Proceedings of 2018 International Conference on Emerging Trends and Innovations in Engineering and Technological Research (ICETIETR)

Optimal Power Flow using Distributed Generation

and Conservation Voltage Reduction techniques
for Micro-Grids
Z. Haidar A. Waqar NM. Adamali Shah
Electrical Engineering Department Electrical Engineering Department Electrical Engineering Department
King Saud University, King Saud University, King Saud University,
Riyadh, KSA Riyadh, KSA Riyadh, KSA
[email protected] [email protected] [email protected]

Khalid Al-Mutib
Electrical Engineering Department
King Saud University,
Riyadh, KSA
[email protected]

Abstract--Voltage control is critical in power systems The optimization techniques which are implemented for
operation and control. With the increase in using distributed OPF are two types: classical methods and evolutionary
generation DG, it becomes essential to ensure that voltage methods. Classical methods involve linear [2] and
stability is main-tained. This work studies using DG and nonlinear program-ming (LP) [3] , quadratic programming
conservation voltage reduction CVR for improving voltage (QP) [4], and interior point method (IM) [5]. Above
profile, reducing genera-tion cost, and power loss. The size mention method have certain drawback, due to its non-
and location of the DG was optimized based on the voltage
convexity nature of OPF, the classi-cal methods have some
profile criterion. The buse’s voltages were reduced to the
lower band to decrease generation cost and total power loss.
deficiency in finding the global min-ima. Gradient-based
Optimal Power flow using PSO algo-rithm was implemented methods overcome this is- sue; but, sometime cannot
to investigate the power system voltage profile, 30 bus IEEE meet the inequality constraints [6].Because of this
systems is used in this work. The results showed that limitations, evolutionary methods were developed to
accompanying DG with CVR can reduce genera-tion cost solve the OPF problem effectively. Evolutionary
and power loss by 13.14 and17.22 methods consist of genetical algorithm(GA) [7]
[8],evolutionary pro-gramming (EP) [9], particle
Keywords: Distributed generation, conservation voltage swarm optimization (PSO)[10][11],simulated
reduction, optimal powerflow, PSO annealing(SA)[12],differential al-gorithm (DE) [13],
shuffle frog leaping algorithm (SFLA) [14]. These
methods were comprehensively re- viewed in [15].
Some researchers investigated hybridizing two methods in
I.INTRODUCTION OPF. Roy et al [16] used a modified shuffle frog leaping
algorithm (SFLA) with GA in optimizing an economic
1.1. Optimal Power Flow(OPF) load dispatch problem. Other researchers used chaos
Optimal Power Flow (OPF) study targets to optimize cer- optimization with linear IM method [17]. Genetic
tain objective by changing the variables of power sys- algorithm (GA) with Fuzzy logic was also implemented
tem, without disturbing constrain of the system. Therefore, in OPL [18]. A hybrid method of PSO, GA and Fuzzy
OPF is one of the main areas of study in power systems. Logic techniques was im-plemented in OPF [19].
This study consists of nonlinearity with continuous and Placing and sizing distributed gen-eration is another
discrete variables which is extremely con strained. In 1962, attention grabbing area of research. The placement of
Carpen-tier published the first paper in OPF [1], the OPF distributed generation in distribution networks was
problem has been done by researchers. optimized using GA [20]. A selection approach of the

978-1-5386-5744-7/18/$31.00©2018IEEE 1
 Proceedings of 2018 International Conference on Emerging Trends and Innovations in Engineering and Technological Research (ICETIETR)

proper buses forth optimal integration of DG was pro- that constrain the optimization process which include
posed to decrease trans- mission power losses [21]. Most volt- age magnitude, generators active and reactive power,
of OPF studies used sys- tems with constant loads. How- and angle and lines limits; u is the system controlled pa-
ever, variable loads were also treated in [22], [23], and rameters which control such as generators voltage, active
[24]. power, and shunt VAR compensations it is a vector quanti-
ty, load MW and MVAr (load shedding), x is also a vector
1.2. Particles Swarm Optimization (PSO) Tech nique quantity which has the dependent variables such as PQ
bus power, that consist of voltages magnitude and phase
R.C.Eberhart and J.Kennedy was the first to introduce angles, generator reactive power outputs for bus voltage
PSO [25]. As described by the authors, PSO are based control, and transmission lines parameters.
upon the social-psychological metaphor; entities of the
population (referred to as particles) adapts by returning 1.4. Fuel Cost Objective Function
stochastically toward previously successful re- gions. PSO
operators based on two main divisions : position and ve- One of the main part of objective function in this study
locity update. In each iteration interval, each par ticle is is the fuel cost function for total generation, which is ex-
directed toward their previous best position and the global pressed as [28]:
best position. For a new velocity value is computed for
NG 2
each particle based on its current velocity, F(P)= i i+ iPi+ iP ) i (8)
and the distance from the global best position. The new where
generated velocity value is used for computing the next and NGcorrespond to the number of generators. To attain
position of the particle in search space. This procedure is min- imum generation cost for a particular load, the non-
then repeated a set number of times or until a minimum linear equality constrains are givenas:
error is achieved[26].Equation(1),and(2)shows the PSO-
process corresponding to the velocity and position[25]: Pi−PGi+PDi=0 (9)
Qi−QGi+QDi=0 (10)
vij vij+c1r1 (pij−xij)+c2r2(pg−xij) (1) t
The i h bus bar active and reactive power generation equa-
tion (Piand Qi) are obtained as:
xij=xij+vij (2)
NB
Where c1, c2 r1, r2 are ran- Pi=|Vi |Vi||Yij|cos
i=1 i j ij) (11)
dom variables with values evenly distributed between Qi=|Vi NB
|Vi||Yij|sin i j ij) (12)
i=1
them, pg is the global best position, xij is current position
,pij the current best position and vijis the velocity of parti-
cles.
the angle of ith
iis ijis the admittance be-
tween buses i and j, and NB correspond to the number of
1.3. Formulation of Optimal Power Flow buses. In all duration, the total generation should meet the
total load and its transmission losses, this is the main con-
The general mathematical form of OPF is as follows strain, which is given as:
[27]:
NG ND
minimize(x,u) (3) i=1 PGi− PDi+PL=0
i=1 (13)
for nonlinear equality constraints is given as where ND correspond to the number of load buses and PL
is the total transmission losses. The above equations must
G(x,u)=0 (4)
be solved with the following generators constraints:
And for inequality constraints is given as PminP≤Gi ≤ Pmax
QminQ≤Gi Qmax ≤ Vmin
H(x,u)<0 (5) ViVmax≤ ≤
min≤ i≤ max
umin<u<umax (6)
where,Qminand Qmaxand Pminand Pmax,indicates the min-
xmin<x<xmax (7) imum and maximum reactive power and active power
whereF(u,x)is the objective function to be minimized(or constraints for each generator respectively. such as Vmi-
maximized); H(x, u) represents the nonlinear power nand Vmax min maxindicate the minimum and max-
flow equations;G(u,x) represents the control parameters’ imum voltage magnitude and voltage angle constraints for
limits each bus respectively.

2
 Proceedings of 2018 International Conference on Emerging Trends and Innovations in Engineering and Technological Research (ICETIETR)

1.5.Conservation Voltage Reduction

Power utilities have used Conservation voltage reduc-

tion (CVR) since the 1970s [29]. Basically, CVR is lower-
ing the supply voltage at the consumer terminal to the
lower band. The aim of implementing CVR is to reduce
the peak demand and losses. ANSI Standard C84.1 [30]
states that thevoltageattheconsumersidemust-
be5%ofthenominal value. Hence, the voltage can be set at
any value within this bandwidth [-5, +5]% of the voltage
nominal value. How- ever, reducing the voltage to the
lower band width results in the above mentioned benefits.
The typical percent of this voltage reduction is 2%-3%
[31]. The benefits of implementing CVR is reported in the
literatures. In [32] 1% re- ductioninvoltageresultsin0.58-
0.98% reduction in active power.
In[33],applying1%reductioninvoltagedecreased energy
demand by 0.25 -1.3%. BC Hydro conducted CVR on the
Fig 1. IEEE 30 Bus single line diagram
distribution system and reported a 1% reduction in ener-
gyanda1.1%reductioninpeakdemand,foravoltage reduction
by 1%[34]. Of the base case to check the voltage violation at each
This paper presents an optimization of size and loca- case. The buses where voltage violation occurred repre-
tion of distributed generation (DG) with implementing sented the candidate buses for inserting DG. Then, DG is
CVR technique. PSO algorithm is used as an optimization inserted to correct the voltage violation .For economic
technique and the well-known 30 bus IEEE system was considerations, the number of the inserted DGs was in-
used. The effect of distributed generation on the genera- creased gradually starting from 1 bus only, 2 buses only
tion scheduling and active power losses was studied with and 3 buses, etc. If the system voltage is within limits
and without applying CVR. The contribution of this study during 10% of load changingwithDGat1bus,then DG will
is that it introduces the use of CVR at the generation level not inserted at the other buses and the optimal location is
not at the distribution level as it is the traditional practice. achieved.
It investigates the benefits of applying CVR in the gen-
eration level. The study showed that DG can help in ap-
plying CVR at the generation level as it does in the dis-
ribution level. It showed that DG with CVR can signifi-
cantly enhance the power systems performance economi-
cally and technically. The improvement on the voltage
pro- file, system lines MVA relief and reduced cost opera-
tion were demonstrated. The layout of IEEE 30 bus sys-
tem is shown in Fig. 1. Full details of this system compo-
nents can be found in [9]. The total load of the original
IEEE 30 bus system is 283.4 MW. In this system there are
6genera- tors available at the bus number 1, 2, 5, 8, 11,
and 13. The load flow analysis of the system is done
Fig 2. Voltage profile of the base case
based on Newton-Raphson method at each time step.PSO
algorithm was implemented to find the optimal generation
scheduling during the study. However, the voltage of all buses -at this case- is still
within limits. More improvement is possible if DG is used
RESULTS AND DISCUSSION at the buses where the voltage is low. The location of the
DG is chosen based the voltage profile. Hence, the buses
The voltage profile of the IEEE 30 bus system is shown 19, 26 and 30 are chosen to locate the DG in this study.
inFig. 2. Thebusvoltageisloweratthebuses26and The PSO method is used to determine the optimal size of
30. These buses are subjected to voltage violation if the theDG.Fig.3showsthebusesvoltageprifileofthesystem for
connected load demand increases. To choose the optimal three cases: the base case, with DG and with DG and
place for the DG, The buses load was varying by±10% CVR. At the base case (the dotted line),thebuses26and30

3
 Proceedings of 2018 International Conference on Emerging Trends and Innovations in Engineering and Technological Research (ICETIETR)

Fig 3. Voltage profile of the system with DG, and GG with CVR.

Fig 5. Active power loss with DG, and DG with CVR

Fig 4. Generation cost with DG, and DG with CVR

have the lowest values of voltage and therefore, these

buses have the probability to violate the system voltage
limits if any disturbance or demand increase occur. There-
fore, these buses with bus 9 (which has the lowest voltage
after buses 26 and 30) are chosen to locate the DG to Fig 6. Lines loading in p.u. with DG, and DG withCVR
boost the voltage profile of the system.
The total grid loss and the generation cost of the base case-
are13.24MWand818.43$/h respectively. The solid line
with triangular in the figure shows the improvement of the
voltage profile when DG was implemented. The voltage of
the buses 26 and 30 was lifted to above 0.98 and 0.99 re-
spectively. Moreover, all the PQ buses voltages were im-
proved. The generation cost andpower-
lossdecreasedto731.8$/hand11.5MW which represents
10.58 % and 13.14 % decrease respectively.
The solid line with squares represents the voltage with
DG and CVR. The voltage was reduced and kept within
0.98 and 0.95. It should be noted that, the applying CVR
will not be possible without voltage violation if DG was
not used. This is because of the lower values of voltage at
buses 26 and 30. The total generation cost and power loss Fig7.LinesloadingaverageswithDG,andDG withCVR
are 684.8 and 10.96 respectively, which reduces the cost
and loss by 13.14 % and 17.22 %. The improvement in
the cost and power loss is summarized in Figures 4 and 5
respectively.
Fig. 6 shows the systems lines MVA loading for the

4
 Proceedings of 2018 International Conference on Emerging Trends and Innovations in Engineering and Technological Research (ICETIETR)

three cases. It is obvious that using DG -in general- de- [7] A. G. Bakirtzis, P. N. Biskas, C. E. Zoumas, and V.Petridis, ”Op-
creases the MVA loading which enhances the systems se- timalpowerflowbyenhancedgeneticalgorithm,”IEEE transactions
on Power Systems, vol. 17, Issue. 2, pp. 229– 236, May2002.
curity and delays congestion problems. The MVA will de-
[8] D. Devaraj and B. Yegnanarayana, ”Genetic-algorithm- based op-
crease more at some lines than others. For example, line timal power flow for security enhancement,” IEE Proceedings-
numbers 24,25,36 has more MVA relief because they are- Generation, Transmission and Distribution, vol. 152, pp. 899-
connected or close to the buses where DG is connected. 905,2005.
The effect of the applying CVR with the DG is shown in [9] Y.R. Sood, ”Evolutionary programming based optimal power flow
red line. It is shown that the applying CVR could further and its validation for deregulated power system analysis,” Interna-
tional Journal of Electrical Power Energy Systems, vol. 29, pp. 65-
improve the performance of the system by reducing the 75,2007.
line’s MVA loading. However ,This effect is not equal for [10] M. Abido, ”Optimal power flow using particle swarm opti- miza-
all lines. The average lines loading p .u. decreased from tion,” International Journal of Electrical PowerEnergy Systems,
0.19 for the base case to 0.17 and 0.16 for the case of DG vol. 24, pp. 563-571,2002.
and DG with CVR respectively and this is shown in Fig.7. [11] N. Mo, Z. Zou, K. Chan, and T. Pong, ”Transient stability con-
strained optimal power flow using particle swarm opti- misation,”
IET Generation, Transmission Distribution, vol. 1, pp. 476-
CONCLUSIONS 483,2007.
[12] Y.-J. Jeon and J.-C. Kim, ”Application of simulatedanneal- ing and
tabu search for loss minimization in distribution systems,” Interna-
This paper has DG with CVR on the performance of tional Journal of Electrical Power Energy Systems, vol. 26, pp. 9-
micro-grids. An optimization of fuel cost conduction on 18,2004.
30 bus IEEE system for three cases are: the base case, [13] A. A. El Ela, M. Abido, and S. Spea, ”Optimal power flow using
with DG and DG with CVR. The place of DG was deter- differential evolution algorithm,” Electric Power sys- tems re-
search, vol. 80, pp. 878-885,2010.
mined on voltage profile improvement and the size of DG
[14] T. Niknam, M. rasoul Narimani, M. Jabbari, and A. R. Malekpour,
was a optimized to achieve the most economic conditions. ”A modified shuffle frog leaping algorithm for multi-objective op-
CVR technique was implemented to further improve the timal power flow,” Energy, vol. 36, pp. 6420-6432,2011.
system by reducing the power loss and the cost of genera- [15] S.Frank,I.Steponavice,andS.Rebennack,”Optimalpower flow: a
tion. It was shown that the accompanying DG with CVR bibliographic survey I,” Energy Systems, vol. 3, pp. 221-258,2012.
reduce the generation cost and power loss by 13.14 % and [16] P. Roy, P. Roy, and A. Chakrabarti, ”Modified shuffled frog leap-
ing algorithm with genetic algorithm crossover for solving eco-
17.22%respectively. nomic load dispatch problem with valve-point effect,” Applied
Soft Computing, vol. 13, pp. 4244-4252, 2013.
References [17] L. Shengsong, H. Zhijian, and W. Min, ”A hybridalgorithm forop-
timalpowerflowusingthechaosoptimizationandthe linear interior
[1] M. Huneault and F.D. Galiana, “A survey of the optimal power- point algorithm,” in Power System Technol-
flowliterature,”IEEEtransactionsonPowerSystems, vol. 6, pp. 762– ogy,2002.Proceedings.PowerCon2002.InternationalCon- ference
770, May1991. on, 2002, pp.793-797.
[2] K. Zehar and S. Sayah, “Optimal power flow with environ- mental [18] A.Saini,D.K.Chaturvedi,andA.Saxena,”Optimalpower flow solu-
constraint using a fast successive linear program- ming algorithm: tion: a GA-fuzzy system approach,” International journal of
Application to the algerian powersystem,” Energy Conversion and emerging electric power systems, vol. 5,2006.
Management, vol. 49, pp. 3362- 3366, November2008. [19] S. Kumar and D. Chaturvedi, ”Optimal power flowsolution us-
[3] M. Shahidehpour and V.Ramesh, “Nonlinear program- ming algo- ingfuzzyevolutionaryandswarmoptimization,”Interna- tional Jour-
rithms and decomposition strategies for OPF,” IEEE/PES tutorial nal of Electrical Power Energy Systems, vol. 47, pp. 416-
on optimal power flow,1996. 423,2013.
[4] J.S. Lipowski and C. Charalambous, “Solution of optimal load [20] R. Rajaram, K. S. Kumar, and N. Rajasekar, ”Power sys- tem re-
flow problem by modified recursive quadratic- programming configuration in a radial distribution network for re- ducinglosses-
method,” IEE Proceedings-Generation, Transmission and Distribu- andtoimprovevoltageprofileusingmodified plant growth simulation
tion, vol. 128, pp. 288–294, September1981. algorithm with DistributedGenera- tion (DG),” Energy Reports,
[5] F. Capitanescu, M. Glavic, D. Ernst, and L. Wehenkel, “Interior- vol. 1, pp. 116-122,2015.
point based algorithms for the solution of optimal powerflowprob- [21] L. Arya, A. Koshti, and S. Choube, ”Distributed genera- tionplan-
lems,”ElectricPowersystemsresearch,vol. 77, pp. 508–517, ningusingdifferentialevolutionaccountingvoltage
April2007.
[6] D. I. Sun, B. Ashley, B. Brewer, A. Hughes, and W. F. Tinney,
“Optimal power flow by Newton approach,” IEEE Transactions
on Power Apparatus and systems, vol. PAS- 103, Issue. 10, pp.
2864–2880, October1984.

5
 Proceedings of 2018 International Conference on Emerging Trends and Innovations in Engineering and Technological Research (ICETIETR)

stability consideration,” International Journal of Electrical Power

Energy Systems, vol. 42, pp. 196-207, 2012.
[22] K.R.GuerricheandT.Bouktir,”Optimalallocationandsiz- ing of dis-
tributed generation with particle swarm optimiza- tion algorithm
for loss reduction,” Revue des Sciences etde la Technologie, vol.
6, pp. 59-69,2015.
[23] R. S. Rao, K. Ravindra, K. Satish, and S. Narasimham, ”Power
loss minimization in distribution system using net- work reconfig-
uration in the presence of distributed gener- ation,” IEEE transac-
tions on Power Systems, vol. 28, pp. 317-325,2013.
[24] M. Kowsalya, ”Optimal size and siting of multiple dis- tribut-
edgeneratorsindistributionsystemusingbacterialfor- agingoptimiza-
tion,”SwarmandEvolutionaryComputation, vol. 15, pp. 58-
65,2014.
[25] J. Kennedy, J. F. Kennedy, R. C. Eberhart, and Y.Shi, Swarm intel-
ligence: Morgan Kaufmann,2001.
[26] M. Settles, ”An introduction to particle swarm optimiza- tion,”
Department of Computer Science, University of Idaho, pp. 1-
8,2005.
[27] Z. Qiu, G. Deconinck, and R. Belmans, ”A literature sur- vey of
optimal power flow problems in the electricity mar- ket context,”
in Power Systems Conference andExposition, 2009. PSCE’09.
IEEE/PES, 2009, pp.1-6.
[28] S.A.-H.SolimanandA.-A.H.Mantawy,Modernoptimiza- tion tech-
niques with applications in electric powersystems: Springer Sci-
ence Business Media,2011.
[29] R. Preiss and V. Warnock, ”Impact of voltage reduction on energy
and demand,” IEEE Transactions on PowerAppara- tus and sys-
tems, pp. 1665-1671,1978.
[30] ”Voltage Rating for Electric Power Systems andEquipment (60
Hz), ANSI Standard C84.1-2006,”2006.
[31] P. K. Sen and K. H. Lee, ”Conservation voltage reduction tech-
nique: An application guideline for smarter grid,” IEEE Transac-
tions on Industry Applications, vol. 52, pp. 2122- 2128,2016.
[32] E. Diskin, T. Fallon, G. O’mahony, and C. Power,”Conser- vation
voltage reduction and voltage optimisation on Irish distribution-
networks,”inIntegrationofRenewablesintothe Distribution Grid,
CIRED 2012 Workshop, 2012, pp.1-4.
[33] R. Beck, ”Distribution efficiency initiative,” NorthwestEn- ergy
Efficiency Alliance,2007.
[34] V. Dabic, C. Siew, J. Peralta, and D. Acebedo, ”BC hydro’s experi-
enceonvoltageVARoptimizationindistributionsys- tem,” in Trans-
mission and Distribution Conference and Ex- position, 2010 IEEE
PES, 2010, pp.1-7.