Overview of Mathematical Methods for Energy

Management Optimization in Smart Grids

Lamyae Mellouk1, Abdessadek Aaroud1, Driss Benhaddou2, Khalid Zine-Dine1, Mohammad Boulmalf3
1
LAROSERI Lab., ChouaibDoukkali University, FS, El Jadida, Morocco
[email protected], [email protected], [email protected]
2
University of Houston, USA
[email protected]
3
University International of Rabat, Morocco
[email protected]

Abstract—Electricity demand increases every year and there is a Analytic optimization methods include : gradient-based
need to improve the existing power grid. Smart Grid concept was optimization algorithm, quadratic programming, interior point
developed by Institutes and researchagencies to address this need. method and nonlinear programming.
This new concept can be generally considered as an intelligent
power grid system that implements distributed energy. Gathering, This paper aims to answer to the question, why opt to use
coordinating operation and control of multiple distributed energy Genetic algorithm than other methods. To do this we group
resources, energy storages, loads and set of users pose a great various fitness based on genetic algorithm used in
challenge. Effective managementof smart gridwill have a benefit theliterature, compares parameters of these function, discuss
of reduced cost of energy, reducedstress on the electric energy and summarize equality and inequality constraints
infrastructure, and improved energy efficiency. This paper
reviews the research and studies on mathematical models used to This paper is structured as follows. In Section II, we present
optimize energy management in smart grid. The paper compares an overview about complexity classes and NP-hard problem
and discussesvarious fitness functions used in the literature. mean? In Section III, we review the Genetic Algorithm. We
Genetic Algorithm is explored and brief overviews of other
then give a brief overview to the other management methods
methods arediscussed.
used in Microgrid in Sections IV, .in Sections IV we conclude
Keywords-component; Genetic Algorithm;Microgrid; Optimal this paper and present some perspective.
Reactive Power Dispatch (ORPD)
II. COMPLEXITY AND NP-HARD PROBLEM?
I. INTRODUCTION A. Complexity classes
A key power generation paradigm enabled by Smart Grid In practice, heuristic techniques use the enumeration of all
(SG) isthe distributed generation (DG). DG takes advantage of possible solutionsto find the optimum solution. As the set of
distributed energy resource (DER) systems (e.g. solar panels feasible solutions increases, it is sometimes very hard (or takes
and small wind turbines), which are often small-scale power an exponentially long time) to find the best solution. Thus, the
generators (typically in the range of 3 kW to 10,000 kW), in time to research for the optimal solution is a very important
order to improve the power quality and reliability [1]. factor.Moreover, it is necessary to study the intrinsic difficulty
of solving mathematical problems, expressed as the
The intermittent and fluctuant generation associated with complexity of the problem. Complexity class describes how
renewable energy sources, lead tothe incorporation of demand/ difficult or easy a problem can be solved.For example, if the
response between source of energy and end users, this made complexity of the function is polynomial in regards to the size
ancillary facilitiesstorage and effective management of energy of the input data, then we say the algorithm runs in polynomial
vital to large scale deploymentof DG. From computation point time and the problem is class P.
of view, management of generation, consumption, and storage
becomes a multi-variable optimization problem. B. NP-hard problem
The time to research the optimal solution in combinatorial
In order to solve the optimization problems related to
optimization problems are considered to be so difficult that the
management objectives, such as improving energy efficiency,
time to find a solution is non-deterministic (NP). NP-hard
controlling emission, reducing cost, profiling demand,optimal
problems are the class of problems that can be resolved in
reactive power dispatch (ORPD) and maximizing
"nondeterministic polynomial time" and therefore are very
utility,researcher have developed multiples mathematical
difficult to solve. We can reduce Problem B (NP) to Problem
tools, modeland solve these optimization problems with an
A (P) if, given a solution to Problem A, we can easily verify
increase complexity that involveuncertainty introduced by
that it is alsoa solution to Problem B. In this case, "easily"
energy. Monte Carlo, game theory and Genetic Algorithms
means "in polynomial time ".
(GA)and other methods canbe used to achieve
thesemanagementobjectives. Thesemethods are classified in to C. Why use Intelligent Algorithms or heuristic?
analytic and heuristic optimization methods[2]. Some combinatorial optimization problems can be solved in
polynomial time as a function of the size of input data.
However, when the exact solution methods are not applicable,

978-1-4673-7894-9/15/$31.00 ©2015 IEEE

the approximations algorithms (or heuristics) are used to find accordingtomathematical model identified as a non-linear
approximate solutions in a reasonable time. mixed-integer problemby the optimization of GA.
A heuristic is an optimization algorithm that aims to find a In [14], Celli used Ant colony approachto reduce total
feasible solution for the objective function, but without production cost, thisobjective is achieved ThroughOptimal
guarantee of optimality. Meta-heuristic methods are methods Power Flow (OPF), by determining optimal generator output,
that employ heuristics techniques with guidance through identified as a non-linear problem. The paper comparedGA
thesearch space in order to exploit its best capabilities with a with Ant colony and found out that Ant colony approach
powerful mechanism to achieve better solutions. requires high cost to generate solutions, despite that it offers
best results than GA. The computation time for a single
Load distribution problems, Reactive power flow and voltage
iteration, by Ant Colonyrequires very low number of iterations
stability in smart grid are non-linear, multi-objective or multi-
compared to the number of iterations performed by theGA.
constraints optimization problem.To solve optimization
We propose hybridizationof GA with an exact method, to
problems in smart grids, algorithms such as linear
evolve the GA by a search for a local optimum and to improve
programming LP, Davidon Fletcher-Powel (D.F.P)are
convergence time.
explored in several articles. However,these methods allows
only local minimum. Heuristic algorithms, such as Genetic The objective function in [9]describes the reactive power
Algorithms, Bee colony optimization, Particle swarm optimization, Mohsenian-Rad proposes single-objective
optimization algorithm, Ant colony, simulated annealing and optimization and the reactive power output at generatoris
Tabu search are promising methodsto solve these kids of reduced by using GA and compared to Linear
problem. Programming.The problem is treated as a single objective
problem. The best trade-off relations between cost and
D. Meta- heuristic Classification
emission cannot directly be found.Pugazhenthi1[2] compared
In this paper we classify these methods according to their evolutionary algorithm like harmony search algorithm,
evolution to find the optimal solution. Simplified Genetic Algorithm(SGA) and Particle Swarm
Individual approach: Start from one initial solution S(0) and Optimization (PSO) for Optimal Reactive Power
evolve through a trajectory which converges to a solution S Dispatch(ORPD) problem. They showed advantages of this
(n) that is an optimum.Some example of these methods algorithm in decreasing transmission loss, reducing voltage
are:Descent method (local min / max), Annealing simulates deviation and increasing the voltage stability index.
and Tabu Method.
Faisal [6] determines the optimal use of the natural resources
Population approach or evolutionary methods start with available, by using the GA approach to solve the multi-
several initials solution.These approaches requests intensive objective problem.The purpose is to minimize the total
computations to find solutions (computing in parallel, operating cost based on sold power produced by Wind turbine
distributed).Some examples of these methods are Genetic and Photovoltaic Cell while taking in consideration
Algorithm, Ant colony, Bee colony optimization, and particle environmental factors. Objective functions yields a set of
swarm optimization algorithm. Pareto optimal solutions, however, one solution is needed to
be selected, is called the best compromise solution. The
III. LITERATURE REVIEW ABOUTGENETIC
weighted sum method used to handle multi-objective
ALGORITHM
optimization.
Vijayakumar [13] proposed distributed algorithm and system
reconfiguration via power-electronics and switches for the Elsied[4]used GA to simultaneously solve the cost function of
real-time accessibility control, local state maintenance, and the transmission and availability of renewable energy as
security enhancement. An embedded intelligence is integrated competing objectives.
into the power-electronics to enable the system Table I to TableV, respectively,summarize the existing fitness
reconfiguration. Genetic Algorithm (GA) is chosen to address function and there equality and inequality constraints, discuss
the reactive power management and optimum transmission and summarize objectives functions, summarize theparameters
device placement. taken into accountbyeachauthors to treat their fitness function,
The execution time of the resource scheduling is decreased,the summarize theparameters takeninto accountbyeachauthors to
faults are effectively detected, prevented, and mitigated, and treat equality and inequality constraints.
energy resource scheduling is modeled
 TABLE I. A BRIEF COMPARISON BETWEEN THE EXISTING FITNESS FUNCTION AND THERE

Fitness Function Papers

𝐧𝐫𝟏 𝐧𝐫𝟐 𝐧𝐫𝐦

In [13]
𝐦𝐢𝐧 ƒ 𝐬, 𝐜 = 𝐦𝐢𝐧 𝐏𝐄𝟏 𝐓𝟏,𝐭 𝐠 𝐄𝟏 𝐓𝟏,𝐭 + 𝐏𝐄𝟐 𝐓𝟐,𝐭 𝐠 𝐄𝟐 𝐓𝟐,𝐭 + ⋯ + 𝐏𝐄𝐦 𝐓𝐦 , 𝐭 𝐠 𝐄𝐦 (𝐓𝐦 , 𝐭)
𝐓𝟏 =𝟏 𝐓𝟏=𝟐 𝐓𝐦 =𝟏

∀𝐭 ∈ 𝟏, … , 𝐓

𝒏𝑮
In [11]
𝑴𝒊𝒏 𝒇 𝑷𝑮 = 𝒇𝒊 𝑷𝑮𝒊
𝒊=𝟏

𝒏
𝑭(𝒙) = 𝑽𝒍 = (𝑳𝒋𝟐 )
𝒋=𝒈+𝟏 In [9]

𝐍
𝐂𝐅(𝐏) = 𝐢=𝟏(𝐂𝐢 𝐅𝐢 (𝐏𝐢)+𝐎𝐌𝐢 (𝐏𝐢 ) + 𝐒𝐓𝐂𝐢 + 𝐃𝐂𝐏𝐄𝐢 − 𝐈𝐏𝐒𝐄𝐢 )
In [6]
𝐍

𝐄(𝐏) = 10-2(𝛂𝐢 + 𝛃𝐢 𝐏𝐢 +𝛄 𝐏𝐢 )+ζ i exp(𝛌𝐢 𝐏𝐢 )

𝐢=𝟏

ug
Di t PDERi t CDERi t + Di t SUCDERi t
T In [4]
+
i=1
ƒ x = ug
t=1
Dj t PESSj t CESSj t + Pgr t Cgr t − Pgs Cgs (t)
j=1

TABLE II. CONSTRAINTS FOR EACH PRECEDENT FITNESS FUNCTIONS

Equality
Papers Inequality constraints State variables
constraints

Voltage limits at communication links. Ploss: Total power losses in the distribution
∆P1=PGt−PLt−Pl
In [13] The limits of the loading factor k. Lines are the 5% value of the former
∆Q1=QGt− QLt−Ql
Generator input power limits. parameter.

𝒏𝑮
𝒏𝒄
𝑷𝑮𝒊 – 𝑷𝒄𝒉𝒋 − 𝑷𝑳 = 𝟎
𝒋=𝟏
𝑷𝒎𝒊𝒏 𝒎𝒂𝒙
𝑮𝒊 ≤ 𝑷𝑮𝒊 ≤ 𝑷𝑮𝒊
𝒊=𝟏

P loss is calculated by two different methods

In [11] 𝒏𝑮
𝒏𝒄
𝐐𝐦𝐢𝐧 𝐦𝐚𝐱
𝐆𝐢 ≤ 𝐐𝐆𝐢 ≤ 𝐐𝐆𝐢 𝑸𝑮𝒊 – 𝑸𝒄𝒉𝒋 − 𝑸𝑳 = 𝟎
𝒋=𝟏
𝒊=𝟏

In [9] 𝐦𝐢𝐧
𝐕𝐆𝐢 𝐦𝐚𝐱
≤ 𝐕𝐆𝐢 ≤ 𝐕𝐆𝐢 𝐍 𝐦𝐢𝐧
𝐕𝐆𝐢 𝐦𝐚𝐱
≤ 𝐕𝐆𝐢 ≤ 𝐕𝐆𝐢
𝑷𝒊 = 𝐕𝐢 𝐕𝒋 𝐆𝒊𝒋 𝐜𝐨𝐬 𝛅𝐢𝐣 + 𝐁𝒊𝒋 𝐬𝐢𝐧 𝛅𝐢𝐣
𝐐𝐦𝐢𝐧 𝐦𝐚𝐱
𝐆𝐢 ≤ 𝐐𝐆𝐢 ≤ 𝐐𝐆𝐢 𝐣=𝟏 𝐐𝐦𝐢𝐧 𝐦𝐚𝐱
𝐆𝐢 ≤ 𝐐𝐆𝐢 ≤ 𝐐𝐆𝐢
𝐍
𝐦𝐢𝐧
𝐓𝐆𝐢 𝐦𝐚𝐱
≤ 𝐓𝐆𝐢 ≤ 𝐓𝐆𝐢 𝑸𝒊 = 𝐕𝐢 𝐕𝒋 (𝐆𝒊𝒋 𝐬𝐢𝐧 𝛅𝐢𝐣 + 𝐁𝒊𝒋 𝐜𝐨𝐬 𝛅𝐢𝐣)
𝐣=𝟏

𝐨𝐧
(𝐓𝐭−𝟏,𝐢 − 𝐌𝐔𝐓𝐢 ) (𝐔𝐭−𝟏,𝐢 − 𝐔𝐭,𝐢 )≥ 0
𝐍
𝐨𝐟𝐟
(𝐓𝐭−𝟏,𝐢 − 𝐌𝐃𝐓𝐢 ) (𝐔𝐭−𝟏,𝐢 − 𝐔𝐭,𝐢 )≥ 0
𝐏𝐢 = 𝐏𝐋 − 𝐏𝐏𝐕 − 𝐏𝐖𝐓 − 𝐏𝐛𝐚𝐭𝐭
In [6] 𝐢à𝟏
𝐏𝐢𝐦𝐢𝐧 ≤ 𝐏𝐢 ≤ 𝐏𝐢𝐦𝐚𝐱 ∀𝐢 = 𝟏, … , 𝐍
 Ug
ul ug
max
EDERi t PDERi t ≤ EDER
PL t − PDERi t
i=1
Us l=1 i=1
Us
max
EESSi t PESSi t ≤ EESS PDERj (t) − Pgr t + Pgs t = 0
In [4]
j=1
max j=1
Eg (t)Pg (t) ≤ Egrid

TABLE III. DISCUSS AND SUMMARIZE OBJECTIVES FUNCTION

Papers Objectives function Discussion

Fault detection, degree of power saving due to power
optimization, Memory usage, consumer-GW Address the reactive power management and optimum transmission device placement.
In [13]
capacity constraints of energy storage devices and emission of power generation units
(gateway) communication overhead, and consumer
were not included
computational overhead.

Comparing GA with Ant colony, Ant colony approach is the high cost solutions
generation. The computation time for one iteration by ACA leads to very low number of
In [11] Optimization of cost function by determining the
iterations compared to the number of iterations performed by the genetic algorithm.
Optimal power flow
Hybridization of the GA with an exact method, Evolve the GA by a search for a local
optimum and to improve convergence time.

Enhance voltage stability, Decrease power loss by Problem is treated as a single objective problem. The best trade-off relations between
In [9]
minimizing l-indice load buses cost and emission cannot directly find

Minimize the operation cost ($/h).

Reliability of the choice of decision maker is not accurate. Authors find the best
Minimize the emissions (kg/h).
In [6] compromise solution between operational costs, emission but without addressing voltage
Ensure that the load is served according to the
stability.
constraints
Optimal distribution of powers, and reduce pollutants without providing an overview about time convergence to find the optimum and voltage
In [4]
emission. stability has not been studied

TABLE IV. SUMMARIZE THEPARAMETERS TAKEN INTO ACCOUNTBYEACHAUTHORS TO TREAT THEIR FITNESS FUNCTION

Reliability
Generating Emission Storage
Papers Fuel Cost cost(maintenance/v Start-up Cost Purchase/sell Cost
Cost Cost Cost/charge discharge
oltage stability)

In[13]


In[11]

In [9] 

In [6]     

In [4]     
 TABLE V. SUMMARIZE THEPARAMETERS TAKEN INTO ACCOUNT BY EACH AUTHORS TO TREAT EQUALITY AND INEQUALITY
CONSTRAINTS

Reactive
Active Power Start-up Transformer taps Voltage Active power Reactive power
Paper power
balance time position constraints generate generate
balance
   
In[13]

   
In[11]

In [9]    

In [6]   

In [4] 

In [12], Huangdeveloped an online adaptive electricity

IV. BRIEF OVERVIEW TO THE OTHER scheduling algorithm for smart energy management in MGs
MANAGEMENT METHODS USED IN MICROGRID by jointly considering renewable energy penetration.
TheLyapunov optimization method was applied to solve the
A. GAME THEORY problem with an efficient online electricity scheduling
Ibars[7] proposed capacity management technique that algorithm, which has deterministic performance bounds.
involves consumers to manage their own demands in order
V. DISCUSSION AND PERSPECTIVES
to minimize a cost Function. The system is modeled as a
Optimization techniques should target specific constraints
non-cooperative network congestion game and finds the
such as power quality, operation efficiency, User satisfaction
local optimum points of the Nash equilibrium, and also
and reliabilityin the model while achieving optimal
shows that it is possible to arrive at a global solution of the
network. However, cooperation of users is not always operation and efficient load distribution in smart grids. In
guaranteed. addition, it is important to explore more effective and
efficient algorithms to solve the optimal load distribution
Mohsenian-Rad [8] presents an autonomous and distributed and optimal reactive power problems of MG. Some of the
demand-side energy management system among users based traditional optimization methods may not be suitable for MG
on game theory; the strategies of the players are the daily optimal load distribution or reactive power, which are a
schedules of their household appliances and loads. The nonlinear, multi-constraint, and multi-objective optimization
global optimal performance in terms of minimizing the problem.Intelligent algorithms, such as genetic algorithms,
energy costs is achieved at the Nash equilibrium of the bee colony optimization, and particle swarm optimization
formulated energy consumption scheduling game.In [5] algorithm which may at the same timeoptimize several
Belganaproposestwo trading schemes, the first mechanism is objectives like as reducing emissionof pollutant, reducing
a pricing mechanism for interconnected smart cost function and maximizing capacity line and keeping
microsourcesbased on potential game theory. The second system power balance charge–discharge. These algorithms
algorithm is an analytic model usingStackelberg game are promising methods to solve management of MG. The
approach based on a bi-level hybrid multi-objective selection of an appropriate energy management model in
evolutionary algorithm (BL-HMOEA) to find optimal practical smart grid applications is of great importance and
strategies that maximize the profit of utilities and minimize further research is needed to tackle this challenging issue.
carbon emission as well as total power loss.
NOMENCLATURE AND SYMBOL
B. MONTE CARLO
P Active power generation.
Ghiani[10] proposed an algorithm in order to find the g Generation cost.
optimal combination of MGs, the algorithm uses a m Number of energy resources.
SequentialMonte Carlo simulation technique ,it maximizes E Various energy resources.
the sum of the savings in both the cost of energy purchasing ∆P1 Global Active power at (t).
and the cost of service interruptions. PGI Active power generated at (t).
PL Load Active power at ( t).
C. NEURAL NETWORKS (NN) Pl Power losses.
Celli [14] developed a novel energy management system Q Reactive Energy.
f(PG) Total cost function.
(EMS)and autonomously determines the correct
nG Generator number.
dispatchinghour by hour, using neural networks.A result by PGi Total power generation.
test case has showed significant saves in the global energy QGi Reactive power.
bill by using typical Italian energy and fuel price data. PGi Active power generated at i.
Pchj Power consumed at node j.
D. THE LYAPUNOV OPTIMIZATION TECHNIQUE PL Total energy lost.
nG Generator number.
nc Consummator number.
Q Reactive Energy. [3] European SmartGrids Technology Platform;Vision and strategy for
L Indices of all load buses. europe’s electricity networks of the
Vl Voltage stability. future,www.smartgrids.eu/documents/vision.pdf, 2006.
CF(P) Operating costs in $/h. [4] M. Elsied, A. Oukaour, H. Gualous and R. Hassan ,“Gestion de
Ci Fuel costs in $/l forthe DG. l'énergie et optimisation du système multisourcesbasée sur
Fi (Pi ) Fuel consumption rate by generator i. l'algorithme génétique”, Symposium de GénieElectrique 2014, Jul.
OMi(Pi) Operation and maintenance cost of the generating in $/h. 2014, Cachan, France <hal-01065314>.
Pi Real power output from generating unit i. [5] A.Belgana, M.Maier, LLe and F. Labeau, “Générationdistribuée dans
N Total number of generating units. un reseauélectrique intelligent : une approche suivant la théorie des
STCi Start-up costs of the unit generator i $/h. jeux“, IEEE Transactions on Smart Grid, Octobre 2014.
DCPEi Daily purchased electricity of unit i.
[6] Faisal A. Mohamed and Heikki N. Koivo,”Modelling and
IPSEi Daily income for sold electricity of unit i.
Environmental/Economic Power Dispatch of MicroGrid Using
E(P) Emission of pollutants.
MultiObjective Genetic Algorithm Optimization” , Chapter 16 in
αi; βi; γi; δi, λi Emission characteristics of the ith unit.
Fundamental and Advanced Topics in Wind Power, Dr. Rupp
Di State ON = 1/OFF = 0 of generator.
Carriveau (Ed.), ISBN: 978-953-307-508-2, published July 05, 2011
PDERi Distributed energy resource (kW). under CC BY-NC-SA 3.0 license.
CDERi Total operating cost of DER.
PESSj Charge -discharge battery. [7] C. Ibars, M. Navarro, and L. Giupponi, “Distributed Demand
Pgs/Pgr Sell/purchase power from/to up-grid. Management in Smart Grid with a Congestion Game“, in Proc. 1st
CESSj Total cost of storage element. IEEE International Conf. on Smart Grid Communications,
SUCDERi Startup cost of distributed resources. Gaithersburg, Maryland, USA, Oct. 2010.
Vgi Generator output voltage. [8] A.Mohsenian-Rad, V .W.S. Wong, J. Jatskevich, R . Schober and A.
Ti Transformer taps position. Leon-Garcia,”Autonomous Demand-Side Management Based
Qi Switchable VAR compensator. onGame-Theoretic Energy Consumption Scheduling”, IEEE
Vi,Vj Voltage at bus i, j. Transactions on Smart Grid,Vol.1, no. 3, pp.320 - 331,December
Bij and Gij Conductance and scuptance of the line connecting bus i 2010.
δij and j. [9] R.Katuri, A. Jayalaxmi, G.Yesuratman and D. Yaddanalapalli,
T Phase angle difference of voltage. “Genetic algorithm Optimisation of Generator Reactive Power”,
U Off/On time for unit i at(t-1). AASRI Conference on Power and energy systems,pp. 192 – 198,2012.
Pi On/Off on the (0, 1) status. [10] E.Ghiani,S. Mocci and F.Pilo,“Optimal reconfiguration of distribution
PPV Real power output of unit i. networks according to the MicroGrid paradigm”, International
Pwt Output power of photovoltaic. Conference on Future Power Systems. Amsterdam, pp.1– 6,
Pbatt Output power of wind turbine in kW. DOI:10.1109/FPS.2005.204290, November 2005.
EDERi,EESSj,Eg Output power of battery in kW. [11] Y. Mimoun, M. Rahl, M. Abid and M. Kandouci, “Optimisation
Pgs/Pgr Respectively total emission of air pollutants. D’écoulement Des Puissances Par AlgorithmesIntelligents“, Rev.
Sell/Purchase power from/to upgrid. Roum. Sci. Techn. –Électrotechn. et Énerg, Bucarest, pp. 3–12, 2007.
[12] Y. Huang, S.Maoand R.M. Nelms, “Adaptive Electricity Scheduling
in Microgrids“, IEEE Transactions On Smart Grid, vol. 5, NO.
1,pp.270-281,January 2014.
[13] D. Vijayakumar and V.Malathi, “A real-time management and
REFERENCES evolutionary optimization scheme for a secure and flexible smart grid
[1] X. Fang, S. Misra, G. Xue , and D. Yang, “Smart grid – the new and towards sustainable energy”, International Journal of Electrical Power
improved power grid: Asurvey“,IEEE Communications Surveys & Energy Systems.Elsevier, Vol. 62, pp 540–548, November 2014.
&Tutorials, vol. 14,no. 4,pp 944–980, 2012. [14] G.Celli,F. Pilo, G.Pisanoand G.GSoma,“Optimal participation of a
[2] S.P ugazhenthi, R. Sures, and j.Baskaran., “A Comparative Study Of microgrid to the energy market with an intelligent EMS“, In the
Varios Evolutionary Methods For Optimal Reactive Power Seventh International Power Engineering ConferenceIPEC,
Dispatch“, National Conference on Research Advances in Singapore,vol.2, pp. 663 – 668,November 2005.
Communication, Computation,Electrical, Science andStructures
(NCRACCESS), ISSN 2348-8379,2015.