Technol Econ Smart Grids Sustain Energy (2 0 1 6 ) 1 :1 0

DOI 10.1007/s40866-016-0009-y

REVIEW PAPER

Artificial Intelligence Versus Conventional Mathematical

Techniques: A Review for Optimal Placement of Phasor
Measurement Units
Kinfe Negash1 · Baseem Khan1 · Estifanos Yohannes1

Received: 11 May 2016 / Accepted: 21 July 2016

© Springer Science+Business Media Singapore 2016

Abstract The Phasor Measurement Unit (PMU) is a con- the various methods is also presented based on the various
temporary metering device installed on system to enable operation and cost related factors.
the power system monitoring and control. PMUs are most
accurate and advance time synchronized technology which Keywords Phasor measurement unit · Conventional
provides measurements of voltages at the buses and also mathematical techniques · Artificial intelligence techniques
current phase values which are connected to those buses
where these PMUs are located. PMU placement on each
bus of a system is difficult to achieve either due to cost Introduction
factor or due to non-existence limitation of communication
facilities in some parts of the system. Moreover, as a con- The Modern deregulated electricity market is highly depen-
sequence of Ohm’s Law, when a PMU is placed at a bus, dent on the real time information. A reliable, accurate,
neighboring busses also become observable. This implies timely and complete set of measurement serves as a foun-
that a system can be made observable with a lesser number dation for the secure real time grid operation. For this Mon-
of PMUs than the number of busses. It is necessary to deter- itoring and control of these grids in order to ensure reliable
mine the minimum number of PMUs for entire observability power deliveries are dependent on sophisticated information
of the power network. The optimal placement of PMUs and control systems commonly referred to as SCADA/EMS
(OPP) problem can be solved by various techniques. Hence systems, where system states estimation depends on unsyn-
in this paper a comprehensive review is presented on the chronized measurement. These measurements have errors
OPP problems under the characterization of heuristic and such as measurement and telemetry bias. To overcome these
mathematical methodologies. Also a comparison between limitations in the SCADA, Wide Area Monitoring Protec-
tion and Control (WAMPAC) system is used. This system
consist Phasor Measurement Units (PMUs) as fundamental
components which gives synchronized and real-time volt-
ages and currents phasor measurement. Global Positioning
System (GPS) provides reference timing signals to achieve
synchronization of sampling voltage and current waveform
 Baseem Khan
baseem.khan04@gmail.com with respect to this reference time.
A PMU directly measure the voltage Phase of the bus
Kinfe Negash
kinfemichael.negash@gmail.com where these PMUs are placed and also measure the current
phases of a few or all the branches connected to that bus.
Estifanos Yohannes
estif83@gmail.com
In recent years rapidly increases use of PMUs to improve
monitoring, needed to place these PMUs on all of the buses
1 Institute of Technology, Hawassa University, Hawassa, for full observability of the network. It is also impossible to
Ethopia place these units on entire system buses because PMUs and
1 0 Page 2 of 13 Technol Econ Smart Grids Sustain Energy (2 0 1 6 ) 1 :1 0

communication services are very costly. Thus determination

of the optimal number of PMUs and its location for overall
observability of the system is very important.
A proper methodology is required to find the optimum
number of the PMUs which will fully observe the power
network. To solve the Optimal PMUs Placement (OPP)
problem a number of mathematical methods have been
employed. Among these methods linear Programming, non-
linear programming, dynamic programming are the com-
mon optimization methodologies are proposed to solve
this problem. Problems such as difficulties of obtaining
local minima and handling constraints in conventional tech-
niques are overcome by other optimization methodology
like heuristic techniques, such as depth first search, iterated
local search, simulated annealing, tabu search, genetic algo- Fig. 1 Layout of PMU along with GPS time stamped signals [26]
rithms, differential evaluation, imperialistic, particle swarm
optimization and ant colony optimization etc. aim is to completely observe the network with an optimum
This paper reviews the research work and studies that number of PMUs. The problem for n-bus system is formu-
have been done in the area of optimal placement of pha- lated and solved by integer programming method [51] as
sor measurement units (PMUs) that are presented to solve given below:
the typical optimal placement problem of PMUs. The for-

n
mulation of this problem is described in Section “Optimal min Wi x i
PMU Placement (OPP) Problem Formulation”. The heuris- i
tic methods to solve the OPP problem are discussed in
Subject to f (x) ≥ 1̂
Section “Artificial Intelligence Techniques”, the mathemat-
ical method is discussed in Section “Mathimatical Methods” Where x = binary decision variable vector
and Section “Comparative Analysis of Mathematical and i = bus number, whose entries are,
Artificail Intellegence Techniques” concludes this paper. xi = 1 if a PMU is located at bus i,
xi = 0 otherwise.

Optimal PMU Placement (OPP) Problem Wi = the cost of the PMU installed at the i th bus,
Formulation f (x) = vector function, having non-zero entry if
corresponding bus voltage is observable and
PMU is an advanced measurement device installed at a bus 1̂ = vector having all ones entries.
(or node) of the system, measuring the voltage and cur-
rent phasors of the line (or branch) it is associated with. The nonlinear constraint expressions are created consid-
Fig. 1 shows PMUs purely isolated to form a Wide Area ering the placement and types of available measurements.
Monitoring System (WAMS). GPS time stamped measure- Assume the phasor value of voltage at the bus where PMU
ment signals are fed to a Phasor Data Concentrator (PDC) located and values of current phasors along the branches
by using PMUs. The PDC collects and sorts the phasor which connected to that bus will be easily accessible. The
measurements and signal processor converts data of PMUs other adjacent bus voltages will also be accessible. Deter-
into useful information which is visible on Human Machine mine the solution vector which is a set of minimum xi
Interface (HMI).The operator can easily access the criti- and satisfy above equation. The constraint function can be
cal information of the power system state. Some rules can defined with the help of Binary Connectivity Matrix A
be used for the placement of PMUs which are given in which gives the information about bus connectivity of power
[26] like, assigning one voltage measurement at the bus network. The elements of matrix A is defined as,
where PMU is located, one branch current measurement, A(mn) = 1 if busm = busn
one voltage and current pseudo measurement. = 1 if there is connection between bus m and busn.
The PMUs can be placed at planned buses to completely
observe the total network. These located PMUs are measur- = 0 Otherwise.
ing the voltage phase value of that bus and current phase The constraint equations are considered for the three cases:
values of the lines which are connected to the same bus. The (1) PMU measurements only, (2) PMU measurements and
Technol Econ Smart Grids Sustain Energy (2 0 1 6 ) 1 :1 0 Page 3 of 131 0

injections (i.e. zero injections) and (3) PMU measurements,

injections and flows.

Artificial Intelligence Techniques

Imperialistic Competition Algorithm (ICA)

ICA is an evolutionary optimization algorithm which is

inspired by socio-political process of imperialistic com-
petition. This algorithm has proven its superior ability to
achieve faster convergence and better solution compared to
the other evolutionary algorithms.
In the proposed method in [1], the optimal placement
of PMUs and minimization of the cost of associated CI Fig. 2 the food finding iterative process of a fruit fly swarm [4]
(communication infrastructure) are carried out simultane-
ously by using ICA. That is to say the optimal placement
of PMUs, the optimal location of the PDC and minimal The Fruit Fly Optimization Algorithm (FOA)
communication paths between PMUs and the PDC are
obtained simultaneously in order to make the network as The fruit fly optimization algorithm (FOA) is a new method
fully observable as possible in terms of the state estimation. for finding global optimization based on the food find-
ing behavior of the fruit fly. The fruit fly itself is superior
Cellular Learning Automata (CLA) to other species in sensing and perception, especially in
osphresis and vision. The osphresis organs of fruit flies can
Cellular learning automata (CLA) algorithm is based on the find all kinds of scents floating in the air; it can even smell
usage of learning automata (LA) to the state transition prob- food source from 40 km away. Then, after it gets close to
ability adjustment of cellular automata (CA). This method the food location, it can also use its sensitive vision to find
starts by specifying the internal state of every cell. Then, the food and the company’s flocking location, and fly towards
reinforcement signal for each LA is determined on the basis that direction too. Fruit flies are a kind of insect, which live
of the rule of CLA. Finally, as for supplied reinforcement in the temperate and tropical climate zones and eat rotten
signal and action chosen by the cell, each LA updates its fruit [4]. Fig. 2 shows the food finding iterative process of a
action probability vector. The desired state will be reached fruit fly swarm.
by continuing this process [85]. In [2] proposed a CLA The authors in [4] develop a modified fruit fly method for
algorithm to tackle the optimization problem, introducing solving PMU placement problems. This method is found to
new CLA local rules to enhance the optimization process in speed up the convergence.
the formulation of a modified multi-objective PMU place-
ment method by first calculating worth of observability for Genetic Algorithm (GA)
a set of PMUs in contingency, by ranking the prospective
contingencies. Genetic algorithm (GA) is adaptive heuristic search algo-
rithm that repeats the process of natural evolution. This
Chemical Reaction Optimization (CRO) process is used to generate solutions to optimization and
also search problems. Modeling natural selection is the base
Chemical reaction optimization (CRO) is a heuristic algo- of genetic algorithm (GA) which does not need any sec-
rithm based on population processed by obtaining a lower ondary functions such as derivatives computation. Some
energy stable state by simulating the action and reaction positive characteristics of GA which make it more usable in
of molecules in a chemical reaction, which aimed to reach optimization problems are as follows:(a) probability of local
the minimum state of free energy. Authors in [3] proposed minimum trapping is decreased,(b) computations of going
chemical reaction optimization (CRO) method to solve the from one state to another is declined, and (c) evaluation of
problem of maximizing the power system observability by the fitness of each string guides the search.
placing phasor measurement units (PMUs) in a multistage A genetic algorithm is proposed in [5] in order to min-
manner where the problem is constrained by the financial imize the number of PMU placement sites. The proposed
budgets available at each installation stage. method is numerical observability by using the minimum
1 0 Page 4 of 13 Technol Econ Smart Grids Sustain Energy (2 0 1 6 ) 1 :1 0

condition number of the normalized measurement matrix as [16] presented a new optimization algorithm for optimal
criteria. The sequential elimination is used to find the essen- PMU configuration based on combination of graph theory
tial measurements for the completely determined condition. and genetic algorithm. The method for PMU allocation has
The sequential addition is used to select the redundancy shown a decent result, obtaining the same or smaller num-
measurements under the contingency. ber of PMUs to fully observe the power system compared
Reference [6] investigates the application of immunity with the references. What is more important is that the com-
genetic algorithm (IGA) for the problem of optimal place- putation cost is greatly reduced, which mainly attributes to
ment of phasor measurement units (PMUs) in an electric the topology constraint analysis.
power network. The problem is to determine the placement
sites of the minimal set of PMUs which makes the system Greedy Algorithm
observable.
The authors in [7] presents a non-dominated sorting The greedy algorithm is an optimization methodology that
genetic-based algorithm (NSGA), which can successfully follows the problem solving heuristic of making the locally
solve PMU placement problem with two competing objec- optimal choice at each stage with the hope of finding the
tives, such as minimization of the number of PMUs and global optimum.
maximization of the measurement redundancy. The opti- In [17], Optimal PMU placement problem was consid-
mization is performed without any preference information ered in the power system to improve the state estimation
given with respect to the objectives. results. The PMU placement problem has been formulated
In [8] genetic algorithms to find the optimal placement as an optimal experiment design problem, with a class of
of PMUs has been generated using a MATLAB program well-known optimality criteria. It has been showed that
.The genetic algorithm method was tested on three network- the design criteria are usually related to a key property,
tests, and the results were compared with other algorithms sub modularity, which often allows efficient greedy algo-
to prove that an optimal placement of PMUs ensures a rithms. A greedy PMU placement algorithm that has been
total observability of the network and thus confirming the showed achieving an approximation ratio of(1 − 1e ) for
efficiency of genetic algorithm. any PMU budget K, which is the best guarantee among
The genetic algorithm in [9] was used to calculate the polynomial-time algorithms has been presented in [18]. In
best PMU location that meets the prespecified threshold [19] a greedy algorithm is proposed to find unique optimal
of the LODE index for system observability reliability. In PMU placement with the best state estimation performance
this work multiple solutions for optimal PMU placement in three-phase network. A demonstration that the greedy
in the normal system state are generated and the solutions algorithm is efficient and taking PMU phase mismatch into
are ranked using the observability reliability indices. Then, account will help in the improvement of the PMU placement
the best solution is compared with the optimal PMU place- was presented in [20].
ment solution considering the full observability under N-1
conditions of transmission branch failures. In [10] authors Differential Evolution (DE)
presents a genetic algorithm based method for optimal meter
placement problem by which measurement system can be Differential evolution (DE) concept employs N-dimensional
optimally determined and upgraded to maintain network element vectors to minimize ongoing space functions.
observability. The proposed algorithm yields a measure- Mutation, crossover, and selection are the principle oper-
ment configuration that withstands any single branch outage ators utilized to carry out the global optimization. This
and/or loss of single measurement, without losing network heuristic method could be widely used in different cost
observability. In [11, 12] genetic algorithm (GA) is pro- function problems such as non-differentiable, non-linear
posed for PMU and SCADA placement for state estimation and, multi-modal functions. Parallel computations, easy
identification. The results obtained in [11] are indicative of usage, and good convergence properties are other benefits of
the fact that the optimal placement of PMU increases the this approach [85]. In [21] authors presents a multi-objective
accuracy of the obtained estimates and efficiency of the optimal model of PMU placement with the objectives of
bad data detection algorithms. In [13] Genetic algorithm minimization of the number of PMU and maximization of
was developed to solve the proposed new model that uti- the N-1 measurement reliability by using DE algorithm.
lizes network observability rules and determine the optimal
investment decision for the placement of PMUs in the power Particle Swarm Optimization (PSO)
grid. Reference [14] proposes an optimal PMU allocation
method for substations based on a specialized Genetic Algo- Particle swarm optimization (PSO) is a population-based
rithm (GA) while in [15] the same method has been used stochastic optimization technique in which each poten-
for complete and incomplete observability. The authors in tial solution (called a particle) is assigned a randomized
Technol Econ Smart Grids Sustain Energy (2 0 1 6 ) 1 :1 0 Page 5 of 131 0

velocity and then flows through the problem hyperspace. promising nodes are identified as those having least/highest
PSO has been found to be extremely effective in solving a score of an evaluation function. While searching towards the
wide range of engineering optimization problems [22, 83]. least optimal cost path, the BFS algorithm has the capabil-
A modified discrete binary version of particle swarm ity to change its search path from the current search path to
algorithm was used in [21, 25] as an optimization tool in the most promising path (Kumar, 2008). This characteristic
finding the minimal number of the required PMUs for the makes the BFS method superior over other graph theoretical
complete system observability. At first, the PMU placement methods for optimal PMU placement problem.
was solved with the goal of minimizing the total number of The authors in [33] utilize this method for the optimal
required PMUs for the complete system observability. Then, placement of PMUs to make the power system topologically
the effect of PMU loss or a branch outage was taken into observable. The proposed method determines the optimal
consideration, and a placement scheme was obtained which locations of PMUs in two stages. Stage I determines the sub-
maintains complete sys-tem observability during the con- optimal locations of PMUs using best first search, which
tingency conditions. In [23, 24] PSO optimization is also is an intelligent search algorithm with an ability to change
used for obsevability with the objective of minimum cost of its search path from the current node to the most promis-
PMU. ing node. From the results of stage I, the redundant PMU
locations are identified and eliminated in stage II using
Iterated Local Search (ILS) pruning. The proposed method is extended to take into
account the presence of conventional flow measurements in
Searching a smaller subspace, which is defined as local the system. It is also found to be effective in handling both
optima, instead of the whole solution space, is the main single as well as multiple flow measurements connected
viewpoint of iterated local search (ILS) method. By utilizing to a bus.
an embedded heuristic, a sequence of solutions is provided
in which the best solution is obtained if one were to utilize Simulated Annealing (SA)
repeated random trials of that heuristic. The Iterated Local
Search (ILS) metaheuristic in [26, 28, 31] is used to mini- Simulated annealing (SA) is a procedure for solving com-
mize the size of the PMU configuration needed to observe plicated combinatorial optimization in which the current
the network .The algorithm is tested on IEEE test networks solution is randomly altered. The new solution is the worse
with 14, 57 and 118 nodes and compared to the results alteration with the probability that is reduced as the compu-
obtained in previous publications. tation proceeds. An optimal solution for a large combina-
In [29] authors use a local search heuristic method to torial optimization problem needs a fit perturbation mecha-
derive the minimum number of PMUs. In addition to the nism, cost function, solution space, and cooling schedule to
practical model, this study also considers the ideal model, be solved by SA. Sufficiency of SA can be found by search-
in which all load nodes are assumed to be zero injection. ing a large-scale system and obtaining good speed in terms
The numerical studies on various IEEE power test systems of finding an optimal or near-optimal solution [84].
demonstrated the superior performance of the proposed In [34] an algorithm dealing with communication con-
algorithm in both the models in regard to computational strained PMU placement is modeled within the framework
time and solution quality. In [30] the empirical observabil- of simulated annealing. Here the pragmatic constraint of not
ity gramian is applied to quantify the degree of observability placing PMUs at buses without communication support is
of a power system under a specific PMU configuration and recognized.
an optimal PMU placement method for dynamic state esti- The proposed method in [35] considers project practical
mation is proposed by maximizing the determinant of the view. It used the concept of indirect observation, the MST
empirical observability gramian. It is effectively and effi- and SA method. And it also improved the state estimation
ciently solved by the NOMAD solver which implements the with SCADA/EMS measurement and PMU measurement.
Mesh Adaptive Direct Search algorithm. The method has been declared as effective, and it also have
some advanced points such as strong maneuverability and
Intelligent Search Technique small investment.

An intelligent-search technique called best first search Tabu Search (TS)

(BFS) method utilizes information about each node by com-
puting an evaluation function at each of them while moving Tabu search (TS) is an adaptive algorithm that utilizes many
towards the goal state. It helps the BFS algorithm to search other methods such as linear programming algorithms and
and explore only the most promising nodes, thereby mak- heuristic concepts. This procedure is presented to solve
ing the search process, simpler and faster one. The most the combinational optimization problems in scheduling and
1 0 Page 6 of 13 Technol Econ Smart Grids Sustain Energy (2 0 1 6 ) 1 :1 0

covering. Tabu list which is one of the main elements of Integer Programming (IP)
TS consists of the number of recently visited states plus
a number of unwanted states. Other main elements of TS Integer programming (IP) is a mathematical programming
are aspiration, diversification, and definition of a state and method of solving an optimization problem having integer
the surrounding area. There is a reset in TS when it is not design variables, while the objective function and the con-
converging [84]. straints are linear, nonlinear, or quadratic, thus leading to
A TS method to obtain a solution for the OPP problem integer linear programming (ILP), integer nonlinear pro-
by checking network observability with a numerical algo- gramming (INLP) and integer quadratic programming (IQP)
rithm is introduced in [36, 37]. The impact of three different algorithms, respectively[49, 79, 84].
TS initialization schemes, together with the two TS param- In [40] an IP was used to achieve optimal PMU place-
eters (TL and TSI), on the optimal solution is investigated. ment for observability in a manner that can also detect and
The experimental results verify the superiority of the pro- eliminate bad data in critical and critical measurement pairs,
posed RTS method over the MTS method. The effectiveness which otherwise can render the state estimation problem
and flexibility of the proposed scheme is demonstrated by unsolvable or bad data undetectable. In [32], a new method
the simulation results tested on the four different IEEE test is proposed for optimal PMU placement in power systems
systems. suffering from random component outages. The original
Now in the next section a review of mathematical tech- ILP method is valid to get a PMU placement solution which
niques used for optimal placement of PMU is presented. can make the whole system topologically observable with
minimum number of PMU.
In [41] authors proposed a new algorithm to determine
Mathimatical Methods optimum number of PMUs for fully observability of the
system under any contingency. The suggested algorithm
These methods are classified into three broad categories: operates in four stages; the essential measurement for totally
Exhaustive search, Integer Programming, Integer Quadratic known condition is found out in first stage. In next stage,
Programming (IQP). it selects redundant measurement set under any contin-
gencies. Above both stages uses least condition number
Exhaustive Search of normalized measurement matrix to select the proper
measurement location. Further, minimum redundant mea-
Exhaustive search is a general optimization technique that surement set is chosen from these measurements using
systematically enumerates all possible candidates for the the binary integer programming technique. Finally these
solution and selects the candidate that satisfied the con- measurements i.e. essential and minimum redundant mea-
straints at the optimum value of the objective function. Its surements are arranged such that locations of PMUs are
main advantage is that it guarantees the finding of the global minimum. For this purpose heuristic method is used. The
optimum. However, it is not suitable for large-scale systems IEEE test systems results show proposed placement method
with huge search space [84]. which ensures the system observability under the given con-
An exhaustive binary search method is implemented in tingency. Also it provides the reliable measurement to the
[38] to determine the minimum number of PMUs needed system.
to make the system observable. In case there is more than The contributions of [42] by using ILP are: new OPP
one placement set having the same minimum number of model is devised in which dc lines are modeled for observ-
PMUs, a method is proposed to select the one resulting in ability analyses. The new OPP model seeks the minimal cost
the most preferred pattern of measurement redundancy. Due of the PMU placement while the cost of each PMU is a func-
to its exhaustive nature, the method gives the global opti- tion of its measurement channels. This practical assumption
mal solution, and hence the results for a number of standard is emphasized in our optimization model by the allocation
test systems are reported to provide benchmark solutions for of more than one PMU to a single node because a PMU
researcher’s investigating various methods of optimal PMU would not have many channels due to technical limitations.
placement. The performance of the proposed model is examined on the
A contribution in [27] is the development of an exhaus- IEEE test systems and a real-size power system. An Inte-
tive binary search algorithm to determine the optimal place- ger Linear programming is used to determine the optimal
ment of PMUs as well as calculate the channels availabil- number and location of PMUs to make the system mea-
ity, the availability of measurement system of each PMU surement model observable and thereby it can be used for
placed, and the global avail ability of measurement and power system state estimation. When installing PMUs in a
maximize the average identification capability of multiple power system that already has SCADA measurement units,
line outages in [39]. it was suggested that the minimum amount of PMUs first be
Technol Econ Smart Grids Sustain Energy (2 0 1 6 ) 1 :1 0 Page 7 of 131 0

calculated and then install those PMUs in phases, where in A different formulation which is numerical and uses inte-
each phase a certain number of PMUs are installed in loca- ger programming allows easy analysis of network observ-
tions providing the highest possible level of obsevability ability [51]. An efficient and comprehensive formulation
chosen from the optimal locations[43, 44]. based on mixed integer linear programming method for the
In [45], OPP problem is solved using Integer Linear optimal placement of PMUs taking into account the channel
Programming considering sensitivity analysis. As the cost avail-ability and different contingency conditions associated
of one PMU is very high, so it is required to place min- with power systems, i.e. single line outage and single PMU
imum number of PMU’s without neglecting the safety loss were also considered in [52].
and security of the system. So the most sensitive buses In [53] mixed integer linear programming (MILP) frame-
in given power system based on load changes, are found work was formulated for the proposed model that simulta-
and PMU’s are optimally placed such that all the sensi- neously optimizes two objectives functions: (i) minimizes
tive buses are observed. A technique is proposed in [46] the number of PMUs and (ii) maximizes the expected value
for placing the PMUs in multiple stages over a given of system’s redundancy (or minimizes the un-observability
time period that ensures complete power system observ- of the system). Incorporating the impact of zero-injection
ability even under a branch outage or a PMU failure. The buses, the formulation is used to evaluate the probability
proposed technique is based on integer linear program- of un-observability of buses resulted from line outages and
ming (ILP) and a multi-criteria decision-making (MCDM) PMU loss.
approach. Integer Linear Programming method is used in An Integer linear programming based method for PMU
first stage to find the optimum number of PMUs for the placement has been considered to ensure complete system
complete observability of the system. These PMU place- observability under critical contingencies. In case of opti-
ments include the contingency cases such as single branch mum placement difficulties two indices have been proposed
outage and single PMU outage. A multi-criteria decision- to choose the optimal solutions; namely Bus Observabil-
making (MCDM) subsequently developed to prioritizing the ity Index (BOI) which gives the total number of PMUs
optimal PMU locations. This criteria include three indices, observing a given bus and to get the sum of all BOI for the
viz. tie-line oscillations observability index (TOI), volt- system, System Observability Redundancy Index (SORI) is
age control area observability index (VOI) and bus voltage used. An optimal PMU placement scheme is also be pro-
observability index (BOI). Finally, select the PMU loca- posed considering different PMU channel capacities. This
tions in each stage by ranking the optimal locations using has been also indicated to help the power system planner
the proposed indices in the MCDM process. he number of to greater extent from application-cost point of view for
PMU’s are further reduced and hence reducing the system PMU-PDC installation in the WAMS based power network
cost. at the planning stage. In addition to the cost factor, differ-
The authors in [47] addresses the various aspects of opti- ent criteria are suggested for the proper allocation of PMUs
mal PMU placement using binary integer linear program- in a given system. Network observability, state estimation
ming(BILP). Cases with and without zero injection buses accuracy and robustness present samples of such criteria
have also been considered. Minimum PMU placement prob- [54–60].
lem may have multiple solutions, therefore, two indices viz. In [61], authors proposed Integer Linear Programming
BOI (Bus Observability Index) and SORI (System Observ- with a fast and novel practical model to determine the opti-
ability Redundancy Index) have been proposed to further mum number of PMU. It consist power system contingency
rank these solutions. An integer-linear programming (ILP) conditions i.e. line outages and measurement losses. The
model is developed in [48] to find the minimum number proposed model is flexible than existing model because
and location of synchrophasors subject to the feasibility of other situation can be included separately or all together
proposed backup protection scheme. in power systems. Also, model consist measurement limi-
In [50] authors present an optimization model to deter- tations i.e. communication constraints of power networks.
mine optimum number of PMUs in power network. The The main importance of this method; it take low time getting
proposed model consists of observability necessities based the result also it gives local optimal solution for large-scale
on a set of probabilistic criteria. The nonlinear mathemati- power.
cal expression related with the probability of observability The problem of PMU placement is formulated and solved
index is changed to a linear model by a useful lineariza- as an IP problem [62]. The solution will provide the mini-
tion technique. This linear model further described as a mum number of strategically located PMUs that will elim-
Mixed Integer Programming (MIP) problem. This proposed inate measurement criticality in the entire system. This
mathematical model is therefore well-suited with the Mixed implies that any bad data appearing on any single measure-
Integer Programming solution tools and also it tested on ment will be detectable. It is shown that depending on the
large scale power systems. measurement configuration and the system topology, this
1 0 Page 8 of 13 Technol Econ Smart Grids Sustain Energy (2 0 1 6 ) 1 :1 0

goal can be achieved by using only a few extra PMUs to full observability and incomplete observability. The pro-
transform several critical measurements into redundant ones posed formulation improved the non linearity with and
in the system. without conventional power flow and power injection mea-
A new multi-stage PMU placement approach to maxi- surements.
mize the network observability during time horizons was In [71] outages and system states is identified with
proposed. Linear Integer Programming (LIP) is widely used the help of optimal sensors locations. In [72], an optimal
to solve the OPP problems. An optimal placement of syn- phasor measurement unit (PMU) placement model consid-
chronized phasor measurement units (PMUs) to ensure ering power system controlled islanding so that the power
complete power system observability in the presence of non- network remains observable under controlled islanding con-
synchronous conventional measurements and zero injec- dition as well as normal operation condition has been pro-
tions is also presented.[63–65]. posed using ILP. An ILP model for the optimal multistage
A Binary Integer Programming technique for instan- placement of PMUs is proposed in [73]. The approach finds
taneous PMUs placement and conventional power flow the number of PMUs and its placement in separate stages,
measurements for fault observability in power systems pro- while maximizing the system observability at each period
posed in [66]. The formulation is initially based on a non- of time. The model takes into account: the available budget
linear programming problem is unsolvable therefore initial per stage, the power system expansion along with the mul-
formulation results in a binary non-linear integer program- tistage PMU placement, redundancy in the PMU placement
ming problem that is transformed into a corresponding ILP against the failure of a PMU or its communication links,
problem through Boolean implications. The transformation user defined time constraints for PMU allocation, and the
provides a high scale of scalability. zero-injection effect.
In [67] using the Binary linear Integer Programming In [74], an integer linear programming (ILP) framework
method, local redundancy of the existing measurement is for the optimal placement of phasor measurement units
considered in the PMU placement strategy. Besides consid- (PMUs), in the presence of conventional measurements is
ering the number of PMUs implemented yearly limited, a presented. Furthermore, by the proposed method, the power
multistage placement strategy is proposed according to the system is to be remaining completely observable during
improvements on local redundancy. all possible single contingencies for lines and measurement
A new method has been proposed in [68] that approached devices. In [75] authors proposed a technique based on sin-
the PMU placement problem with purpose to ensure that gular value decomposition (SVD) to define the role of each
all WAMPAC applications attain complete measurement placement in dynamic stability assessment and find the best
observability. It presented an optimal method of placing PMU places to achieve this goal. Since observability is
PMUs for three categories of WAMPAC applications. The still a very important factor, the integer linear programming
author weighted the application by all three criteria and method is used to solve the optimal PMU placement prob-
categorize by the type of measurement shared. Category 1 lem for full system observability. The contribution of the
consists of applications that require phasor measurement authors in [76] is to present a PMU placement methodology
relative to a network’s stability. Category 2 requires a bus’ with observability redundancy; that is, to get more observ-
normal voltage and 3 requires measurement under faulted ability times of buses with the same number of PMUs. The
conditions. The phasing strategy ensures that initial PMU method employs MILP, to make it possible to explore the
placements enable Category 1 and subsequently 2 and 3. global optimal solution unlike nonlinear programming.
In [69] authors propose a new PMU placement algo- The aim of the authors in [77] is to recognize the effect
rithm, which can minimize the number of PMU, considering of channel capacity of PMUs on their optimal placement for
network connectivity, ZIB, N-1 line contingency, system complete power system observability. Initially, the conven-
reliability and voltage stability level. By introducing the tional full observability of power networks is formulated.
concept of system reliability, the proposed method can Next, a modified algorithm based on integer linear program-
effectively reduce the number of PMU while ensuring the ming model for the optimal placement of these types of
observability for major contingencies. The proposed method PMUs is presented.
also employs the Voltage Stability Index (VSI) to weigh the In [78] authors presents a novel integer linear program-
priority level of bus, which is used to screen the severer ming approach, which is shown to be equivalent to the
buses during contingencies. ES-based PMU placement. The linearity of circuit equa-
Based on the integer linear programming formulation tions ensures that there is a one-to-one relationship between
proposed for optimal PMU placement, the authors in the solutions of the proposed equivalent integer linear pro-
[70] presents a generalized integer linear programming gramming method (EILPM) and those of ES. Therefore,
formulation for cases including redundant PMU placement, impractical ES is translated into EILPM, which can be
Technol Econ Smart Grids Sustain Energy (2 0 1 6 ) 1 :1 0 Page 9 of 131 0

solved in a few seconds or minutes. Furthermore, additional constraint. In this technique all design variables take only
constraints for observability preservation following single an integer value [85]. An integer-quadratic-programming-
PMU or line outages can easily be developed in the pro- based method in [80] is used to determine the minimum
posed EILPM. Moreover, the limitation of communication number and the optimal locations of the PMUs to ensure
channels is also handled by rendering the nonlinear terms complete topological observability of the system. In [81] an
into linear ones. integer quadratic programming approach is used to deter-
mine the optimal locations of PMUs. The optimization
Branch and Bound Method process tries to attain dual objectives: 1) to minimize the
number of PMUs needed to maintain complete observabil-
In [86] authors presented the optimal positioning of the ity of the system for normal operating conditions as well
phasor measuring units for complete monitoring of power as for the outage of a transmission line or PMU and 2)
network in Khouzestan province in Iran. The (ILP) inte- to maximize the measurement redundancy at all buses in
ger linear programming based branch and bound method is the system. The method was applied on IEEE test systems
used to find the optimal number and position for PMUs. In considering the outage of a single transmission line or a sin-
[87] authors proposed an optimal phasor measurement units gle PMU. The proposed method can be used to determine
(PMUs) placement algorithm for power system observabil- PMU locations when conventional measurements, such as
ity and also increasing the performance of secondary volt- line flows and power injection measurements, are available.
age control scheme. The OPP is formulated to minimize the While the authors in [82] presents a new method for optimal
number of PMU installations subject to full network observ- placement of phasor measurement units (PMU) for com-
ability and monitoring pilot buses of the system to improve plete power system observability by formulating the optimal
secondary voltage control performance. The branch and PMU placement (OPP) problem as a quadratic minimiza-
bound optimization method is used to solve the OPP prob- tion problem with continuous decision variables subject to
lem which is suitable for problems with integer and boolean nonlinear observability constraints. The optimal solution
variables. is obtained by an unconstrained nonlinear weighted least
squares (WLS) approach.
Integer Quadratic Programming (IQP)

Quadratic Programming (QP) is solving the optimization Comparative Analysis of Mathematical

problem which has quadratic objective function and linear and Artificail Intellegence Techniques

S. Properties Mathematical Technique Artificial Intelligence

No. Techniques

1. Problem size Mathematical models are typically Artificial intelligence (AI) is

used to model a system when the a computational method that
system is not so complicated. Some attempts to imitate, in a very
complex problems have defied simplistic way, the human
solution using conventional mathe- cognition capability so as to
matical techniques. solve PMU placement prob-
lem
2. Data requirement Conventional techniques first use AI models are dependent
the principles (e.g., physical laws) on data that rely on the
to derive the relationships of the data alone to determine the
system, which usually justifiably structure and parameters
simplified with many assumptions that solve a system problem,
and require prior knowledge about with fewer assumptions
the nature of the relationships about the physical behavior
among the data. of the system.
3. Handling capability In the real world, it is prone The prior information of the
of non-linearity that complex and highly non-linear nature of the non-linearity is
problems are encountered. If the not required for AI models.
relationship between x and y is non-
1 0 Page 10 of 13 Technol Econ Smart Grids Sustain Energy (2 0 1 6 ) 1 :1 0

linear, mathematical analy-

sis can be applied success-
fully only if prior knowl-
edge of the nature of the
non-linearity exists.
4. Assumptions For most traditional math- There is less need to either
ematical models, the lack simplify the problem or
of physical understanding incorporate assumptions in
is usually supplemented by AI modals.
either simplifying the prob-
lem or incorporating several
assumptions into the mod-
els.
5. Problem solving approach Mathematical models rely AI techniques are data-
on assuming the structure of driven approaches in which
the model in advance, which the model development
may be sub-optimal. Hence is based on training of
many mathematical models input-output data pairs to
fail to simulate the complex determine the structure and
engineering problems parameters of the model.
6. Up gradation of results In the conventional mathe- AI models can always be
matical techniques this up updated to obtain better
gradation of the past results results by presenting new
is not possible. training examples as new
data become available
7. Precision & Accuracy As compare to AI tech- Artificial intelligence tech-
niques, mathematical mod- niques provide greater preci-
els not provide highest level sion and accuracy due to less
of accuracy due to various chances of error.
assumptions incorporated in
the modeling.
8. Efficiency Conventional mathematical Artificial intelligence tech-
techniques require more niques can be utilized in
time to solve the complex carrying out repetitive and
problem as compare to AI time-consuming tasks effi-
techniques. ciently.
9. Transparency & model uncertainty In mathematical modeling AI techniques are still fac-
such as MIP The existence ing classical opposition due
of integer decision variables to some inherent inadequacy
makes the problem hard to that need further attention in
solved using conventional the future including the lack
log barrier. of transparency, knowledge
extraction, and model uncer-
tainty
10. Prior information In mathematical techniques Special attention should be
are the process are mathe- paid to incorporating prior
matically proved. knowledge about the under-
lying physical process based
on engineering judgment or
human expertise into the
learning formulation.
Technol Econ Smart Grids Sustain Energy (2 0 1 6 ) 1 :1 0 Page 11 of 131 0

Conclusions 9. Dehghania M, Goelb L, Lic W (2014) PMU based observability

reliability evaluation in electric power systems. Electr Power Syst
Res 116(2014):347–354
Under smart gird environment the placement of phasor mea-
10. El-Zonkoly A (2006) Optimal meter placement using genetic algo-
surement units is very critical because accurate measure- rithm to maintain network observability. Expert Systems with
ment of active power transferred is required. Hence in this Applications 31(1):193–198
paper a comprehensive review of PMU placement methods 11. Gamm AZ, Kolosok IN, Glazunova AM, Korkina ES (2008) PMU
placement criteria for EPS state estimation, Electric Utility Dereg-
is presented. Optimal PMU placement (OPP) problem has
ulation and Restructuring and Power Technologies, 2008. DRPT
been solved by utilizing numerous optimization methods. 2008, Third International Conference on, Nanjuing:645–649
To obtain solutions for the OPP problem, two major tech- 12. Jamuna K, Swarup KS (2011) Optimal placement of PMU and
niques including mathematical methods and heuristic algo- SCADA measurements for security constrained state estimation.
Electr Power Energy Syst 33(2011):1658–1665
rithms exist. The correct methodology is required to find
13. Mousavian S, Feizollahi MJ (2015) An investment decision model
the optimal placement of PMUs in power network. The pre- for the optimal placement of phasor measurement units. Expert
sented review of heuristic and conventional/mathematical Systems with Applications 42(21):7276–7284
optimization techniques could largely help researchers in 14. Müller HH, Castro CA (2013) Optimal substation PMU place-
terms of employing new concepts to solve the OPP problem. ment method for the two-level state estimator. Electric Power
and Energy Conversion Systems (EPECS), 2013 3rd International
Further economic point can also be added to make opti- Conference on, Istanbul:1–6
mal PMU placement problem more practical to the industry 15. Sajan KS, Tyagi B (2011) Optimal placement of PMU with opti-
stakeholders. This review can be helpful to researchers mal branch current phasors for complete and incomplete observ-
ability. 2011 IEEE Power and Energy Society General Meeting,
for making improvement in already existing methods and
San Diego, CA, pp 1–5
develop new heuristic optimization approaches for multi- 16. Zhao Y, Yuan P, Ai Q, Lv T (2015) Optimal PMU place-
objective optimal PMU placement considering the con- ment considering topology constraints. Electr Power Energy Syst
straints. Additionally security assessment with PMU’s is 73(2015):240–248
17. Li Q, Negi R, Ilić MD (2011) Phasor measurement units place-
also a important issue which could also be incorporated.
ment for power system state estimation: A greedy approach. 2011
IEEE Power and Energy Society General Meeting, San Diego, CA,
pp 1–8
References 18. Li Q, Cui T, Weng Y, Negi R, Franchetti F, Ilic MD (2013) An
Information-Theoretic Approach to PMU Placement in Electric
Power Systems. IEEE Trans Smart Grid 4(1):446–456
1. Beg Mohammadi M, Hooshmand RA, Haghighatdar Fesharaki F 19. Yang Y, Roy S (2013) PMUplacement for optimal three-phase
(2016) A New Approach for Optimal Placement of PMUs and state estimation performance. Smart Grid Communications (Smart
Their Required Communication Infrastructure in Order to Mini- Grid Comm.), 2013 IEEE International Conference on, Vancou-
mize the Cost of the WAMS. In: IEEE Transactions on Smart Grid, ver, BC:342–347
vol 7, pp 84–93 20. Yang P, Tan Z, Wiesel A, Nehorai A (2015) Placement of PMUs
2. Mazhari SM, Monsef H, Lesani H, Fereidunian A (2013) A Considering Measurement Phase-Angle Mismatch. IEEE Trans
Multi-Objective PMU Placement Method Considering Measure- Power Delivery 30(2):914–922
ment Redundancy and Observability Value Under Contingencies. 21. Peng C, Sun H, Guo J (2010) Multi-objective optimal PMU
In: IEEE Transactions on Power Systems, vol 28, pp 2136–2146 placement using a non-dominated sorting differential evo-
3. Wen MHF, Xu J, Li VOK (2013) Optimal Multistage PMU Place- lution algorithm. Electr Power Energy Syst 32(2010):886–
ment for Wide-Area Monitoring. In: IEEE Transactions on Power 892
Systems, vol 28, pp 4134–4143 22. Hajian M, Ranjbarb AM, Amraee T, Mozafari B (2011) Opti-
4. Bhuvana Ramachandran A, Bellarmine TG (2013) Improving mal placement of PMUs to maintain network observability
observability using optimal placement of phasor measurement using a modified BPSO algorithm. Electr Power Energy Syst
units. Electr Power Energy Syst 56(2014):55–63 33(2011):28–34
5. Rakpenthai C, Premrudeepreechacharn S, Uatrongjit S, Watson 23. Rather ZH, Chen Z, Thø gersen P, Lund P, Kirby B (2015) Realis-
NR (2007) An Optimal PMU Placement Method Against Mea- tic Approach for Phasor Measurement Unit Placement: Consider-
surement Loss and Branch Outage. IEEE Trans Power Delivery ation of Practical Hidden Costs. In: IEEE Transactions on Power
22(1):101–107 Delivery, vol 30, pp 3–15
6. Aminifar F, Lucas C, Khodaei A, Fotuhi-Firuzabad M (2009) 24. Su C, Chen Z (2010) Optimal Placement of Phasor Measure-
Optimal Placement of Phasor Measurement Units Using Immunity ment Units with New Considerations. 2010 Asia-Pacific Power
Genetic Algorithm, vol 24 and Energy Engineering Conference, Chengdu, pp 1–4
7. Milosevic B, Begovic M (2003) Non-dominated sorting genetic 25. Wang JM, Li C, Zhang J (2012) Optimal Phasor Measurement
algorithm for optimal phasor measurement placement. In: IEEE Unit Placement by an Improved PSO Algorithm. 2012 Asia-
Transactions on Power Systems, vol 18, pp 69–75 Pacific Power and Energy Engineering Conference, Shanghai,
8. Allagui B, Aribia HB, Abdallah HH (2012) Optimal placement pp 1–4
of Phasor Measurement Units by genetic algorithm, 2012 First 26. Baldwin TL, Mili L, Boisen Jr MB, Adapa R (1993) Power Sys-
International Conference on Renewable Energies and Vehicular tem Observability With Minimal Phasor Measurement Placement.
Technology. Hammamet:434–439 IEEE Trans Power Syst 8(2):707–715
1 0 Page 12 of 13 Technol Econ Smart Grids Sustain Energy (2 0 1 6 ) 1 :1 0

27. Albuquerque RJ, Paucar VL (2013) Evaluation of the PMUs Mea- 46. Sodhi R, Srivastava SC, Singh SN (2011) Multi-criteria decision-
surement Channels Availability for Observability Analysis. In: making approach for multistage optimal placement of phasor mea-
IEEE Transactions on Power Systems, vol 28, pp 2536–2544 surement units. In: IET Generation, Transmission & Distribution,
28. Hurtgen M, Maun J-C (2010) Optimal PMU placement using Iter- vol 5, pp 181–190
ated Local Search. Electr Power Energy Syst 32(2010):857–860 47. Biswal A, Mathur HD (2015) Identification of optimal locations of
29. Liao C-S, Hsieh T-J, Guo X-C, Liu J-H, Chu C-C (2015) Hybrid PMUs for WAMPAC in smart grid environment. Advancements in
search for the optimal PMU placement problem on a power grid. Power and Energy (TAP Energy), 2015 International Conference
Eur J Oper Res 243(3):985–994 on, Kollam:369–374
30. Qi J, Sun K, Kang W (2015) Optimal PMU placement for power 48. Zare J, Aminifar F, Sanaye-Pasand M (2015) Synchrophasor-
system dynamic state estimation by using empirical observability Based Wide-Area Backup Protection Scheme with Data Require-
Gramian. 2015 IEEE Power & Energy Society General Meeting, ment Analysis. In: IEEE Transactions on Power Delivery, vol 30,
Denver ,CO, pp 1–1 pp 1410–1419
31. Saha Roy BK, Sinha AK, Pradhan AK (2012) An optimal PMU 49. Thorp JS, Phadke AG, Karimi KJ (1985) Real Time Voltage-
placement technique for power system observability. Electr Power Phasor Measurements for Static State Estimation. In: IEEE Power
Energy Syst 42(2012):71–77 Engineering Review, vol. PER-5, no. 11, pp 32–33
32. Tai X, Marelli D, Rohr E, Fu M (2013) Optimal PMU place- 50. Aminifar F, Fotuhi-Firuzabad M, Shahidehpour M, Khodaei A
ment for power system state estimation with random component (2011) Probabilistic Multistage PMU Placement in Electric Power
outages. Electr Power Energy Syst 51(2013):35–42 Systems, vol 26
33. Venkatesh T, Jain T (2015) Intelligent-search technique 51. Xu B, Abur A (2004) Observability analysis and measurement
based strategic placement of synchronized measurements for placement for systems with PMUs. Power Systems Conference
power system observability. Expert Systems with Applications and Exposition, 2004. IEEE PES 2:943–946
42(10):4768–4777 52. Abiri E, Rashidi F, Niknam T, Salehi MR (2014) Optimal PMU
34. Nuqui RF, Phadke AG (2005) Phasor measurement unit placement placement method for complete topological observabilit of power
techniques for complete and incomplete observability. In: IEEE system under various contingencies. Electr Power Energy Syst
Transactions on Power Delivery, vol 20, pp 2381–2388 61(2014):585–593
35. Yang Y, Shu H, Yue L (2009) Engineering practical method for 53. Aghaei J, Baharvandi A, Rabiee A, Akbari MA (2015) Probabilis-
PMU placement of 2010 Yunnan power grid in China. 2009 Inter- tic PMU Placement in Electric Power Networks: An MILP-Based
national Conference on Sustainable Power Generation and Supply, Multiobjective Model. In: IEEE Transactions on Industrial Infor-
Nanjing, pp 1–6 matics, vol 11, pp 332–341
36. Koutsoukis NC, Manousakis NM, Georgilakis PS, Korres GN 54. Sodhi R, Srivastava SC, Singh SN (2009) Optimal PMU place-
(2013) Numerical observability method for optimal phasor mea- ment to ensure system observability under contingencies, 2009
surement units placement using recursive Tabu search method. In: IEEE Power & Energy Society General Meeting, Calgary, AB:1–6
IET Generation, Transmission & Distribution, vol 7, pp 347–356 55. Dua D, Dambhare S, Gajbhiye RK, SOman SA (2008) Opti-
37. Peng J, Sun Y, Wang HF (2006) Optimal PMU placement for full mal Multistage Scheduling of PMU Placement An ILP Approach.
network observability using Tabu search algorithm. Electr Power IEEE Trans Power Delivery 23(4)
Energy Syst 28(2006):223–231 56. Kavasseri R, Srinivasan SK (2011) Joint Placement of Phasor and
38. Chakrabarti S, Kyriakides E (2008) Optimal Placement of Pha- Power Flow Measurements for Observability of Power Systems.
sor Measurement Units for Power System Observability. In: IEEE IEEE Trans Power Syst 26(4)
Transactions on Power Systems, vol 23, pp 1433–1440 57. Rihan M, Ahmad M, Beg MS, Anees MA (2013) Robust and eco-
39. Wu J, Xiong J, Shil P, Shi Y (2014) Optimal PMU placement nomical placement of phasor measurement units in Indian Smart
for identification of multiple power line outages in smart grids Grid. 2013 IEEE Innovative Smart Grid Technologies-Asia (ISGT
2014 IEEE 57th International Midwest Symposium on Circuits Asia) Bangalore:1–6
and Systems (MWSCAS), College Station, TX, pp 354–357 58. Bhonsle JS, Junghare AS (2015) An Optimal PMU-PDC Place-
40. Gou B, Kavasseri RG (2014) Unified PMU Placement for Observ- ment Technique in Wide Area Measurement System, 2015 Inter-
ability and Bad Data Detection in State Estimation, vol 29 national Conference on Smart Technologies and Management
41. Rakpenthai C, Premrudeepreechacharn S, Uatrongjit S, Watson for Computing, Communication, Controls, Energy and Materials
NR (2007) An Optimal PMU Placement Method against Mea- (ICSTM), India. 6 - 8 May 2015. pp 401–405
surement Loss and Branch Outage,” Power Engineering Society 59. Abbasy NH, Ismail HM (2009) A Unified Approach for the
General Meeting, 2007, IEEE, Tampa, FL, pp 1–1 Optimal PMU Location for Power System State Estimation, vol 24
42. Aminifar F, Fotuhi-Firuzabad M, Safdarian A, Shahidehpour M 60. Gupta N, Goyal M, Tripathy P (2012) A novel approach for opti-
(2014) Observability of Hybrid AC/DC Power Systems With mal placement of PMUs with minimum measurement channels,
Variable-Cost PMUs. In: IEEE Transactions on Power Delivery, Power and Energy (PECon), 2012 IEEE International Conference
vol 29, pp 345–352 on, Kota Kinabalu, pp 505–509
43. Gomathi V, Ramachandran V (2011) Optimal location of PMUs 61. Dua D, Dambhare S, Gajbhiye RK, Soman SA (2008) Optimal
for complete observability of power system network, Electrical multistage scheduling of PMU placement: An ILP approach. IEEE
Energy Systems (ICEES), 2011 1st International Conference on, Trans Power Del 23(4):1812–1820
Newport Beach, CA, pp. 314-317 62. Chen J, Abur A (2006) Placement of PMUs to Enable Bad Data
44. Fish AA, Chowdhury S, Chowdhury SP (2011) Optimal PMU Detection in State Estimation. In: IEEE Transactions on Power
placement in a power network for full system observability, 2011 Systems, vol. 21, no. 4, pp 1608–1615
IEEE Power and Energy Society General Meeting, San Diego, 63. Razavi SE, Falaghi H, Ramezani M (2013) A new integer lin-
CA,, pp 1–8 ear programming approach for multi-stage PMU placement. Smart
45. Yammani C, Narsi RK, Maheswarapu S (2015) Optimal place- Grid Conference (SGC), 2013, Tehran:119–124
ment of PMU’s considering sensitivity analysis, Advancements in 64. Wang F, Zhang W, Li P (2012) Optimal incremental placement
Power and Energy (TAP Energy), 2015 International Conference of PMUs for power system observability, 2012 IEEE Power and
on, Kollam, pp 40–44 Energy Society General Meeting, San Diego, CA, pp 1–7
Technol Econ Smart Grids Sustain Energy (2 0 1 6 ) 1 :1 0 Page 13 of 131 0

65. Korres GN, Manousakis NM, Xygkis TC, Löfberg J (2015) Opti- 76. Esmaili M, Gharani K, Shayanfar HA (2013) Redundant Observ-
mal phasor measurement unit placement for numerical observ- ability PMU Placement in the Presence of Flow Measurements
ability in the presence of conventional measurements using Considering Contingencies. In: IEEE Transactions on Power Sys-
semi-definite programming,” in IET Generation. Transm Distrib tems, vol 28, pp 3765–3773
9(15):2427–2436 77. Rashidia F, Abiria E, Niknamb T, Salehia MR (2015) Optimal
66. Kavasseri R, Srinivasa SK (2011) Joint placement of phasor placement of PMUs with limited number of channels for com-
and conventional power flow measurements for fault observ- plete topological observability of power systems under various
ability of power systems. IET Gene Trans Distrib 5(10):1019– contingencies. Electr Power Energy Syst 67(2015):125–137
1024 78. Azizi S, Dobakhshari AS, Sarmadi SAN, Ranjbar AM (2012)
67. Nan X, Beng GH (2013) Optimal PMU placement with local Optimal PMU Placement by an Equivalent Linear Formulation for
redundancy of conventional measurements. Computational Intel- Exhaustive Search. IEEE Trans Smart Grid 3(1)
ligence Applications In Smart Grid (CIASG), 2013 IEEE Sympo- 79. Yuill W, Edwards A, Chowdhury S, Chowdhury SP (2011) Opti-
sium on, Singapore:1–5 mal PMU placement: A comprehensive literature review. 2011
68. Fadiran JI, Chowdhury S, Chowdhury SP (2013) A multi-criteria IEEE Power and Energy Society General Meeting, San Diego, CA,
optimal phasor measurement unit placement for multiple appli- pp 1–8
cations, 2013 IEEE Power & Energy Society General Meeting, 80. Chakrabarti S, Kyriakides E, Albu M (2009) Uncertainty in Power
Vancouver, BC, pp 1–5 System State Variables Obtained Through Synchronized Measure-
69. Putranto LM, Hara R, Kita H, Tanaka E (2014) Voltage stability- ments. IEEE Trans Instrum Meas 58(8)
based PMU placement considering N− 1 line contingency and 81. Chakrabarti S, Kyriakides E, Eliades DG (2009) Placement of
power system reliability Power Engineering and Renewable Synchronized Measurements for Power System Observability.
Energy (ICPERE), 2014 International Conference on, Bali, pp IEEE Trans. Power Delivery 24(1)
120–125 82. Manousakis NM, Korres GN (2013) A Weighted Least Squares
70. Gou B (2008) Generalized integer linear programming formu- Algorithm for Optimal PMU Placement. IEEE Trans Power Syst
lation for optimal PMU placement. IEEE Trans Power Syst 28(3)
23(3):1099–1104 83. Manousakis NM, Korres GN, Georgilakis PS (2012) Taxonomy of
71. Zhao Y, Chen J, Goldsmith A, Poor HV (2014) Identification of PMU Placement Methodologies. IEEE Trans Power Syst 27(2)
Outages in Power Systems With Uncertain States and Optimal 84. Nazari-Heris M, Mohammadi-Ivatloo B (2015) Application of
Sensor Locations. IEEE J Sel Top Sign Proces 8(6) heuristic algorithms to optimal PMU placement in electric
72. Huang L, Sun Y, Xu J, Gao W, Zhang J, Wu Z (2014) Optimal power systems: An upyeard review. Renew Sust Energ Rev
PMU Placement Considering Controlled Islanding of Power Sys- 50(2015):214–228
tem. In: IEEE Transactions on Power Systems, vol 29, pp 742– 85. More KK, Jadhav HT A Literature Review on Optimal Placement
755 of Phasor Measurement Units, 2013 International Conference on
73. Gómez Oscar, Rı́os MA (2013) ILP-based multistage placement Power, Energy and Control (ICPEC):2013
of PMUs with dynamic monitoring constraints. Electr Power 86. Moarref B, Mehdipour A, Yazdanipour M (2012) Using branch
Energy Syst 53(2013):95–105 and bound method and optimal placement of phasor measurement
74. Azizi S, Gharehpetian GB, Dobakhshari AS (2013) Optimal Inte- unit of power network in Khouzestan province in Iran, 2012 IEEE
gration of Phasor Measurement Units in Power Systems Consider- International Conference on Power Electronics, Drives and Energy
ing Conventional Measurements. In: IEEE Transactions on Smart Systems (PEDES). Bengaluru:1–6
Grid, vol 4, pp 1113–1121 87. Mohammadi-Ivatloo B, Hosseini SH (2008) Optimal PMU place-
75. Dehghani M, Shayanfard B, Khayatian AR (2013) PMU Ranking ment for power system observability considering secondary volt-
Based on Singular Value Decomposition of Dynamic Stabil- age control, Electrical and Computer Engineering, 2008. CCECE
ity Matrix. In: IEEE Transactions on Power Systems, vol 28, 2008. Canadian Conference on, Niagara Falls, ON, pp 000365–
pp 2263–2270 000368