Anda di halaman 1dari 6

Loss Reduction of Power Distribution Network

Using Optimum Size and Location of Distributed


Generation
Adnan Anwar, Student Member, IEEE, and H. R. Pota, Member, IEEE

AbstractDistributed generation (DG) units reduce electric


power losses and hence improve reliability and voltage prole. Determination of appropriate size and location of DG is
important to maximize overall system efciency. In this paper,
an optimization method has been presented to determine the
appropriate size and proper allocation of DG in a distribution
network. Results obtained from this method have been compared
with using the repeated load ow method. A new approach to
perform repeated load ow by using simulation engine openDSS
COM server through Matlab programming is also introduced
here. Both optimized method and repeated load ow based
method have been compared for three IEEE distribution test
systems. This analysis shows that using appropriate size and
location of DG, total power loss in primary distribution system
can be reduced signicantly.
Index TermsDistributed generation, Optimum size, Optimum
location, Power loss, Sensitivity analysis.

I. I NTRODUCTION
The growth of electricity demand is increasing rapidly. DG
is one of the better alternatives to full this ever growing
energy demand. Moreover, it reduces system energy loss,
alleviates transmission congestion, improves voltage prole,
enhances reliability and provides lower operating cost. Because of small size compared with conventional generation
units, DG is more exible to install in terms of investment and
time. As a result, integration of Distributed Energy Resources
(DER) with distribution network offers a promising solution;
therefore, an intensive level of research is needed to understand
the impacts of distributed resources on Distribution System.
Before operating distributed and dispersed generation in power
system, different technical, environmental, commercial and
regulatory issues should be analysed properly. Most signicant
technical barriers are protection, power quality, stability and
islanding operation. However, there are some other issues
which should be analysed before to maximize these technical
benets. From previous studies, it has been seen that different
penetration level and various placement of DER will impact
the distribution system differently [1]. Moreover, improper
DG size and inappropriate allocation of DER may lead to
higher power loss than when there is no dispersed generation
in the system at all [2]. Therefore, detail and exact analysis
method is required to determine the proper location and size
of DG more accurately and precisely. In distribution system,
Adnan Anwar and H. R. Pota are with the School of Engineering and Information Technology (SEIT), The University of New South Wales at Australian
Defence Force Academy (UNSW@ADFA), Canberra, ACT 2600, Australia.
E-mail:Adnan.Anwar@student.adfa.edu.au and h.pota@adfa.edu.au.

DG should be allocated in an optimal way such that it will


reduce system losses and hence improve voltage prole [3].
In our study, we will try to focus on optimum location and
size of DG to decrease total system power loss. In most
of the previous researches of DG sizing and allocation, DG
has been connected with grid directly. Signicant risks are
associated in connecting such equipment directly to utility
distribution system. The insulation level of the machines may
not synchronize with the system. Therefore, direct connection
of DG is often discouraged [4]. In our analysis, we have
connected DG through a three-phase step down transformer.
The organization of this paper is as follows; Denition
of DG is discussed in section II. Section III describes
distribution losses with the variation of DG size and location.
Previous analysis methodology is discussed in section IV.
Objective of our analysis is given in Section V. Our analysis
Methodology is shown in Section VI. In Section VII, test
system and analytical tools have been discussed. Simulation
results and discussion have been presented in section VIII.
Evaluation of estimation performances have been analyzed and
compared in Section IX. Finally, the paper is concluded by
brief remarks in Section X.
II. D EFINITION OF DISTRIBUTED GENERATION AND
RATING OF DG ( SIZE )
Generally, Distributed generation means the electric power
generation within distributed network to full the rapid energy
demand of consumers. However, distributed generation can be
dened in a variety of ways.
1) The Electric Power Research Institute (EPRI) denes
distributed generation as generation from a few kilowatts up to 50 MW [5].
2) International Energy Agency (IEA) denes distributed
generation as generating plant serving a customer on-site
or providing support to a distribution network, connected
to the grid at distributed level voltages [6].
3) The International Conference on large High Voltage
Electric Systems (CIGRE) denes DG as smaller than
50-100 MW [5].
Although there are variations in denitions, however, the
concept is almost same. DG can be treated as small scale
power generation to mitigate the consumer energy demand.
Distributed Generation can come from a variety of sources
and technology. To analysis the DER impacts, different types
of generator groups can be considered [7]. Here, we will

III. D ISTRIBUTION L OSSES W ITH THE VARIATION OF DG


S IZE AND L OCATION
In a distribution system power loss varies with numerous
factors. Real power losses of a distribution system depend on
the resistance of distribution lines, core losses of transformers
and motors. As dielectric and rotational losses are so small
compared with line losses, therefore, only line losses are
considered in this analysis. The complex power Sij from node
i to j and Sji from node j to i are

Sij = Vi Iij

(1)


Vj Iji

(2)

Sji =

Where, Vi and Vj are the voltages at node i and j respectively. The line current Iij which is measured at bus i in the
positive direction of i to j and Iji which is measured at bus j
in the positive direction of j to i. Therefore, power loss in any
line between node i and j can be written as the algebraic sum
of power ows determined from (1) and (2) [8].
SLij = Sij + Sji

(3)

After any converged loadow, power loss in any line can


be calculated using (3) and taking the summation of all line
losses, total power loss of the network can be calculated using
equation (4) where n is the number of lines.
Loss =

n


SL (k)

(4)

k=1

For any distribution system, placement of any DG unit will


change the power loss prole of that system. Actually, in
a distribution network, power loss curve with the variation
of power generation at a particular location is approximately
quadratic function because Line Losses I2 R and I S
considering I is the line current, R is the resistance, and S is
the apparent power owing through the line [9]. Therefore, as
the DG size is increased in any location of a power distribution
network, the total system losses are reduced to a minimum
value. With further increasing of DG, losses again start to
increase. This trend of losses with DG size variation is given
in Fig. 1 for a test case to demonstrate the sizing and location
issues of DG. Here, for DG size PDG2 we get the minimum
power loss which is called optimum DG size for that bus.

Power Loss (kW)

consider induction generator as distributed generation source


for our analysis purpose.
As the technical design of each distribution network is
unique, therefore, it can not be answered what should be
the optimum generation capacity or rating of DG [5]. The
maximum size or rating of DG which can be connected to
a distribution network depends on numerous factors, such as
voltage level within the distribution system, power loss prole
and other technical, environmental, commercial and regulatory
issues. In our paper, we will focus on the technical issues only.
As DG offers lots of benet, the penetration of DG in
distribution system is increasing rapidly. Therefore, DG should
be allocated in an optimal way to maximize the system
efciency.

PL1
PL3

PL2

PDG1

PDG2

PDG3

Size of DG (kW)

Fig. 1. Power loss characteristic of distribution system with DG size variation

Actually, the structure of distribution system is such that


power should ow from the substation to the consumer end and
conductor sizes are also decreased gradually [3]. When a DG
is placed in the network, it is desirable that power should be
consumed within the distribution network and thus improves
power prole. Any size of DG more than the optimum size will
create reverse ow of power towards distribution substation.
Therefore, excessive power ow through small sized conductors towards the transmission area will increase the power loss
in distribution network.
IV. P REVIOUS M ETHODOLOGIES
A good number of research work is going on DG integration
with grid and its safe and reliable operation [3], [10-17].
However, only a few studies have been done on DG sizing
and allocation issue. Different methodologies to determine
optimum location and size have been discussed in different
literatures. The 2/3 rule is often used in capacitor allocation
studies in power distribution network. Similar approach can be
performed in DG allocation to reduce system power loss [10].
In the paper [10] authors have used this analytical method and
rule of thumb for analysing the distribution system which is
radial and has uniformly distributed loads. Rule is simple and
easy to use but it cannot provide the proper solution when
the load distribution type is changed. Moreover, it can not be
applied in meshed network.
In [11], analytical approaches for both radial and networked
distribution systems with different types of load conguration
are given. Here, separate algorithms have been used for radial
and meshed networks. To simplify the analysis, authors have
considered only overhead lines for which uniformly distributed
parameters like R and L per unit length are same along the
feeder. Results obtained from the analysis are very quick; however, one generalized algorithm is expected for both radial and
meshed networks. Besides, in practical distribution system,
conductor sizes are gradually decreased from substation to
load centre, therefore, this analysis procedure would be very
complex when line parameters are not uniformly distributed.
One major limitation of this approach is, they have only solved
the location problem for a xed size of generator but they have
not considered DG sizing issue in their analysis.

Another analytical approach has been proposed based on


exact loss formula in [3]. Authors have considered the loss
coefcients constant. Here they have considered both sizing
and location issues. This process takes only two load ows to
determine the location and size of DG. Although the technique
is very fast; however, this methodology can be applied only if
DG delivers real power [12]. This is one major limitations of
this approach. For load ow, authors have considered NewtonRaphson algorithm. Although Newton-Raphson approach has
an excellent convergence characteristics but in distribution
system because of smaller X/R ratio it can not be decoupled.
Moreover, in distribution network, multi-phase, unbalanced
operation, unbalanced distributed load and dispersed generation makes the Newton-Raphson approach unattractive [13].
For selection optimum size and location of DG, several
genetic algorithms (GA) and fuzzy logic based methods have
been discussed in [14], [15], [16], [17]. Although GA provides
almost near optimum output but they are computationally very
demanding and have a slow convergence [3].
As load ow represents the system states, therefore it can
be used for planning the future expansion of power systems.
We can calculate the system loss from the load ow result
and doing the load ow repeatedly, we can easily tell the
location and size of DG for which we get the minimum
power loss of the system. This method is known as exhaustive
load ow (ELF) method. Although this ELF method gives the
exact answer; however, it needs lots of load ow computation.
Therefore, ELF method needs to be optimized to get accurate
answer and less computational time.
In the previous literatures, researchers have considered
radial distribution system but they have not considered three
phase unbalanced system. As, distribution system is three
phase and unbalanced, therefore, more detailed analysis is
needed based on these type of distribution networks.
V. P ROBLEM F ORMULATION
The main goal of our analysis is to determine optimum
size and location of dispersed generation so that it can reduce
the real power loss in a distribution system. This analysis is
important for efcient power system planning and operation.
Distributed generation not only reduces the power loss of
a system but also improves the voltage prole. However,
inappropriate size and allocation of DG can cause low or overvoltage in the distribution system [18]. Therefore, another goal
of our analysis is to check whether the voltage prole remains
within permissible limit. So, voltage constraint becomes,
Vmin V Vmax

(5)

Distribution step-type voltage regulators are used to maintain


line voltages within predetermined limits; as a result end
user can get a constant voltage output [19]. Impact of DG
on voltage regulation with the presence of voltage regulator
is discussed in [20]. The test systems, which we used for
our analysis, comes with voltage regulators which maintain a
permissible voltage limit throughout the analysis. During our
analysis, we considered Vmin=0.94 pu and Vmax=1.06 pu In
section IX-B, we will show how optimum size and location
of DG impacts on voltage level of the interconnecting bus.

VI. P ROPOSED A NALYSIS M ETHOD


In our analysis, repeated load ow for loss reduction
has been performed in a different approach by integrating
openDSS with Matlab which makes the analysis faster, more
accurate and efcient. Base on sensitivity, a new methodology
has been proposed to calculate optimum size and location
of DG. Finally, these two methods have been analysed and
compared.

A. Optimized Algorithm using Sensitivity


For any distribution system, if DG size is varied from PDG1
to PDG2 and their corresponding change in power loss is
respectively PL1 to PL2 , then the sensitivity factor becomes,
dPL
PL1 PL2
=
dPi
PDG1 PDG2

(6)

In our analysis, Sensitivity factors are evaluated for each


bus using equation (6) and the bus with maximum sensitivity
is identied. Only those buses which have sensitivity factors
close to the maximum value have been considered in our
analysis. Thus solution space is reduced to only a few buses.
After that, for each of these buses, power loss has been
determined using large step size of DG variation and then
a quadratic curve has been formed using these few samples.
The minimum value of the curve represents the optimum size
for that bus and corresponding generation is the optimum DG
size. The bus which is responsible for minimum loss of the
system is the appropriate location for DG allocation. The ow
chart of our analysis methodology is given in Fig. 2.
VII. T EST S YSTEM AND A NALYTICAL T OOLS
Our analysis method has been implemented on three IEEE
distribution test systems. These are IEEE 34 node test feeder,
13 Node test feeder and 123 Node test feeder. The rst one,
IEEE 34 Node test feeder is an actual feeder located in Arizona
and its nominal voltage is 24.9 kV [21]. The second one,
IEEE 13 Node test feeder is small but good for test cases. The
third one, IEEE 123 Node test feeder operates at a nominal
voltage of 4.16 kV and contains overhead, underground line
segments with various phasing, unbalanced loading with all
conguration of loads [21]. Single line diagrams of the test
feeders are shown in the appendix. The basic data for these
three test feeders can be obtained from [22]. For analysis purpose, we have developed several Matlab programs to perform
the tasks 3 to 12 shown in ow chart. Only for three phase
unbalanced load ow we have used Open Distribution System
Simulator (OpenDSS). OpenDSS has an in-process COM
server DLL through which the user can design and execute
custom solution modes and features externally [23]. We have
interfaced openDSS COM server with our Matlab program
and load ow results have been imported from openDSS. Then
Matlab programs have been used to nd out optimum size and
location of DG.

Start

300
280

1. Input basic data of Distribution system

3. Calculate sensitivity of each bus


with a small penetration of DG

Power Loss (kW)

2. Base case load ow

260
240
220
200
180
160

4. Make a list of most sensitive buses

140

5. Select a bus from the list and calculate


power loss for large variation of DG size
6. Continue until loss starts to
increase and record each sample

120
0

no

9. Check
whether all
sensitive buses
have been
analysed?

400

600

800

1000

1200

1400

1600

1800

DG Size (kW)

Fig. 3. Power loss prole with the variation of DG size (for 34 Node test
system)

7. Fit those samples into a quadratic curve

95

Actual
Approximate

90

Power Loss (kW)

8. Find the min point of the curve which


represents the min loss of that bus

200

85

80

75

70

65
0

200

400

600

800

1000

1200

1400

1600

1800

Variation of DG Size (kW)

Fig. 4. Comparison of power loss curve using both methodology at bus


number 76 for 123 bus test system

yes
10. Find the bus which has min power loss
11. Find corresponding DG size
12. Print optimum DG size and location
End
Fig. 2.

Flowchart to identify optimum size and location of DG

result obtained from repeated load ow. For the other two test
systems, IEEE 13 Node test system and IEEE 34 Node test
system, we have obtained optimum DG size and location using
exhaustive search and proposed method. Results obtained from
simulation are shown in table I. For all of the three test
systems, results from our proposed method are quite similar
with repeated load ow approach but our method is quite
faster.
TABLE I
C OMPARISON OF THE RESULTS OF TWO METHODS

VIII. S IMULATION RESULTS AND DISCUSSION


Fig. 3 shows the trend of power loss with the variation of
DG size for each bus of 34 Node test system. This gure
has been obtained using repeated load ow method. In that
analysis procedure, DG which is responsible for minimum
loss at each bus has been identied rst and that is optimum
DG size for that bus; therefore, the bus which has minimum
loss among the whole network is considered as the optimum
location for that network. Although repeated load ow is
inefcient due to its large number of load ow computations
and thus called exhaustive approach but it gives us accurate
results. Using this methodology, optimum location of DG for
123 Node test system is bus 76 and optimum DG size is
1.32 MW. Now, using our methodology, for same test system,
we have found the optimum size of DG is 1.32 MW and
optimum location is bus 76 which is similar with the generated

Test
System
13 Bus
34 Bus
123 Bus

Repeated load ow
Optimum
Optimum
Size (MW)
Location
1.28
Bus 675
1.10
Bus 844
1.32
Bus 76

Proposed Method
Optimum
Optimum
Size (MW)
Location
1.29
Bus 675
1.15
Bus 844
1.32
Bus 76

IX. E VALUATION OF ESTIMATION PERFORMANCES


For 123 Node test system, at bus number 76, power loss
curves using both methods have been shown in Fig. 4. We
considered the output of repeated load ow as the actual case
and output using our method is the approximate one. From the
gure we can say, our methodology has a very good match
with the actual case; moreover, it takes less no of solution
space and time.

A. Active power loss reduction by DG

Voltage before adding DG


Voltage after adding DG

Active power loss reduction (PLR) by DG may be dened


as below:
DG
PLoss PLoss
P LR =
100%
(7)
PLoss
where,
PLoss =Power loss of the system before introducing DG
DG
=Power loss of the system after adding DG
PLoss

Volatage Magnitude ( P.U.)

1.1
1.05
1
0.95
0.9
0.85
0.8

Phase3

Phase2

Phase1

For 123 bus test system, before adding any DG, total system
power loss was 95.3 kW. Adding an additional DG of optimum
size on the optimum location, this loss becomes 65.3 kW.
Therefore, active power loss is reduced by 31.5%. Therefore,
overall system efciency improves. Using optimum size and
location, active power loss reduction for the three systems are
summarized in table II.

Voltage variation at the phases of bus 675

(a) 13 Node test system


Voltage before adding DG
Voltage after adding DG

Voltage Magnitude ( P.U.)

1.1

TABLE II
ACTIVE P OWER LOSS REDUCTION IN DIFFERENT TEST SYSTEMS

1.05
1
0.95
0.9
0.85

Test
System
13 Bus
34 Bus
123 Bus

Power Loss (kW)


Before
After
adding DG adding DG
110.4
85.4
272.8
125.5
95.3
65.3

0.8

PLR

Phase1

22.6%
53.9%
31.5%

Phase2
Voltage variation at three phases of bus 844

Phase3

(b) 34 Node test system


Voltage before adding DG
Volatge after adding DG

Volage Magnitude ( P.U. )

1.1

B. Impact on Voltage prole


Voltage range for static voltage stability varies from country
to country and utility to utility. However, ANSI standard
should be maintained to supply a steady-state voltage on
consumer side of a distribution network [19]. The acceptable
steady state voltage range of systems between 1kV and 132
kV is considered within 6% of the nominal voltage [24]. For
123 bus test system, nominal voltage is 4.16 kV and voltage
base is 4.16 kV, therefore, acceptable voltage should be within
0.94 pu to 1.06 pu For the other two test systems, we get the
same voltage range for steady state voltage stability. In our
analysis, using voltage regulator and allocating optimum size
of DG in optimum location voltages remain within acceptable
limit which is shown in Fig. [5].

1
0.95
0.9
0.85
0.8
Phase1

Phase2
Voltage variation at three phases of bus 76

Phase3

(c) 123 Node test system


Fig. 5.

Voltage prole of different test systems

detail dynamic analysis is needed which will be our future


work.

X. C ONCLUSION
The impact of proper allocation and sizing of DG is
very signicant. Power Loss of distribution system increases
overall system cost and has a major impact in power system
management. From our analysis, we can come to know that,
improper DG size and inappropriate DG allocation may cause
a greater system loss than the loss without DG. DG should
be allocated in those locations where they provide higher
reduction of losses. Although DG is usually consumers
property, nevertheless, it is the interest of utilities and
engineers to determine the appropriate size and location for
safe, reliable and stable operation of the distribution system.
Our analysis method identies this sizing and location issues
accurately and precisely in a faster manner. To understand the
impact of proper allocation and size of DG properly, more

1.05

A PPENDIX

Fig. 6.

IEEE 13 Node Test Feeder

[12]
[13]
[14]
[15]

[16]
Fig. 7.

IEEE 34 Node Test Feeder

[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]

Fig. 8.

of distributed generation sources in power systems, IEEE Transactions


on Power Systems, vol. 19, no. 4, pp. 20682076, nov. 2004.
D. Q. Hung, N. Mithulananthan, and R. Bansal, Analytical expressions
for DG allocation in primary distribution networks, Energy Conversion,
IEEE Transactions on, vol. 25, no. 3, pp. 814 820, sept 2010.
M. Srinivas, Distribution load ows: a brief review, in Power Engineering Society Winter Meeting, IEEE, 2000.
G. Celli and F. Pilo, Optimal distributed generation allocation in MV
distribution networks, in IEEE Power Engineering Society International
Conference on Power Industry Computer Applications, 2001, pp. 8186.
M. Ameli, V. Shokri, and S. Shokri, Using fuzzy logic amp; full search
for distributed generation allocation to reduce losses and improve voltage
prole, in International Conference on Computer Information Systems
and Industrial Management Applications (CISIM), oct 2010, pp. 626
630.
L. de Queiroz and C. Lyra, A genetic approach for loss reduction in
power distribution systems under variable demands, in IEEE Congress
on Evolutionary Computation, 2006, pp. 26912698.
J. Savier and D. Das, Impact of network reconguration on loss
allocation of radial distribution systems, IEEE Transactions on Power
Delivery, vol. 22, no. 4, pp. 24732480, oct 2007.
C. Borges and D. Falcao, Impact of distributed generation allocation
and sizing on reliability, losses and voltage prole, in Power Tech
Conference Proceedings, 2003 IEEE Bologna, vol. 2, june 2003, p. 5.
W. Kersting, Distribution feeder voltage regulation control, IEEE
Transactions on Industry Applications, vol. 46, no. 2, pp. 620626, 2010.
L. Kojovic, Impact of DG on voltage regulation, in Power Engineering
Society Summer Meeting, 2002 IEEE, vol. 1, july 2002, pp. 97102.
Distribution system analysis subcommittee report. [Online]. Available:
http://ewh.ieee.org/soc/pes/dsacom/testfeeders/testfeeders.pdf
Distribution
system
analysis
subcommittee
radial
test
feeders
[online].
[Online].
Available:
http://ewh.ieee.org/soc/pes/dsacom/testfeeders/index.html
OpenDSS manual and reference guide. [Online]. Available:
http://sourceforge.net/projects/electricdss
C. Masters, Voltage rise: the big issue when connecting embedded
generation to long 11 kv overhead lines, Power Engineering Journal,
vol. 16, no. 1, pp. 512, feb 2002.

IEEE 123 Node Test Feeder

R EFERENCES
[1] D. Rizy, F. Li, H. Li, S. Adhikari, and J. Kueck, Properly understanding
the impacts of distributed resources on distribution systems, in Power
and Energy Society General Meeting, IEEE, 2010.
[2] O. T. Mithulananthan, Nadarajah and L. V. Phu, Distributed generator
placement in power distribution system using genetic algorithm to reduce
losses, Thammasat International Journal of Science and Technology,
vol. 9, pp. 5562.
[3] N. Acharya, P. Mahat, and N. Mithulananthan, An analytical approach
for DG allocation in primary distribution network, International Journal
of Electrical Power & Energy Systems, vol. 28, no. 10, pp. 669678,
2006.
[4] R. Arritt and R. Dugan, Distributed generation interconnection transformer and grounding selection, in Power and Energy Society General
Meeting, IEEE, July 2008.
[5] T. Ackermann, G. Andersson, and L. Sder, Distributed generation: a
denition, Electric Power Systems Research, vol. 57, no. 3, pp. 195
204, 2001.
[6] IEA
publication,
distributed
generation
in
liberalized
electricity
market,
2002.
(page-19).
[Online].
Available:
http://www.iea.org/textbase/nppdf/free/2000/distributed2002.pdf
[7] T. McDermott, Behavioral models of DR technologies for feeder-level
analysis, in Power Energy Society General Meeting, IEEE, July 2009.
[8] H. Saadat, Power System Analysis. Tata McGraw-Hill, India, 2002, p.
212.
[9] J. ODonnell, Voltage management of networks with distributed generation, Ph.D. dissertation, The University of Edinburg, 2007.
[10] H. Willis, Analytical methods and rules of thumb for modeling DGdistribution interaction, vol. 3, pp. 16431644, 2000.
[11] C. Wang and M. Nehrir, Analytical approaches for optimal placement

Adnan Anwar was born in Bangladesh in 1987.


He received B.Sc in Electrical and Electronic Engineering from Islamic University of Technology
(IUT), Bangladesh in 2008. In 2009, he joined at
the University of Asia Pacic (UAP) as a lecturer.
He is currently pursuing his Masters by Research
at the University of New South Wales, Australian
Defence Forces Academy, Canberra, Australia. His
research interests include computer aided analysis of
power distribution network, optimization of distribution system, power system dynamics and control,
integration of renewable energy sources in distribution system and smart grids.

Hemanshu R. Pota received the B.E. degree from


SVRCET, Surat, India, in 1979, the M.E. degree
from the IISc, Bangalore, India, in 1981, and the
Ph.D. degree from the University of Newcastle,
NSW, Australia, in 1985, all in electrical engineering. He is currently an Associate Professor at the
University of New South Wales, Australian Defence
Force Academy, Canberra, Australia. He has held
visiting appointments at the University of Delaware;
Iowa State University; Kansas State University; Old
Dominion University; the University of California,
San Diego; and the Centre for AI and Robotics, Bangalore. He has a
continuing interest in the area of power system dynamics and control, exible
structures, and UAVs.