be
were developed later in the 70‟s, when various heuristics were found out which will satisfy above equation 1. The constraint
also introduced for specific purpose problems in different vector function is formed using the binary connectivity matrix
fields of science and technique, including electrical A of power system [13].
engineering [11]. The entire procedure of building seven bus constraints in a
This paper introduces two optimization algorithms: Greedy PMU placement problem is explained with the help the seven
Algorithm and Single Vertex Algorithm. After introducing the bus system. The seven bus system is given in Fig. 1 [14].
characteristics of the two algorithms and purpose in detail, the
effect of zero injection buses is studied and the simulation
results of PMU placement are presented. The proposed
methods are tested on IEEE 14-bus test system, New England
(NE) 39- bus system and West Delta Network followed by a
conclusion.
II. OBSERVABILITY PROBLEM

Observability is defined as the ability to uniquely estimate

the states of a power system using given measurements. It is Fig 1: 7 Bus System
well-known that the state estimation cannot work even if
measurements are redundant. Observability analysis is A graph or network is defined by a set of nodes and vertices.
required to decide meter placement in order to maintain In order to determine the topology of the network, an
solvability of the observation equations in various conditions. adjacency matrix A can be defined so that Aij = 1 if there is a
Implementation of PMU presents an opportunity for vertex between nodes i and j. If nodes i and j are not
improving observability analysis and state estimation [6]. connected, Aij = 0 [15].
The problem of optimal PMU placement has been Initially a bus connectivity matrix A is formed with the help of
investigated earlier in [3, 4, 12]. These studies are based on the criterion given before:
the assumption that PMUs will be associated with buses where
they will monitor the bus voltage as well as currents along the
branches (lines and transformers) that are incident to this bus.
Results of such studies indicate that the entire system can be
made observable by placing PMUs at roughly 1/3 of the buses
in the system [12]. This number will decrease if zero injection
buses are taken into account during the placement process.
A. Problem Formulation
The constraint for this can be formed as,
The objective of the PMU placement problem is to accomplish
this task by using a minimum number of PMUs. For an n-bus
system, the PMU placement problem can be formulated as
follows:
n
min w *x
i 1
i i

Subject to…. f (x)  1 (1)

Where x is a binary decision variable vector and i is the bus
number, whose entries are defined as The operator „+‟serves as the logical ‟OR‟ and the use of 1 in
xi  1 if a PMU is installed at bus i, the right hand side of the inequality ensures that at least one of
xi  0 otherwise. the variables appearing in the sum is non-zero. For example,
wi is the cost of the PMU installed at the ith bus and f (x) is a consider the constraints associated with the first bus and the
vector function, whose entries are non- zero if the second bus. The first constraint f 1 implies that at least one
corresponding bus voltage is solvable using the given PMU must be placed at either one of buses 1 or 2 (or both) in
order to make bus 1 observable. Similarly, the second
measurement. 1 is a vector whose entries are all ones.
The expressions for the nonlinear constraints are formed based constraint f 2 indicates that at least one PMU should be
on the knowledge about the locations and types of existing installed at any one of the buses 1, 2, 3, 6, or 7 in order to
measurements. Given a PMU at a bus, it is assumed that the make bus 2 observable [16].
B. Modeling of Zero Injection Buses The constraints corresponding to all other buses will remain
Zero injection busses are the busses from which no current is the same as in (2). The zero injection bus constraint is
being injected into the system. If zero injection busses are also eliminated as it is taken care of by its neighbors [26].
modeled in the PMU placement problem, the total number of III. ALGORITHMS
PMUs can be further reduced [17]. To understand this issue,
consider the four bus example shown in Fig. 2. A. Greedy Algorithm
Fig. 3(a) depicts the system with injections in all the busses. Greedy algorithms are iterative methods used to find the
Fig. 3(b) shows a similar system with zero injection in bus 2 optimal solution of an optimization problem. Basically, this
and injections in bus 1, 3 and 4. For system (a), it can easily be algorithm makes decisions according to one rule: at each
seen that a minimum of two PMUs are required to make the stage, choose to install a PMU at the bus that covers the
system completely observable. These can be placed on any largest number of uncovered buses [19]. They are not as far-
two of the four busses. For example, if a PMU is placed on sighted as dynamic programming and other more sophisticated
bus 1, another PMU is required to make observable bus 4. In optimization algorithms. But this lack of sophistication makes
contrast consider system (b). For a PMU at bus 1, the current greedy algorithms particularly fast, easy to implement, and
in branch 2-4 becomes known as bus 2 is a zero injection bus. adaptable [20].
i.e. I24  I12 . Hence knowing the line parameters, the voltage Given a set of S elements, the greedy algorithm will choose
at bus 4 can be calculated 1 as: V4 = V2 - I12 Z24 . one element at a time based on a greedy choice property until
Hence a separate PMU is not required at bus 4 for (b). an end criterion is met. The elements in S could be locations,
Therefore it is seen that presence of zero injections can help in scheduled activities, paths, or items. Each element can only
reducing total number of PMUs required to observe the system be chosen once and is give a value for its greedy choice
[18]. property. At each stage, the element with the greatest value
will be chosen and removed from the set of candidate
elements. After each stage, the remaining element values will
be updated. This process continues choosing one candidate
element one at a time until the end criterion is met [21]. The
greedy algorithm flowchart is shown in Fig. 3.
B. Single Vertex Algorithm
This is one of the early heuristics developed by Teitz and
Bart (1968) for the p-median problem [22]. Similar to the
greedy algorithm, the single vertex (Random Node) selection
Fig. 2: Optimal PMU placement for a four-bus system (a) neglecting zero heuristic iteratively builds the solution until the visibility
injection constraints and (b) considering bus-2 as zero injection bus.
constraint is satisfied. The solution is built by taking input
Consider again the 7-bus system shown in Figure 1, where bus power system data and randomly selecting node. A PMU will
3 is assumed to be a zero injection bus. In this case, it is easy be placed on the selected node. The visibility matrix is
to see that if the phasor voltages at any three out of the four updated similar to the approach undertaken in the greedy
buses 2, 3, 4 and 6 are known, then the fourth one can be heuristic. Moreover, we also maintain the set of nodes with
calculated using the Kirchhoff‟s Current Law applied at bus 3 that are observed so that we can select a unique node. This
where the net injected current is known. Hence, the constraints random node selection process continues until the entire
associated with these buses will have to be modified power system is observable as shown in Fig. 4.
accordingly as shown below:
f 2  x1  x 2  x3  x6  x7  f 3.f 4.f 6  1
f 4  x3  x 4  x5  x 7  f 2.f 3.f 6  1
f 6  x 2  x3  x 6  f 2.f 3.f 4  1
The operator “.” in the above equations serves as the logical
“AND” Operation. The expressions for fi can be further
simplified by using the following properties of the Boolean
logical AND (.) and OR (+) operators. Given two sets A and
B, where set A is a subset of set B, then A + B = B and A⋅ B =
A. By substituting f 2, f 3, f 4 and f 6 in expressions of f 2 , f 4
and f 6 . The constraints can be written as:
f 2  x1  x 2  x3  x 6  x 7
f 4  x 2  x3  x 4  x5  x 6  x 7
f 6  x 2  x3  x 6  x1.x 4  x 4.x 7

Fig. 3: Greedy Algorithm

 Fig. 4: Single Vertex Algorithm

IV. SIMULATION RESULTS AND ANALYSIS

Fig. 6: IEEE 39- New England System.
The previous algorithms have been tested for the OPP
problem of IEEE 14 bus [29], New England 39 bus system
[30] and Real West Delta Network (WDN) [31]. The
simulation results are given in Tables 2 and 3 for the three
different cases.
A. IEEE 14 bus system
The IEEE 14-bus system has five synchronous machines,
three of which are synchronous condensers used for reactive
power support. Single line diagram of this test system is
shown in Fig. 5.
B. New England 39 bus system
The New England 39-bus system, having 10 generators, 19
loads, and 36 transmission lines, represents a reduced model
of the NE power system. Single line diagram of this test
system is shown in Fig. 6.
C. Real West Delta Network (WDN)
Single line diagram of real West Delta network (WDN) is
shown in Fig. 7.
Fig.7: West Delta Network

The data of the three test systems are shown in Table 1.

TABLE 1
BASIC DATA FOR THE TEST SYSTEMS

Test system No of No of ZI Location of ZI buses

branches buses
14-bus 20 1 7
39-bus 46 12 1, 2, 5, 6, 9, 10, 11, 13,
14, 17, 19, 22
WDN 55 0
52-BUS

Case 1: Without zero injection measurements

Optimum number of PMUs and their location using greedy
algorithm and Single Vertex Algorithm for the test systems are
shown in Table 2.
Fig. 5: IEEE 14-Bus System.
 TABLE 2
PMUS' PLACEMENTS FOR THE TEST SYSTEMS

Algorithm Test No of Location of PMUs

system PMUs
Greedy 14-bus 4 2, 6, 7, 9
Algorithm 39-bus 17 2, 4, 6, 8, 10, 12, 14, 16, 18, 19, 20,
22, 23, 25, 26, 29, 39
WDN 22 1, 2, 3, 4, 5, 6, 8, 14, 17, 18, 21, 28,
52-bus 29, 30, 35, 39, 42, 43, 45, 48, 49, 51
Single 14-bus 4 2, 6, 7, 9
Vertex 39-bus 17 2, 4, 6, 8, 10, 13, 14, 16, 18, 19, 20,
Algorithm 22, 23, 26, 29, 37, 39
WDN 18 1, 3, 4, 5, 6, 8, 11, 13, 17, 20, 21, 30,
52-bus 35, 38, 43, 44, 49, 51
Fig. 8: IEEE 14-bus single vertex algorithm.
In the first case, where no limitations were put on the number
of PMUs devices, the optimum number of devices is found to Comparison between results of different algorithms
be 4 in IEEE 14-bus and 17 in New England 39 bus system.
For the real Egyptian network, the optimum number of PMUs A comparison between the number of the required PMUs in
is found to be 22 in the first technique and 18 in the second the paper approaches and the other approaches used in optimal
technique. PMU placement is performed in Table 4.
TABLE 4
Case 2: With zero injection measurements
COMPARISON OF TOTAL NUMBER OF PMUS OBTAINED BY
Greedy Algorithm and Single Vertex Algorithm have been SEVERAL METHODS
tested for IEEE 14 bus and New England 39 bus systems. The TEST SYSTEM
simulation results are given in Table 3. ALGORITHM 14-BUS 39-BUS
Greedy 3 8
TABLE 3 Single Vertex 3 8
PMUS' PLACEMENTS FOR THE TEST SYSTEMS CONSIDERING ZERO Tabu search [23] 3 10
INJECTION EFFECT Integer linear 3 N/A
programming [17]
Graph theoretic [24] 5 N/A
Algorithm Test No of Location of PMUs
system PMUs Iterated local search [25] 4 N/A
Greedy 14-bus 3 2, 6 , 9 Simulated Annealing [26] 4 9
Algorithm 39-bus 8 3, 12, 16, 20, 23, 26, 29, 37 BPSO [27] 3 N/A
Single Vertex 14-bus 3 2, 6 , 9 Depth First Search [28] 6 16
39-bus 8 7, 12, 16, 20, 26, 29, 36, 37 N/A: not available

Greedy method and Single Vertex method were used to

In IEEE 14-bus system, bus 7 has no direct generator or load modify the PMU configuration in order to minimize the size
connection therefore it is classified as a zero-injection bus and of the configuration, as shown in table 4. The table indicates
bus 8 will be treated as directly connected to bus 4 and bus 9 that the algorithms can eliminate number of PMUs while
when applying the algorithm. modifying the configuration at random.
The IEEE 14-bus system only has one zero injection bus. Table (4) shows that
There are 4 PMUs in initial placement scheme and the optimal - No specified technique could be considered as absolute
scheme needs 3 PMUs to assure the observability conditions. optimal technique for all cases.
The New England 39 bus system has 12 zero-injection buses. - Less number of PMUs will be required if the zero-injections
Initially, the number of PMUs is 17, after optimization, the buses are considered.
PMU number reduces to 8. The Real West Delta Network -The Depth First Search gives the highest PMUs number
(WDN) has no zero injection bus. because it does not consider the zero injections and further its
The number of PMUs versus number of trials for IEEE 14 bus solution is not optimal as in [28].
is shown in Fig. 8 to demonstrate the random way of the
algorithm. Fig. 8 shows how the algorithm goes about V. CONCLUSION
minimizing the number of PMUs in the IEEE networks or the In this paper, two approaches for the optimal placement of
evolution of the minimum length as the algorithm iterates. The phasor measurement units are proposed. They were used as an
algorithm thus tries a chain of configurations which do not optimization tool for finding the minimal number of required
change much at all iterations. At all iterations, it also checks if PMUs for the complete power system observability. The main
the size of the PMU configuration cannot be reduced. contribution of the paper lies in benchmarking the global
optimal PMU placement solutions for widely used test
systems, and also in formulating a strategy to select the best
among competing solutions. Zero injection constraints have a
capability to further reduce PMU requirement. Simulation [15] J. S. Bhonsle & A. S. Junghare, “A Novel Approach for the Optimal
results were performed on several IEEE test systems and one PMU Placement using Binary Integer Programming Technique”, International
Journal of Electrical and Electronics Engineering (IJEEE), Vol. 1, Issue 3,
real system customized from Egyptian network. For the IEEE 2012, pp. 67-72.
standard test systems, the minimum number of PMUs was [16] Ganga Reddy Tankasala, Sridhar Sanisetty, Varun Kumar Vala, “Optimal
compared with those of existing methods. The salient features Placement of Phasor Measurement Units for State Estimation using Artificial
Intelligence Techniques”, International Journal of Scientific & Engineering
of the proposed methods include a low execution time as well
Research, Vol. 3, Issue 2, February 2012, pp. 1-5.
as global optimality that make the methods easy to implement. [17] Devesh Dua, Sanjay Dambhare and Rajeev Kumar Gajbhiye, "Optimal
Multistage Scheduling of PMU Placement: An ILP Approach", IEEE
REFERENCES Transactions on Power Delivery, Vol. 23, No. 4, October 2008, pp. 1812-
[1] A. P. Sakis Meliopoulos and George J. Cokkinides, “GPS Synchronized 1820.
Data Acquisition: Technology Assessment and Research Issues”, Proceedings [18] Sanjay Dambhare, Devesh Dua, Rajeev Kumar Gajbhiye and S. A.
of the 39th Hawaii International Conference on System Sciences, 2006, pp.1- Soman, "Optimal Zero Injection Considerations in PMU Placement: An ILP
9. Approach", 16th PSCC, Glasgow, Scotland, July 14-18, 2008, pp. 1-6.
[2] A. G. Phadke and J. S. Thorp, “History and Applications of Phasor [19] Ming Zhou, Virgilio A. Centeno, Arun G. Phadke, Yi Hu, Damir Novosel
Measurements”, IEEE PES Power Systems Conference and Exposition, 2006, and Hector A. R. Volskis, A Preprocessing Method for Effective PMU
pp. 331-335. Placement Studies, DRPT, Nanjing China, 6-9 April 2008,pp.1-6.
[3] Ranjana Sodhi and S.C.Srivastava, Optimal PMU Placement to Ensure [20] Walter Lassonde & Nathan Jensen Advisor: Professor Samee Khan
Observability of Power System, Fifteenth National Power Systems Group SD0903 , Phasor Measurement Unit Placements for Complete
Conference (NPSC), IIT Bombay, 16-18 December, 2008, pp. 1-6. Observability using Linear-Time, Quadratic-Time, and Subquadratic-Time
[4] B. Mohammadi Ivatloo, “Optimal Placement of PMUs for Power System Heuristics, NDSU ECE wiki, 2010.
Observability Using Topology Based Formulated Algorithms”, Journal of [21] Innocent Kamwa and Robert Grondin, PMU Configuration for System
Applied Sciences, Vol. 9, No. 13, 2009, pp. 2463- 2468. Dynamic Performance Measurement in Large Multiarea Power Systems,
[5] Alessandra B. Antonio and Jose R. A. TorreBo, “Meter Placement for IEEE Transactions on Power Systems, Vol. 17, No. 2, May 2002.
Power System State Estimation Using Simulated Annealing”, Proc. of the [22] Michael Dzator and Janet A. Dzator, Locating Emergency Facilities:
IEEE Porto Power Tech Conference, 10-13 Sep. 2001. Targeting Efficiency and Cost-Effectiveness, Western Australian Centre of
[6] Jiangnan Peng, Yuanzhang Sun, and H. F. Wang, “Optimal PMU Excellence in Industrial Optimization, University of Newcastle,2004 ,pp.1-20.
Placement for Full Network Observability using Tabu Search Algorithm”, [23] Jiangnan Peng, Yuanzhang Sun and H.F. Wang, Optimal PMU
Electrical Power and Energy Systems, Vol. 28, May 2006, pp. 223-231. placement for full network observability using Tabu search algorithm,
[7] B. Milosevic, M. Begovic, “Nondominated Sorting Genetic Algorithm for Electrical Power and Energy Systems Vol. 28, 2006, pp. 223–231.
Optimal Phasor Measurement Placement”, IEEE Transactions on Power [24] Morteza farsadi, Hamid golahmadi and Hadi shojaei, Phasor
Systems, Vol.18, No. 1, Feb. 2003, pp. 69–75. Measurement Unit (PMU) Allocation in Power System with different
[8] Reynaldo F. Nuqui and Arun G. Phadke, “Phasor Measurement Unit Algorithms, Eletrical and electronics Engineering International Conference
Placement Techniques for Complete and incomplete Observability”, IEEE ELECO, 2009,pp.396-400.
Transactions on Power Delivery, Vol. 20, No. 4, Oct. 2005, pp. 2381- 2388. [25] M. Hurtgen and J.-C.Maun, Optimal PMU placement using Iterated
[9] G. B Denegri, M. Invernizzi and F. Milano, “A Security Oriented Local Search, Electrical Power and Energy Systems, 32 (2010) 857–860.
Approach to PMU Positioning for Advanced Monitoring of a Transmission [26] Thawatch Kerdchuen and Weerakorn Ongsakul, Optimal PMU
Grid”, International Conference Power System Technology, 2002, pp. 798- Placement for Reliable Power System State Estimation, GMSARN
803. International Conference on Sustainable Development: Challenges and
[10] Mihai Garvilas, and Gheorghe Asachi, Heuristic and Metaheuristic Opportunities for GMS 12-14 Dec. 2007, pp.1-5.
Optimization Techniques with Application to Power Systems, Selected Topics [27] Charu Sharma1, Barjeev Tyagi, An approach for Optimal PMU
in Mathematical Methods and Computational Techniques in Electrical placement using Binary Particle Swarm Optimization with Conventional
Engineering,Romania, 2010,pp.95-103. Measurements, International Journal of Engineering, Science and Technology
[11] F. Box, A Heuristic Technique for Assigning Frequencies to Mobile Vol. 3, No. 3, 2011, pp. 56-63.
Radio nets, IEEE Transactions on Vehicular Technology, 27, 57–74, 1978. [28] Almoataz Y. Abdelaziz, Amr M. Ibrahim, and Reham H. Salem, Optimal
[12] Roozbeh Emami and Ali Abur, "Optimal Placement of Phasor PMU Placement for Power System Observability, International Journal on
Measurements for Enhanced State Estimation: a case study", 16th PSCC, Power System Optimization, January-June 2011, Volume 3, No. 1, pp. 21– 25.
Glasgow, Scotland, July 14-18, 2008, pp.1-6 [29] Power Systems Test Case Archive. Available [Online] at:
[13] Rahul H. Shewale , Bala krishna Kethineni, Uma P Balaraju, S. K. Bhil http://www.ee.washington.edu/research/pstca/
and Prashant D More, “Optimal Placement of Phasor Measurement Unit for [30] M. A. Pai, Energy function analysis for power system stability, Kluwer
Power System Observability by Heuristic Search Method”, International Academic Publishers; 1989.
Journal of Advanced Technology & Engineering Research (IJATER), Vol. 2, [31] A. A. Abou El-Ela, S. M. Allam and M. M. Shatla “Maximal Optimal
Issue 2, May 2012, pp. 128-133. Benefits of Distributed Generation using Genetic Algorithms,” Electric Power
[14] R. Sudha, Vishwas Vats, Gaurav Pathak, T. Jayabarathi, “Optimal Systems Research, Vol. 80, Issue 7, 2010, pp. 869–877.
Placement of Phasor Measurement Units Using Modified Invasive Weed
Optimization”, International Journal of Soft Computing and Engineering
(IJSCE), Vol. 2, Issue 2, May 2012, pp. 274-277.

View publication stats