Available online at www.sciencedirect.

com

ScienceDirect
Energy Procedia 90 (2016) 137 144

D 5th International Conference on Advances in Energy Research, ICAER 2015, 15-17 December
2015, Mumbai, India

Energy Savings in Radial Distribution Systems with Intermittent

Wind Power and Probabilistic Load Demands
Sooraj Narayan Ka, Ashwani Kumarb, *
a
National Institute of Technology, Kurukshetra-131961, India
b
National Institute of Technology, Kurukshetra-131961, India

Abstract

This paper discusses the energy savings in radial distribution systems with wind turbine integration with probabilistic load power
demands. The energy savings are determined by calculating the difference cost of energy losses of the distribution system
without and with the wind power source. Deterministic load flow (DLF) studies do not incorporate the randomness of the load
power demands. The probabilistic load flow (PLF) method incorporates this randomness to provide more realistic and accurate
results regarding the performance of the system. In this paper, the load and the substation voltages of the distribution system are
considered as random variables for the calculation of energy savings. The main contributions of this paper are: i) Comparison of
annual energy savings and cost of energy losses obtained using DLF and PLF calculations ii) Analysis of 24 hour energy savings
of radial distribution systems with the integration of wind turbines using practical 24 hour wind data iii) Comparison of energy
savings of radial distribution systems for winter and summer seasons. The comparison of energy loss savings is done on
deterministic and probabilistic platforms. The study is conducted for winter and summer seasons using the corresponding hourly
wind data for both seasons and hourly load demand curves for both seasons. This study can help extensively in the realistic
analysis of radial distribution systems with renewable integration. The results are obtained for a standard IEEE 33 bus radial
distribution system.
2016
2016TheTheAuthors.
Authors. Published
Published by Elsevier
by Elsevier Ltd. Ltd.
This is an open access article under the CC BY-NC-ND license
Peer-review under responsibility of the organizing committee of ICAER 2015.
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the organizing committee of ICAER 2015
Keywords: Radial Distribution Systems; Wind Turbine Integration; Probabilistic Load Flow; Energy Savings;

* Corresponding author.Tel: (+91) 9416366091, Fax: (+91) 1744238050.

E-mail address: ashwa_ks@yahoo.co.in

1876-6102 2016 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the organizing committee of ICAER 2015
doi:10.1016/j.egypro.2016.11.178
138 K. Sooraj Narayan and Ashwani Kumar / Energy Procedia 90 (2016) 137 144

1. Introduction

Electric power systems are in a continuous state of evolution and are fast evolving into a completely intelligent
system [1]. Distributed Generation (DG) has procured an expanded enthusiasm for as far back as decade because of
the techno-socio-economical advantages provided by it. The techniques and strategies embraced for the planning
and operation of the customary unidirectional power system frameworks must be altered and remade with the mix of
DG's into the distribution systems. This is chiefly in view of the fact that DG's convert the formerly passive
distribution system into an active one. The different benefits of integrating DGs into the distribution system have
been well examined and explored in the past [3-5]. Among its various benefits, the integration of DG units in
distribution systems provides significant savings in the cost energy losses. The total power losses are reduced with
DG integration and hence the energy losses are also reduced. Accurate estimation of annual energy savings with the
integration of DG resources plays a very important role in the planning of power distribution systems. Wind energy
is one of the fastest growing sources of energy. In the year 2014, the annual installation level rose up to 50GW,
while in 2013, the annual installations were about 35 GW [6]. Wind power integration into the distribution system
also poses some technical challenges. The modeling of these resources is a significantly important part of the
analysis. Since the wind speed is intermittent, probabilistic distribution functions are be used to model these
variables [7]. The impacts of renewable integration on distribution system parameters have to be studied before
integration in order to ensure system security and reliability.
For the analysis of a distribution system to be realistic, the instantaneous dynamism of the system has to be taken
into consideration while performing load flow studies. One of the vital steps in that direction is the load flow
analysis of the distribution system in a probabilistic perspective. The probabilistic analysis can incorporate the
stochastic behavior of the distribution system [8]. It can help provide better and more accurate results regarding
various system variables. Precise and accurate modelling of feeder loads is essential for any study regarding
distribution systems. In addition to modelling the load probabilistically, this work has also considered real time
stochastic intermittency of substation voltages also [9]. In [10], the authors discussed the energy loss savings for a
mesh distribution system in the presence of distributed generation. In the above referenced paper, the authors also
incorporated the impacts of time varying realistic ZIP load models on the energy loss savings of mesh distribution
systems with DG penetration. In [11], the authors discussed the use of capacitor banks for maximizing the net energy
savings in radial distribution systems. The maximization of energy savings is done in the referenced paper by
minimizing the power losses using various algorithms like Cuckoo search and Bat algorithm. In [12], the authors
proposed a new algorithm for the accurate estimation of energy savings in a smart distribution network for Volt/Var
control. The algorithm presented in the referenced paper for energy saving estimation is non-intrusive in nature. In
[13], the authors presented a new methodology for allocating capacitors in distribution systems based on maximizing
the energy loss reduction. The maximization of reduction in energy losses effectively maximizes the increase in
energy savings with capacitor allocation. In [14], the authors presented the use of artificial bee colony algorithm for
the maximization of net annual cost savings of energy loss with optimally allocated capacitor banks in radial
distribution systems. In [15], the authors discussed about the energy savings by using energy efficient distribution
transformers. In [16], the authors discussed the planning of energy saving modification or ESM for distribution
networks. In [17], the authors presented the use of a two-point estimate method for the probabilistic load flow in
order to analyze a distribution system considering wind generation. A new probabilistic load flow method was
proposed based on voltage drop calculation for distribution systems with wind power in [18]. In [19], integration of
wind power and electric vehicles were both considered to carry out a constrained probabilistic load flow of
distribution systems. In [20], the unsymmetrical two point estimate method is used to solve the probabilistic load
flow problem of distribution networks with integrated wind and solar PV modules. The literature review for this
paper has revealed that very few authors have addressed the energy savings in distribution systems with intermittent
wind power integration. Many authors have focused on analysing the specific economic benefits of wind energy
integration in distribution networks like power loss reduction and voltage profile improvement. Even though many
of these works have incorporated the probabilistic nature of the system, the accurate estimation of annual energy
savings from a probabilistic perspective with wind integration needs to be addressed. In this paper, the energy
savings are determined by calculating the cost of losses of the system without and with the wind power source. The
energy losses before and after wind turbine integration are compared to obtain the energy loss savings. This
 K. Sooraj Narayan and Ashwani Kumar / Energy Procedia 90 (2016) 137 144 139

comparison of energy loss savings is done on deterministic and probabilistic platforms. The probabilistic load flow
or PLF is solved using the Monte Carlo Simulation or MCS method. The study is conducted for winter season and
summer season using the corresponding 24 hour wind data for both seasons and 24 hour load demand curves for
both seasons. The main contributions of this paper are: i) Comparison of annual energy savings obtained using DLF
and PLF calculations ii) Analysis of 24 hour energy savings of radial distribution systems with the integration of
wind turbines using practical 24 hour wind data iii) Comparison of energy savings of radial distribution systems for
winter and summer seasons. The cost of energy losses are computed mathematically from the total annual energy
loses of the system. The comparison of energy loss savings is done on deterministic and probabilistic platforms. The
study is conducted for winter season and summer season using the corresponding 24 hour wind data for both seasons
and 24 hour load demand curves for both seasons. The results are obtained for a standard IEEE 33 bus radial
distribution system [21].

Nomenclature

C

Cost of energy losses




Energy savings
(+,%'",

Operating cost of wind turbine for hour ,





Active load power demand at bus %





Reactive load power demand at bus %


.


Power output of wind turbine in MW
-


Wind speed in m/s
-!%


Cut-in speed of the wind turbine in m/s
-!(


Cut-out speed of the wind turbine in m/s
-*


Rated speed of the wind turbine in m/s
*


Rated power output of the wind turbine in MW



Voltage stability index

2. Cost of energy losses, energy savings and cost of wind

2.1. Cost of energy losses

Based on the mathematical equation presented in [10], the annual cost of energy losses () can be calculated
as follows:


= (Total Real Power Loss)*(L+$ ) $ (1)

where

: Annual demand cost of power loss ($/kW) , 
: Annual cost of energy loss ($/kWh) , +$
: Loss factor .
Loss factor +$
can be defined in terms of load factor ($) as:

+$
= &!$+ 1&! $2 (2)

The values taken for the calculations are [10]:&!=0.2, $=0.47, N57.6923 $/kW, N
0.00961538 $/kWh

2.2. Energy savings

The energy savings or the savings in cost of energy losses is obtained as follows:


140 K. Sooraj Narayan and Ashwani Kumar / Energy Procedia 90 (2016) 137 144


N
.(M
$ (3)

Where .(
is the cost of energy losses without wind turbine integration and 
is the cost of energy
losses with wind turbine integration.

2.3. Cost of wind

The cost of operating the wind turbine for each hour ,
((+,%'",) can be calculated as follows [22]:


(+,%'",N..L ..,

(4)
where
.
Fixed costs of the wind turbine
.
Variable cost coefficient of the wind turbine
.
Coefficient denoting the operational availability of the wind turbine for hour ,
(1- available, 0-not available)
.
Power output of wind turbine for hour ,
in MW.
The values of the coefficients taken for determining the hourly costs of wind turbine are [22]:
.N12.8 $, .N75 $/MWh, .N1 for all the 24 hours of study.

3. Modeling of Wind Turbine

3.1. Wind power source modeling

The uncertainty of wind turbine output at any location mainly arises due to the variation in wind speed and air
density. Since wind speed varies frequently, it is considered to be a random variable in the radial power flow
calculation. In this paper, the Weibull distribution has been used to sample wind speed. The Weibull distribution
function is a two parameter function which is used to describe wind speed mathematically as [8]:
; ? 3?
$ - N R S;31 #/)R S; , Q - Q 
(5)
9 9 9
where
- is the wind speed in m/s,
& is the shape parameter and !
is the scale parameter.
The power output from a wind turbine is given by the following equation [7]:
 - Q -9:
&1 - L &2  -9: O - O -=
@ N (6)
=  -= O - O -9<
 - P -9<
where @ is the power output of wind turbine in MW, -9:
is the cut-in speed of the wind turbine in m/s,

-9< is the
cut-out speed of the wind turbine in m/s,

-= is the rated speed of the wind turbine in m/s, = is the rated power
6
output of the wind turbine in MW,

&1 N J and
&2 N M&1  -9: .
?J 3?HI

3.2. Probabilistic load modeling

The distribution system load demands are uncertain and this uncertainty is handled using probability distribution
functions. The probabilistic nature of load at each bus in a distribution system can be incorporated into load flow
studies by visualizing the loads as random variables distributed with a variance about a mean value. The load
demands at each bus are assumed to be random variables with Gaussian or Normal distribution [7].
1 R6DI 34EDI SC
$ 5: N #/) M (7)
BEDI 2A 2BEDI C

1 R7DI 34FDI SC
$ 5: N #/) M
(8)
BFDI 2A 2BFDI C

where, 5: and 5: are the active and reactive load demands at bus number %. 65:  75:  065: and 075: are the
mean and standard deviation values of load active and reactive powers respectively. The PLF is done incorporating
 K. Sooraj Narayan and Ashwani Kumar / Energy Procedia 90 (2016) 137 144 141

all the uncertain system variables using MCS [18].

3.3. Probabilistic substation voltage modeling

The substation voltage may also vary instantaneously similar to the load variation. This instantaneous variation in
the substation voltage can be modelled using the normal distribution function [9].
1 R8K
K 34GK
K SC
$ >
> N #/) M (9)
BGK
K 2A 2BGK
K C

where, >
> is the nominal substation bus voltage (1.00 p.u.) and 8>
> , 08>
> are the mean and standard deviation
values of substation voltage respectively. Optimal location of wind turbine has been obtained finding voltage
stability index (VSI).

4. Results and discussion

The results are obtained for the standard IEEE 33 bus radial distribution system given in [9]. The base power of
the system is 100 MVA and the base voltage is 12.66 KV. The total connected active power load is 3.72 MW and
reactive power load is 2.30 MVAR. The deterministic load flow or the DLF of the distribution system is carried out
using the forward/backward load flow method [9]. The wind turbine selected for this study has the following
specifications: -9: N2.5 m/s, -= =13 m/s, -9< =25 m/s and = =0.6 MW [10]. A total of 20 wind turbines are assumed
to be integrated into the test distribution system. 24 hour data of a day is used to represent the two seasons, winter
and summer, of the year. The mean and standard deviation of 24 hour wind speed data is obtained for Cape
Comorin, a site in South India, for the months of January and May. The mean and standard deviations of wind speed
for summer and winter seasons used in this paper are shown in Fig. 1.

Mean wind speed (m/s) for summer Standard deviation of wind speed (m/s) for summer
10
Wind speed (m/s)

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time (Hours)
Fig.1. Mean and standard deviations of wind speed for summer and winter seasons

Fig. 2 shows the cost of energy losses in $ obtained from DLF and PLF calculations in summer and winter
seasons of the 33 bus RDS. Fig. 3 shows the comparison of total real power loss or TPL values obtained from DLF
and PLF calculations in summer and winter seasons of the 33 bus RDS. From Fig. 2, it is evident that there is a
significant reduction in the cost of energy losses of the system in both summer and winter seasons after integrating
the wind turbines.
From Fig. 3, it can be observed that the minimum cost of energy losses is obtained for hour 5 in summer season
without wind turbine placement. The value of this minimum cost of energy losses for hour is 4966.203$ using DLF
and 4973.194$ using PLF. The corresponding real and reactive power losses are 61.83085 kW and 41.86635 kVAR
respectively. The minimum bus voltage for this hour is 0.948197 p.u. The minimum cost of energy losses for
summer season after wind turbine placement is also obtained for hour 5. The value of this minimum cost of energy
142 K. Sooraj Narayan and Ashwani Kumar / Energy Procedia 90 (2016) 137 144

losses for hour is using 4719.639$ DLF and 4727.551$ using PLF. The corresponding real and reactive power losses
are 58.77682$ and 46.03493$ respectively.
Summer without wind integration (DLF) Summer with wind integration (DLF)
Winter without wind integration (DLF) Winter with wind integration (DLF)
20000 Summer without wind integration (PLF) Summer with wind integration (PLF)
Cost of energy loss ($)

15000

10000

5000

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time (Hours)
Fig. 2. Cost of energy losses ($) of 33 bus RDS

Summer without wind integration (DLF) Summer with wind integration (DLF)
Winter without wind integration (DLF) Winter with wind integration (DLF)
Summer without wind integration (PLF) Summer with wind integration (PLF)
250

200
TPL (kW)

150

100

50

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time (Hours)

Fig. 3. Total real power loss (TPL) for the 33 bus RDS

The seasonal energy savings are shown graphically in Fig. 4. From Fig. 4, it can be observed that there is a
significant variation in the energy savings of the test system for winter and summer seasons. The maximum energy
savings for summer season is obtained for hour 15. This value amounts to 5207.86$ from PLF calculations and
5207.25$ from DLF calculations. For the same hour, the energy savings in winter is 4013.38$. The maximum
energy savings for winter season is obtained for hour 10. This value amounts to 4780.92$ from PLF calculations and
4779.25$ from DLF calculations. For the same hour, the energy savings in summer is 4529.87$. A minimum energy
savings of 26.33$ is obtained in hour 4 for summer season. The minimum energy savings for winter season is
obtained for hour 6 and the value is 1068.787$.The minimum energy savings for winter is higher than the minimum
energy savings of summer. From Fig. 11, the energy savings are higher in winter from hours 1 to 10 and for hour 20
 K. Sooraj Narayan and Ashwani Kumar / Energy Procedia 90 (2016) 137 144 143

while it is higher in summer from hours 11 to 19 and from hours 21 to 24. The variation in wind speed and the load
patterns influence the net energy savings.
Summer (DLF) Summer (PLF) Winter (DLF) Winter (PLF)
6000

5000
Energy savings($)

4000

3000

2000

1000

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time (Hours)
Fig. 4. Energy loss savings for summer and winter seasons of 33 bus RDS

5. Conclusion

This paper studied the impacts of intermittent wind power penetration on the cost of energy losses and the energy
savings of radial distribution systems. A comparison was made between the results obtained from deterministic and
probabilistic load flow calculations. The effects of probabilistic loads and seasonal wind speeds/load patterns on the
cost of energy losses and the energy savings of the distribution system were also discussed in detail. The results
were analyzed for a standard 33 bus radial distribution test system.
From the studies conducted, the following observations were made:
1. There is a significant reduction in the cost of energy losses of the system with the integration of wind turbines.
This results in increased annual energy savings. The reduction in cost of energy losses is due to the reduction in total
real power loss with wind integration.
2. The cost of energy losses calculated from PLF is higher than that calculated from DLF. Since the PLF
incorporates the realistic randomness of the various uncertainties of the system, it is more accurate than the DLF
method. Hence, the DLF underestimates the power losses and the cost of energy losses compared to the PLF
calculations.
3. There is a notable difference in the cost of energy losses and the energy savings obtained for winter and summer
seasons. The load profile and wind speeds are different for winter and summer seasons and hence the energy savings
also vary accordingly.
4. The single maximum energy savings is higher for summer season compared to winter season. But the single
minimum energy savings is higher for winter season compared to summer season.
There is also a significant improvement in the voltage profile of the system after wind turbine integration for both
winter and summer seasons. This study is important as it provides better insights into the impacts of intermittent
wind speeds and random load variations on the energy savings of distribution systems.
Acknowledgements
This work is carried out under the project grant SR/S3/EECE/0035/2012, SERB, New Delhi. The author
acknowledges DST New Delhi for grant of the project.

References

[1] Dugan, R.C., Mcdermont, T.E. Distributed generation. IEEE Ind.Applicat.Mag.,2012, 8, (2), pp.19-25
[2] Chiradeja, Pathomth, Ramakumar R.. An approach to quantify the technical benefits of distributed generation. IEEE Transactions on Energy
Conversion., 2004, 19, (4), pp.764,773.
144 K. Sooraj Narayan and Ashwani Kumar / Energy Procedia 90 (2016) 137 144

[3] Gil, H.A, Joos, G. Models for Quantifying the Economic Benefits of Distributed Generation. IEEE Transactions on Power Systems, 2008, 23,
(2), pp.327,335.
[4] Siano, P., Ochoa, L.F., Harrison, G.P., Piccolo, A. Assessing the strategic benefits of distributed generation ownership for DNOs. IET
Generation, Transmission & Distribution, 2009, 3, (3), pp.225-236.
[5] Sheen, J., Tsai, M., Wu, S. A benefit analysis for WT allocation in apower distribution system. 2013, Int. J. Energy Conversion and
Management, 68, pp. 305-312
[6] GWEC, Global Wind Energy Statistic:2015 <www.gwec.net> [accessed 14.04.2015]
[7] Takle. E. S, Brown J. M. Note on the use of Weibull statistics to characterize wind speed data,. J. Appl. Meteor., 1977, 17, pp 556-559.
[8] Briceno W. C. Vincente., Caire r., and Hadjsaid N. Probabilistic load flow for voltage assessment in radial distribution systems with wind
power. Int. J. Electr. Power Energy Syst., 2012, 41, pp. 27-33
[9] Hengsritawat V.,Tayjasanant T., Nimpitiwan N. Optimal sizing of photovoltaic distributed generators in a distribution system with
consideration of solar radiation and harmonic distortion. Int. J. Electr. Power Energy Syst., 2012, 39, pp. 36-47
[10] Murthy V. V. S. N., Kumar. A, Mesh distribution system analysis in presence of distributed generation with time varying load model. Int.
Journal of Elect. Power Energy Syst., 2014, 62, pp.836-854,
[11] Injeti S. K., Thunguntla V. K., Shareef. M. Optimal allocation of capacitor banks in radial distribution systems for minimization of real
power loss and maximization of network savings using bio-inspired optimization algorithms. Int. Journal of Elect. Power Energy Syst., 2015,
69, pp. 441-455.
[12] Chanda S. ,et al,. Implementation of non-intrusive energy saving estimation for Volt/VAr control of smart distribution system. Int. Journal of
Elect. Power Syst. Research, 2015, 120, pp. 49-46.
[13] Khodr. H. M., Olsina F. G., De Oliveira-De Jesus P. M., Yusta J. M.. Maximum savings approach for location and sizing of capacitors in
distribution systems. Int. Journal of Elect. Power Energy Syst., 2008, 78, ( 7) , pp. 1192-1203.
[14] El-Fergany A.A. Involvement of cost savings and voltage stability indices in optimal capacitor allocation in radial distribution networks
using artificial bee colony algorithm. Int. Journal of Elect. Power Energy Syst., 2014, 62, ( 7) , pp. 608-616.
[15] De Keulenaer. H., Chapman. D., Fassbinder. S. The scope for energy saving in the EU through the use of energy-efficient electricity
distribution transformers. Electricity Distribution, 2001. Part 1: Contributions. CIRED. 16th International Conference and Exhibition on (IEE
Conf. Publ No. 482) , 2001, 4, pp.5 2001.
[16] Yong-jun Zhang, Meng-lin Wang. Energy saving modification planning for distribution systems. Advances in Power Conversion and Energy
Technologies (APCET), 2012 International Conference on , 2012,1, (5), pp.2-4.
[17] Morteza Aien, Reza Ramezani and S. Mohsen Ghavami, Probabilistic load flow considering wind generation uncertainty,
Engineering,Technology & Applied Science Research.2011,1,pp. 126-132.
[18] W.C Briceno Vincente, R. Caire and N. Hadjsaid, Probabilistic load flow for voltage assessment in radial distribution systems with wind
power ,Int. J. Electr. Power Energy Syst.,2012, 41, pp. 27-33.
[19] Vlachogiannis J.G, Probabilistic constrained load flow considering integration of wind power generation and electric vehicles, IEEE
Transactions on Power Systems,2009,.24(4), pp. 1808-1817.
[20] Alireza Soroudi, Morteza Alen and Mehdi Ehsan, A probabilistic modeling of photo voltaic modules and wind power generation impact on
distribtuion networks, IEEE Systems Journal,2012, 6(2), pp. 254-259.
[21] Rakesh Ranjan & Das, Simple and Efficient Computer Algorithm to Solve Radial Distribution Networks, Electric Power Components and
Systems, 2010,31(1), pp. 95-107.
[22] Dukpa. A., Duggal. I., Venkatesh. B., Chang. L, Optimal participation and risk mitigation of wind generators in an electricity market
Renewable Power Generation, IET , 2010, 4, (2) , pp.165,175.
[23] Zhipeng Liu, Fushuan Wen, Ledwich. G, Optimal Siting and Sizing of Distributed Generators in Distribution Systems Considering
Uncertainties, Power Delivery, IEEE Transactions on , 2011, 26, (4), pp.2541,2551.
[24] Emingolu U. and Hocaoglu M.H, A voltage stability index for radial distribution networks. Universities Power Engineering Conference,
2007, pp.408-413.
[25] Ayodele T. R., Jimoh A. A., Munda J. L., Agee J. T. Wind distribution and capacity factor estimation for wind turbines in the coastal region
of South Africa. Int. J. Energy Conversion and Management., 2012, 64,pp. 614-625.
[26] < http://rredc.nrel.gov/solar/new_data/India/about.html> [accessed 1.04.2015]
[26] Wong. P., Albrecht. P., Allan. R., Billinton. R., Chen. Q., Fong. C. Haddad. S., Li. W. Mukerji. R., Patton. D., Schneider. A., Shahidehpour.
M. Singh. C. The IEEE Reliability Test System-1996. A report prepared by the Reliability Test System Task Force of the Application of
Probability Methods Subcommittee. Power Systems, IEEE Transactions on , 1999, 14, (3), pp.1010-102 .