Development of controller for economic load dispatch by generating units und

International Journal of Electrical Engineering and Technology (IJEET), ISSN 0976 –
6545(Print), ISSN 0976 – 6553(Online) Volume 4, Issue 4, July-August (2013), © IAEME
159
DEVELOPMENT OF CONTROLLER FOR ECONOMIC LOAD DISPATCH
BY GENERATING UNITS UNDER VARYING LOAD DEMANDS
Sanjay Mathur
Ph.D Scholar, Mewar University, Gangrar, Chittorgarh, Rajasthan, India
Shyam K. Joshi
Ph.D Scholar, IIT Delhi, New Delhi, India
G.K. Joshi
Professor,& Head Deptt. of Electrical Engg., MBM Engg. College, JN Vyas University,
Jodhpur, Rajasthan, India
ABSTRACT
The paper presents a simulink model of controller for feeding power to the load by the generator, in a
group of generators according to power demand imposed by the conditions of economic load
dispatch on the generating plant. The knowledge base that correlates throttle opening of governor
with specific power demand has been derived using experience based training of a feed forward
network and the same has been used to operate a proposed feedback controller. The controller
ensure that the power delivered by the generator equals the power demand on a specific generator for
a given load state, while maintaining economic load dispatch. The simulink model of the feedback
controller shows that the power delivered by a generator operating in parallel with other generators is
same as the one provided by ANN trained modal. The work can be extended for developing a real
time controller that enables the generator to supply power equal to power demand determined by the
conditions of economic load dispatch.
Keywords: Feedback Controller, Simulink, Economic Load Dispatch, Feed Forward Network,
Knowledge Base.
INTERNATIONAL JOURNAL OF ELECTRICAL ENGINEERING &
TECHNOLOGY (IJEET)
ISSN 0976 – 6545(Print)
ISSN 0976 – 6553(Online)
Volume 4, Issue 4, July-August (2013), pp. 159-171
© IAEME: www.iaeme.com/ijeet.asp
Journal Impact Factor (2013): 5.5028 (Calculated by GISI)
www.jifactor.com
IJEET
© I A E M E

160
1.1 INTRODUCTION
The Aim of the present work is to develop a controller, which can manage the requisite
amount of fuel supply to a generator so that it supplies the power equal to the power demand
developing upon the generator as a result of conditions of economic load dispatch. A dedicated fuel
supply control is needed for each generator among the group of generators in the plant. For this
purpose the throttle valve / shutter of the governor is coupled with the shaft of turbine feeding
mechanical power to the generator rotor. Higher the load demand larger would be the throttle
/shutter opening and higher would be the fuel supply leading to more power generation to meet the
increased load demand and vice- versa.
The concept of flux control of speed of a separately excited D.C. motor has been used to
control fuel supply that enabled power supply equal to power demand. It is therefore certain that a
specific power demand can be supplied if the field current (If) of the D.C. shunt motor is of specific
value. This is because the field current decides the size to which the throttle should open and
therefore the fuel supply that should be given to the generator.
The knowledge that correlates the specific power demand to the size of field current (If) has
been developed by using the experience of operators. Also the data base of this kind has been
developed using the Artificial neural network working on feed forward network approach. Having
developed the knowledge-base a feedback controller has been developed, where the field current (If
)ref. keeps changing with changing values of power demand on the generator.
In order to develop a real time controller the developed feedback controller has been
converted into a simulink with an intuitively developed transfer function. The simulink based
controller has been given different values of field currents viz (If ) ref and the corresponding power
generated has been estimated. It is found that the power generated agrees with the power demands,
supplying of which could be made possible by using specific field current (If) ref to be given to the
field of a separately excited D.C. motor for control of throttle opening. The time response for field
current (If ) ref = 5A has been plotted that yields power equal to the one provided by the knowledge
base due to ANN.
The paper has been organized in 04 sections. Section I, covers the basic controller model.
Section II deals with development of knowledge base, for setting fuel rate supply for specific power
demand: The ANN Approach. Section III deals with feed back controller model for supplying
specific power demand by setting the specific value of field current (If) ref. Section IV : covers the
development of simulink for establishing the control action for feeding specific power demand.
SECTION-I
2.1 BASIC CONTROLLER MODEL
The basic controller model is given in figure1. It controls the fuel supply rate (α) to ensure
that the generator delivers a specific power demand ‘P’ & helps in maintaining economic load
dispatch.

161
Figure 1. (a). Control scheme for Fuel supply rate to meet given load demand
a
Ndc
Figure 1. (b) Shutter speed control by separately excited D.C. Motor
For this purpose it is necessary to know the size of field current (If ) for enabling the
generator to deliver given power (P) for every value of load demand. i.e. what would be (If ) for
given (P). This knowledge has been obtained by training the ANN with the physical values of power
(P) and the field current (If ) that gives this power. This is because the field current gives the opening
speed of shutter Ndc and therefore the fuel supply rate (α) which in turn decides the power (P) to be
generated. How does the load demand (P) affect the fuel supply rate (α) through the field current (If )
of d.c. motor is given as under
Field Current
If
Vdc
+
-
VTG
Eb
Armature Current
Ra
Nd.c.
Liters/sec Torque
To Load (P)
VTG

162
The idea to control the fuel supply rate (α) follows the following algorithm.
Figure 2. Control strategy for fuel supply rate( α) as per load demand (P)
Thus the controller works to adjust the fuel supply rate (α) in correspondence with the
specific power demand (P) as determined by the conditions of economic load dispatch for every state
of load demand.
Shutter opens larger
If load demand (P) is increased
The speed (N) of generator goes low
(VTG ) goes low
(If ) of the motor goes low
Flux (φ) of the motor goes low
The speed (Ndc) goes higher
Shutter opens smaller
Power supplied by the generator =
power demand on the generator
Shutter settles to specific size
Fuel supply matches with
increased power demand
Fuel supply rate (α) slows down
The generator speed N =synchronous speed Ns
The flux φ goes higher
Speed of DC motor ’Ndc’ goes lower
Fuel supply rate (α) goes higher
Generator Speed (N) goes higher
(VTG) the techogenerator voltage goes higher
The field current (If ) goes higher

163
SECTION-II
3.1 KNOWLEDGE BASE TO TRAIN ANN FOR FUEL SUPPLY RATE (α) THAT GIVES SPECIFIC
POWER GENERATION ‘P’
ANN has been trained to get a knowledge base for field current (If) for given values of load demand
(P). The data base for fuel supply rate (α) to generate power (P) has been obtained by the experience
of working personnel from various thermal power stations.
Table 1: Training data for ANN based on the experience of working personnel of various thermal
stations
Sr. No. ( If )
Expected power
“P” as provided by
ANN
1. 0.05 40
2. 0.1 40.8
3. 0.15 41.6
4. 0.2 42.4
5. 0.25 43.2
6. 0.3 44
7. 0.35 44.8
8. 0.4 45.6
.
.
.
.
.
.
.
.
.
200. 10 199.2
Table 2: Testing data: as provided by ANN after training as in Table 1
Sr.
No.
( If )
Expected power “P”
as provided by ANN
201. 10.05 200
202. 10.1 200.8
203. 10.15 201.6
204. 10.2 202.4
205. 10.25 203.2
206. 10.3 204
207. 10.35 204.8
208. 10.4 205.6
209. 10.45 206.4
210. 10.5 207.2
211. 10.55 208
212. 10.6 208.8
213. 10.65 209.6
214. 10.7 210.4
215. 10.75 211.2
216. 10.8 212

164
217. 10.85 212.8
218. 10.9 213.6
219. 10.95 214.4
220. 11 215.2
221. 11.05 216
222. 11.1 216.8
223. 11.15 217.6
224. 11.2 218.4
225. 11.25 219.2
226. 11.3 220
227. 11.35 220.8
228. 11.4 221.6
229. 11.45 222.4
230. 11.5 223.2
231. 11.55 224
232. 11.6 224.8
233. 11.65 225.6
234. 11.7 226.4
235. 11.75 227.2
236. 11.8 228
237. 11.85 228.8
238. 11.9 229.6
239. 11.95 230.4
240. 12 231.2
241. 12.05 232
242. 12.1 232.8
243. 12.15 233.6
244. 12.2 234.4
245. 12.25 235.2
246. 12.3 236
247. 12.35 236.8
248. 12.4 237.6
249. 12.45 238.4
250. 12.5 239.2
251. 12.55 240
252. 12.6 240.8
253. 12.65 241.6
3.2 DEVELOPMENT OF ANN PLATFORM FOR OBTAINING KNOWLEDGE BASE
3.2.1 Authentication of ANN standards
In order that the ANN formulated on MATLAB works with high degree of confidence it is
checked for its ability of performance, training states, Regression. A normal feed forward network &
its features in the ANN training window are shown in figure 3

165
Figure 2. ANN Training Window
The performance of success of ANN is given in figure 4.
Figure 3. Performance of ANN
The error regarding training, testing and validation converges to its best values which shows
the authoritative confidence in using ANN for certain test results after proper training.
The training states are given in figure 5

166
Figure 4. Training States of ANN
The regression is shown in figure 6
Figure 5. Regression of ANN
SECTION-III
4.1 DEVELOPMENT OF FEEDBACK CONTROLLER FOR FUEL SUPPLY RATE (α )
THAT GIVES SPECIFIC POWER GENERATION ‘P’
In order to develop a controller that enables a generator working among a group of generators to
deliver specific power ‘P’, the fuel supply rate (α) has been controlled by controlling the size of
opening of throttle/shutter of Governor. For this purpose the controller has been given the knowledge
base as developed by ANN. The proposed Feedback controller is given in Figure 7

167
Figure 6. Block diagram of Controller to Power by Fuel supply rate (α) /Field current ( If )
For every new state of load the new power demand is thrown on the output of a generator. It
therefore required new (If ) ref to be set at the input of the controller. This follows the knowledgw
base given by ANN. If due to change in load state Pdemand becomes higher. It is therefore if P demand is
greater than P demand previous than (If)ref shall be greater than (If)ref previous causing ∆I to be larger and the
shutter will open with larger area leading to a higher rate of fuel supply rate ߙ and therefore more
power output G. When the power demand is supplied fully the ∆I=0 and the shutter will be set to
new opening and new fuel supply rate α. This would match with increase power demand. This
procedure is repeat every time the power demand changes occures on the controller .
SECTION-IV
5.1 SIMULINK MODEL OF CONTROLLER AND ITS TESTING
While developing the simulink for controller the transfer function has been developed by
taking T = 0.3 secs. With the justification that despite all the non- linerities the controller operates in
the linear zone. The transfer function for shutter, turbine and generator has been chosen to be
)13.0(
1
+S
each. Also the feedback path transfer function is taken as
5.7
1 intutively. The entire transfer
function has been multiplied to gain K. Thus based on empirical relations the transfer function has
been taken as .
T.F. = )1......(..........
133.133 1
2
1
2
1
3
+++ STTSTS
K
For T1 = 0.3
)2..(..........
133.19.02.03.0
.. 23
+++
=
STSS
K
FT
The MATLAB programme for obtaining the step response of the system is given below
n= [0 0 0 23]
d= [0.30.27 0.9 1.133
step (n,d);
grid on;
title (‘plot of the unit step response of G(s)=([23]/[0.3s^3+0.27s^2+0.9s+1.133])
xlabel (‘Time(secs)’);
ylabel(‘Amplitude’);
If ref
Shutter Turbine Generator
Power (P) / If
Converter
Power (P)
If actual
Fuel supply rate (α) Ns

International Journal of Electrical E
6545(Print), ISSN 0976 – 6553(Online) Volume 4, Issue 4, July
A simulink for the controller has been developed which gives power output P for specific field
current (If) / fuel supply rate (α) as shown in
Figure 7. Simulink for controller to control power by fuel supply
The simulink is tested for every
is given in Fig.e 9
Figure 8.
International Journal of Electrical Engineering and Technology (IJEET), ISSN 0976
6553(Online) Volume 4, Issue 4, July-August (2013), © IAEME
168
shown in figure 8
Simulink for controller to control power by fuel supply rate (α) /Field current ( I
The simulink is tested for every value of field current (If) but only the sample case for I
Figure 8. The time response for If =5A
ngineering and Technology (IJEET), ISSN 0976 –
August (2013), © IAEME
) /Field current ( If )
t only the sample case for If = 5A
Gain
24/23/22

169
Table 3 Shows the power output of the generator for given value of field current (If)
Table 3 The field current (If) & Power output (P) of Generator as a result of controller operation
Sr. No. If Power P (MW) Controller
Controller Gain Co-
efficient (K)
1 6.3 138 24
2 7 149 24
3 7.5 159 24
4 8 168.5 24
5 8.5 180 24
6 9 189 24
7 10 201 23
8 11 220 23
9 12 235 22
10 12.5 240 22
It is found that as the field current (If) is increased the value of K needs to be reduced so that
the controller delivers the desired response as suggested by the knowledge base of ANN. Training &
Testing. The Error between execution of controller and one suggested by ANN is shown in Table 4
Table 4: Error between Execution of Controller and one Suggested by ANN
Sr. No. ( If )
Power output
suggested by the
ANN
Power output of
generator due to
Controller
Error
Controller
Gain
1 6.3 140.1 138 2.1 24
2 7 152 149 3 24
3 7.5 159.2 159 0.2 24
4 8 167.2 168.5 -1.3 24
5 8.5 175 180 -5 24
6 9 183 189 -6 24
7 10 198.2 201 -2.8 23
8 11 215.2 220 -4.8 23
9 12 231.2 235 -4.8 22
10 12.5 239.25 240 -0.75 22
The error has been plotted in Figure 10
Figure 9. Plot of error between Controller and ANN
The error between controlling power of controller and ANN-Knowledge base is within 5%.
Hence the design can be extended for developing real time controller.
-50
0
50
100
150
200
250
300
( If ). 6.3 7 7.5 8 8.5 9 10 11 12
ANN2
Controller
Error

170
CONCLUSION
In order that the generator delivers specific power demand (P), the generator needs to be
fueled with specific fuel supply rate (α). It therefore needs to develop a controller which does this
work. For the purpose even if the controller is developed it cannot work unless the proper knowledge
base is developed for opening of throttle providing fuel supply rate (α) that gives the desired power
generation (P). The present work has contributed the development of (a) A knowledge base for
operating a controller (b) Basic model of controller (c) Feedback controller and (d) A simulink
model of controller.
This has been found that the throttle opening as decided by the field current (If) for separately
excited D.C. motor enables, the generator to generate power ( P) as per load demand posed on the
generator to implement economic load dispatch for every state of load. The results for step response
of the simulink model of controller has been found for varying values of field current (If). However
the values of increasing If requires the value of K to reduce. The error between controller output and
ANN – Knowledge base has been found to be within 5%.
FUTURE SCOPE
It is possible to extend the work for developing a real time controller for implementating
economic dispatch.
REFERENCES
[1] Park, J.H. ; Kim, Y.S. ; Eom, I.K. ; Lee, K.Y. “Economic load dispatch for piecewise
quadratic cost function using Hopfield neural network”, IEEE Transactions on Power
Systems, Volume: 8 , Issue: 3 , 1993 , Page(s): 1030 - 1038
[2] Scalero, R.S. ; Grumman Melbourne Syst., FL, USA ; Tepedelenlioglu, N, “A fast new
algorithm for training feed forward neural networks”, Signal Processing, IEEE
Transactions, Vol. 40 , Issue- 1, Jan 1992, Page No.
[3] F. N. Lee, A. M. Breipohl, "Reserve constrained economic dispatch with prohibited operating
zones", IEEE Trans. Power Syst., vol.8, no.1, pp.246-254, 1993.
[4] D. C. Walters, G. B. Sheble, "Genetic algorithm solution of economic dispatch with valve
point loading", IEEE Trans. Power Syst., vol.8, no.3, pp.1325-1332, 1993.
[5] R. H. Liang, "A Neural-based redispatch approach to dynamic generation allocation", IEEE
Trans. Power Syst., vol.14, no. 4, pp.388-1393. 1999.
[6] P. H. Chen, H. C. Chang, "Large-scale economic dispatch by genetic algorithm", IEEE Trans.
Power Syst., vol. 10, no.4, pp.1919-1926, 1995.
[7] D.C.Walters, G.B.Sheble. Genetic algorithm solution of economic dispatch with valve point
loading, IEEE Trans. Power Syst, 1993,8(3):1325-1332
[8] Gaing Zl. Particle swarm optimization to solving the economic dispatch considering the
generator constraints. IEEE Transactions on power systems, 2003, 18(3):1187-1195
[9] Simon Haykin, “Neural Networks A Comprehensive Foundation”, Pearson Prentice Hall
Publication 2nd edition, ISBN 978-81-7758-852-1, pp. 23-26,665-67,755-57
[10] LiMin Fu, “Neural Networks in Computer Intelligence”, McGraw Hill Education Pvt ltd.,
Thirteenth reprint 2010, ISBN-13: 978-0-07-053282-3, ISBN-10: 0-07-053282-6, pp. 18- 19,
8-9.
[11] T Yalcinoz, H Altun, U Hasan, “Environmentally constrained economic dispatch via neural
networks”, International Conference on Electrical and Electronics Eng. Eleco 99, 176-180

171
[12] J. Kumar Jayant, and Gerald B. Sheblé, "Clamped State Solution of Artificial Neural Network
for Real-Time Economic Dispatch," IEEE Transactions on Power Systems, vol. 10, no. 2, May
1995, pp. 925-931.(7)
[13] Bharathkumar S, Arul Vineeth A D, Ashokkumar K and Vijay Anand K, “Multi Objective
Economic Load Dispatch Using Hybrid Fuzzy, Bacterial Foraging-Nelder–Mead Algorithm”
International Journal of Electrical Engineering & Technology (IJEET), Volume 4, Issue 3,
2013, pp. 43 - 52, ISSN Print : 0976-6545, ISSN Online: 0976-6553.
[14] Vijay Kumar, Dr.Jagdev Singh, Dr.Yaduvir Singh and Dr.Sanjay Sood, “Design &
Development of Genetic Algorithms for Economic Load Dispatch of Thermal Generating
Units”, International Journal of Computer Engineering & Technology (IJCET), Volume 3,
Issue 1, 2012, pp. 59 - 75, ISSN Print: 0976 – 6367, ISSN Online: 0976 – 6375.
AUTHORS PROFILE
Sanjay Mathur did his B.E. in Electrical Engineering from Amravati University in
1998 and M.E. from M.B.M Engg. College Jodhpur. He has worked as Asstt. Prof in
the Deptt. of electrical Engg at M.E.C.R.C., Jodhpur, Rajasthan, India then worked as
associate professor at Techno India NJR Institute of Technology, Udaipur. Currently
he is Ph.D scholar at Mewar University, Gangrar, Chittorgarh, Rajasthan, India. His
area of interests are Circuit Analysis, Economic Operation of Generators, Artificial Intelligence,
Programming languages and Electrical Machines. He has authored a book titled “Concepts of C”. He
is also technical consultant of Techlab Instruments.
Shyam K Joshi is currently a pursuing Ph.D from Deptt of Electrical Engg, IIT
Dehi He has obtained M.E (Hons.) in Electrical Engg. with specialization in Control
Systems & B.E . (Hons) in Electronics & Communication Engg. Game Theory,
Biological Neural Network , Networked Dynamical Systems, happens to be his ares
of research interest. Till date he has around 12 publications in various International
Journals , International conferences and Seminars. He is Member of International Association of
Computer Science & Information Technology – Singapore.
G K Joshi did his B.E., M.E. and Ph.D. in Electrical Engineering from M.B.M.
Engineering college Jodhpur, Jai Narayan Vyas University, Jodhpur. He has worked
till now as a lecturer, Sr. lecturer, reader, professor and Principal of Engineering
College I.E.T. Alwar. Presently he is head deptt. Of electrical engineering MBM
Engineering college JNVU Jodhpur. He has guided 03 Ph.D, 23 M.E. dissertations, 30
M.E. seminars, 50 technical papers in national, international conferences and journals. Prof. Joshi is
a technical paper reviewer of Institution of Engineers (I). He is a member editorial board of IJCEE,
International Journal for Computer & Electrical Engineering. He is a fellow of Institution of
Engineers (I). He is a life member of ISTE. He has completed many projects under U.G.C. and
AICTE grants and established a high voltage lab of 400KV standard with non-destructive testing
facilities. His area of research is residual life estimation of dielectrics, applications of soft computing
viz. fuzzy, neuro, GA, evolutionary algorithm to practical problems. His subjects of interest are high
voltage engineering, pattern recognition, instrumentation, power systems and electrical machines. He
is presently guiding 6 Ph.D scholars and 4 M.E. students dissertations. He has organized many
international conferences and has been a key note speaker in several international conferences. His
keynote address on estimation of residual useful life of dielectrics using partial discharges” was rated
excellent in the International conference on signal Acquisition and Processing (ICSAP-2011) held at
Singapore on 26-28 Feb. 2011.

Development of controller for economic load dispatch by generating units und

More Related Content

What's hot (19)

Similar to Development of controller for economic load dispatch by generating units und (20)

More from IAEME Publication (20)

Recently uploaded (20)

Development of controller for economic load dispatch by generating units und