Optimization of Unit Commitment Problem using Classical Soft Computing Technique (PSO)

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 04 Issue: 11 | Nov -2017 www.irjet.net p-ISSN: 2395-0072
© 2017, IRJET | Impact Factor value: 6.171 | ISO 9001:2008 Certified Journal | Page 814
OPTIMIZATION OF UNIT COMMITMENT PROBLEM USING CLASSICAL
SOFT COMPUTING TECHNIQUE (PSO)
Sanjeev Kumar1, Harkamal Deep Singh2
1M.Tech. Research Scholar, Department of EEE, IKGPTU University, Punjab
2Assistant professor, Department of EEE, IKGPTU University, Punjab
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract: In electrical power system network of transmission
and distribution, unit commitment is a complicated decision-
making process, which link to the arrangement of generators
over a desire of time periods to satisfy power system load
demand (industrial and agriculture), operational constraints
and system reliability. A classical soft computing (particle
swarm optimization) is a technique used to apply for the
search space of a given problem to discover out the
parameters required to max. or min. a particular objective.
This research paper presents the way out to short term (one
day) unit commitment of thermal electrical power System
using PSO Algorithm.
Keywords: Unit Commitment problem (UCP), ParticleSwarm
Optimization (PSO).
I. INTRODUCTION
The unit commitment problem finds out hourly start-upand
shut down schedule as well as power output for the
generating units over an assured time period. The
optimization schedule of units minimizes the total
operational cost while satisfying all system constraints and
load demand of generating units. In a Unit Commitment
Problem, the main aim is to get the minimum total operating
cost by a accurate scheduling of the units ON/OFF status of
the generators subject to the power system and physical
constraints. For a short-term (one day) unit commitment
problem such as daily or hourly arrangement of generators,
the units operator needs to run the model in real-time. The
operator should have instant right to use to information
concerning which generators should be operated when
emergency situations came up or how to-do list around
planned maintenance of generating units. Modern Soft
Computing Techniques Particle Swarm Optimization is
applied to solve the unit commitment problem.
II. PROBLEM FORMULATION
Unit commitment is a multifarious decision making process
because of the many constraints that mustnot bedesecrated
when finding optimal or close to optimal commitment
schedules. Mathematically, the UnitCommitmentProblemis
a mixed-integer, non-linear, combinatorial optimization
problem. The optimal solution of above complex UCP in
power system can be obtained by classical soft computing
global search techniques. The objective function of the short
term thermal Unit Commitment Problem is combination of
the fuel cost, start-up cost and shut-down cost of the
generating units andmathematicallycanbeexpressedas[1]:
( 1) ( 1)
1 1
[ ( )* *(1 )* *(1 )* ]
H NG
NH i ih ih ih i h ih ih ih i h
h i
Cost FC P U STUC U U SDC U U 
 
     (1)
Where,
NHCost
is the total operating cost over the scheduled
horizon
( )i ihFC P
is the fuel cost function of units
( 1)i hU 
is the ON/OFF status of ith unit at
( 1)
th
h 
hour.
ihU
is the ON/OFF status of ith unit at hth hour.
U is the decision matrix of the ihU
variable. for i=1,2,3,........NG.
ihP
is the generation output of ith unit at hth hour.
ihSTUC
is the start-up cost of the ith generating unit at hth
hour.
ihSDC
is the shut-down cost of the ith generating unit at the
hth hour.
NG is the number of thermal generating units
{0,1}ihU  and ( 1) {0,1}i hU  
H is the number of hours in the study of time horizon.
(for Short-Term unit Commitment, H is generally takenas8-
12 Hours or one day. For general unit commitment
scheduling H is taken as 24 hours and for long term unit
commitment, Time horizon H may be taken as one week ,
one month, three month, six month or one year duration.
(a) Fuel Cost, ( )i ihFC P
The fuel cost function ofthethermal unit
( )i ihFC P
isexpressed
as a quadratic equation:
2
1
( ) ( ) $ / .
NG
ih i ih i ih i
i
FC P a P bP c hrs

   (2)

Where, ia ($/MW2h), ib ($/MWh) and ic ($/h) are fuel
consumption coefficients of ith unit.
(b) Start-up cost, ihSTUC
Startup cost is warmth-dependent. Startup cost is the cost
concerned in bringing the thermal generators unit online.
Startup cost is expressed as a function of the no. of hours the
generating units has been shut down. Mathematically, the
start-up cost can be represented as a step function:
, )
, )
i i i ii
ih
i i i i
HSC if MDT DT MDT CSH
STUC
CSC if DT MDT CSH
    
  
   
where, DTi is shut down duration, MDTi is Minimum down
time, HSCi is Hot start up cost, CSCi is Cold start up cost and
CSHi is Cold start hour of ith unit.
(c) Shut down cost , ihSDC
Shut down costs are defined as a fixed amount for each
unit/shutdown. The typical value of the shut down cost is
zero in the standard systems. This cost is considered as a
fixed cost.
III. PARTICLE SWARM OPTIMIZATION
PSO is a swarm-based intelligence algorithm subjective by
the social behavior of animals such as a flock of birds finding
a food source or a school of fish protecting them-self from a
marauder. It is classical soft computing technique first
described by James KennedyandRussell C.Eberhartin1995.
They found the idea from two separate concepts, idea of
swarm intelligence based off the surveillance of swarming
habits by certain kinds of animals (such as fish & birds) and
field of evolutionary computation.
A particle in this technique is analogous to a bird or fish
flying through a search (problem) space. The movement of
each particle is co-ordinate by a velocity which has both
magnitude and direction. Each particle (unit)positionatany
instance of time is influenced by its best position and the
position of the best particle in a problem space. The
performance of a constituent part is measured by a fitness
value, which is problem precise.
IV. MATHEMATICS INVOLVED IN PSO
The Particle Swarm Optimization algorithm works by
separately maintaining a no. ofcandidate(particle)solutions
in the search space. For the period of each iteration of the
algorithm, each particle solution is considered by the
objective function being optimized, determining fitness of
that solution. Each candidates solution can be consideration
of as particle ‘flying’ through the fitness landscape finding
the max./min. of the objective function. At the start, the PSO
The PSO algorithm just use to calculate its candidate
solutions, and operates upon the resultant fitness values.
Each particle sustains its position, collected of the candidate
solution and its calculated fitness, and its velocity. In
addition, it considered the bestfitnessvalueithascompleted
thus far during the operation of the algorithm, referred to as
the individual best fitness, and candidate solution that
achieved this fitness, referred to as the individual best
position or individual best candidate solution. At last, the
algorithm keeps the best fitness value achieved among all
particles in the swarm, called the global best fitness and
candidate solution that achieved thisfitnesscalledthe global
best candidate solution or global best position. Fitness
evaluation is performed by supplying the candidatesolution
to the objective function. Individual and global bestfitnesses
and positions are updated by comparing the recently
evaluated fitnesses against the previous individual and
global best fitnesses, and replacing the best fitnesses and
positions as needed. The position and velocity updatestep is
responsible for optimization capability of algorithm. The
velocity of each particle in swarm is updated using the
following equation:
   
    1
1 1 2 2* * * 1,2... ; 1,2...
u u u u u u
i ij ij ij J ijV w V C R Pb P C R G P i NP j NG

       (3)
1 1u u u
ij ij iP P V 
  (4)
P is the current position of
th
j member of
th
i particle at
th
u iteration
1 2,C C are the acceleration constants
w is the weighing function or inertia weight factor
NP is the number of particles in a group
NG is the number of members in a particle
1 2,R R is random number between 0 and 1
The velocity is generally limited to a certainmaximumvalue.
PSO using Eq.(3) is called the gbest model. The particles in
the swarm are accelerated to new positions by adding new
Velocities to their present positions. The new velocities are
calculated using Eq.(5) and new positionsoftheparticlesare
updated using Eq. (6).
    1 1 2 2* 1,2... ; 1,2...new best best
i ij ij ij J ijV w V C R Pb P C R G P i NP j NG       (5)
new new
ij ij iP P V  (6)
Suitable selection of inertia weight ‘ω’ is used to provide a
balance between global and local explorations, which
requires less iterations, on an average, to find a sufficiently
optimal solutions.

The inertia weight w is set according to the following
equation,
max min
max
max
*
W W
W W ITER
ITER
 
   
 
(7)
Where
max
IT is themaximumnumberofiterations(generation)and
IT is the current number of iterations.
The maximum and minimum velocity limit in the
th
j dimension is computed as:
max min max min
max minj j j j
j j
P P P P
V and V
 
 
 
(8)
Where is the chosen number of intervals in
the
th
j dimension.
PSO Algorithm and Flow Chart: figure 1
The PSO algorithm have just three steps, which are repeated
in anticipation of some stopping condition is meet up.
1. Evaluate the fitness of each particle
2. Update individual and global best fitness and positions
3. Update velocity and position of each particle
V. FLOW CHART OF PROPOSED PSO ALGORITHM
Figure-1: Flow chart of proposed PSO.
VI. ALGORITHM FOR STUCP USING PSO
The search procedure for calculating the optimal generation
quantity of each unit is summarized as follows:
1. In the ELD problems the number of online
generating units is the 'dimension' of this problem.
The particles are randomly generated between the
maximum and the minimum operating limits of the
generators and represented using Eq. (3).
2. To each individual of the population calculate the
dependent unit output from the power balance.
3. Calculate the evaluation value of each particle giP in
the population using the evaluation function.
4. Compare each particle's evaluation value with its
pbest . The best evaluation value among them pbest
is identified as gbest .
5. Modify the Velocity of each particle by using the
Equation (5)
6. Check the velocity constraints of the members of
each particle from the following conditions :
1 1 max
min
1 1
min min
max max
,
,
0.5
0.5
r r
ij J ij j
r r
ij J ij j
j j
j j
f V V then V V
if V V then V V
where V P
where V P
 
 
 
 
 
 
7. Modify the position of each particle using the Eq.
(4). 1u
ijP 
must satisfytheconstraints, 1u
ijP 
must
be modified towards the nearest margin
of the feasible solution.
7. If the evaluation value of each particle is betterthan
previous pbest , the current value is set to be pbest .
If the best pbest is better than gbest , the best pbest
is set to be gbest .
8. If the number of iterations reaches the maximum,
then go to step 10. Otherwise, goto step 2.
10. The individual that generates the latest gbest is the
optimal generation power of each unit with the
minimum total generation cost.
VII. TEST SYSTEMS, RESULTS AND DISCUSSION:
In order to show the effectiveness of the Proposed PSO
Algorithm for STUCP, two different types of test systems
have been taken into consideration:
1. The first test system consists of SixGeneratingunits
has been taken from IEEE 30-Bus System with a
time varying load demand for one day.
2. The second test system consists of Ten Generating
Units Model and load data for one day.
The corresponding results has been obtained using Particle
Swarm optimization Technique using Population Size=50

and Maximum Iteration=50. The Flow chart for Single Area
Unit Commitment Problem using PSO is shown in Figure-1.
The MATLAB Simulation software7.12.0(R2011a)isused to
obtain the corresponding results.
Table-I: IEEE 30 bus system characteristics 6 Unit Model
Table-II: Load Demand data for 6 Unit Model
Table-III: Results for 30-Bus System 6 unit Using PSO
UNITS Pmax Pmin A
Rs
B
Rs.
C
Rs
MUi MDi Hcost
Rs
Ccost
Rs
Chour IniState
Unit1 200 50 0.00375 2 0 1 1 70 176 2 1
Unit2 80 20 0.0175 1.7 0 2 2 74 187 1 -3
Unit3 50 15 0.0625 1 0 1 1 50 113 1 -2
Unit4 35 10 0.00834 3.25 0 1 2 110 267 1 -3
Unit5 30 10 0.025 3 0 2 1 72 180 1 -2
Unit6 40 12 0.025 3 0 1 1 40 113 1 -2
Load Demand (MW) U1 U2 U3 U4 U5 U6
166 166 0 0 0 0 0
196
146 0 50 0 0 0
229
167 0 50 0 0 12
267
137 80 50 0 0 0
283.4
153 80 50 0 0 0
272
142 80 50 0 0 0
246
166 80 0 0 0 0
213
133 80 0 0 0 0

Table-IV: Generating Unit characteristics-10 Unit Model
192
112 80 0 0 0 0
161
161 0 0 0 0 0
147
147 0 0 0 0 0
160
160 0 0 0 0 0
170
170 0 0 0 0 0
185
105 80 0 0 0 0
208
128 80 0 0 0 0
232
152 80 0 0 0 0
246
166 80 0 0 0 0
241
161 80 0 0 0 0
236
156 80 0 0 0 0
225
145 80 0 0 0 0
204
124 80 0 0 0 0
182
102 80 0 0 0 0
161
161 0 0 0 0 0
131
131 0 0 0 0 0
TOTAL COST
13423
($)
UNITS Pmax Pmin A
Rs
B
Rs.
C
Rs
MUi MDi Hcost
Rs
Ccost
Rs
Chour IniState
Unit1 455 150 1000 16.19 0.00048 8 8 4500 9000 5 8
Unit2 455 150 970 17.26 0.00031 8 8 5000 10000 5 8
Unit3 130 20 700 16.6 0.002 5 5 550 1100 4 -5
Unit4 130 20 680 16.5 0.00211 5 5 560 1120 4 -5
Unit5 162 25 450 19.7 0.00398 6 6 900 1800 4 -6
Unit6 80 20 370 22.26 0.00712 3 3 170 340 2 -3
Unit7 85 25 480 27.74 0.00079 3 3 260 520 2 -3
Unit8 55 10 660 25.92 0.00413 1 1 30 60 0 -1
Unit9 55 10 665 27.27 0.00222 1 1 30 60 0 -1
Unit10 55 10 670 27.79 0.00173 1 1 30 60 0 -1

Table-V: Load Demand data for 10 Unit Model
Table-VI: Results of 10 unit System Using PSO
Load Demand (MW) U1 U2 U3 U4 U5 U6 U7 U8 U9 U10
700 455 245 0 0 0 0 0 0 0 0
750 455 295 0 0 0 0 0 0 0 0
850 455 370 0 0 0 0 25 0 0 0
950 455 450 0 0 0 20 25 0 0 0
1000 455 370 0 130 0 20 25 0 0 0
1100 255 455 0 130 0 20 25 0 0 0
1150 455 455 0 130 40 20 0 0 0 0
1200 455 410 130 130 25 0 0 0 0 0
1300 455 455 130 130 30 0 0 0 0 0
1400 455 355 130 130 110 0 0 10 10 0
1450 455 355 130 130 162 33 25 10 10 0
1500 455 355 130 130 162 73 25 10 10 0
1400 455 355 130 130 162 80 25 43 10 10
1300 455 355 130 130 162 33 25 10 0 0
1200 455 455 130 130 30 0 0 0 0 0
1050 455 310 130 130 25 0 0 0 0 0
1000 455 260 130 130 25 0 0 0 0 0
1100 455 335 130 130 25 0 25 0 0 0
1200 455 415 130 130 25 20 25 0 0 0
1400 455 455 130 130 162 33 25 10 0 0
1300 455 455 130 130 100 20 0 0 10 0
1100 455 360 130 130 25 0 0 0 0 0
900 455 420 0 0 25 0 0 0 0 0
800 455 345 0 0 0 0 0 0 0 0
TOTAL COST 5,67,330.56 ($)
Table VII: Comparison of Results for ACO and proposed Method
S.NO METHOD TOTAL COST($)
1 FUZZY 571893.00
2 ACO 568815.38
3 PSO 567330.56

VIII. Conclusion
In this paper, researchers have presented the solution of
Short Term Unit CommitmentProblemusingParticleSwarm
Optimization Algorithm. The results for standard IEEE Bus
system consisting of6Generatingunits hasbeensuccessfully
evaluated using PSO. Proposed method result is compared
with ACO technique 10 unit system.
IX. Future Scope
(1) Particle Swarm Optimization is basedontheintelligence.
It can be applied into both scientific research and
engineering purpose use.
(2) Particle Swarm Optimization has no overlapping and
mutation calculation. The search can be carried out by the
velocity of the particle. During the development of several
generations, only the most optimist particle can transmit
information on to the other particles, and the speed of the
researching is very fast.
(3)The calculation in Particle Swarm Optimization is very
simple. Compared with the other developing calculations, it
occupies the bigger optimization ability and it can be
completed easily.
(4) Particle Swarm Optimization adopts the real number
code, and it is decided directly by the solution. The number
of the dimension is equal to the constant of the solution.
REFERENCES
[1] Dimitris N. Simopoulos, Stavroula D. Kavatza, and Costas
D. Vournas‘‘Unit Commitment by an Enhanced Simulated
Annealing Algorithm’’IEEE Transactions on PowerSystems,
Vol.21, No. 1 Year: feb. 2006, pp. 68 – 76.
[2] Shantanu Chakraborty,Tomonobu Senjyu, Atsushi
Yona,Ahmed Yousuf Saber and Toshihisa Funabashi,
“Generation Scheduling of Thermal Units Integrated with
Wind-Battery System Using a Fuzzy Modified Differential
Evolution”, Year 2009 , pp. 1-6
[3] Morteza Eslamian, Seyed Hossein Hosseinian, and
Behrooz Vahidi ‘‘Bacterial Foraging-Based Solution to the
Unit-Commitment Problem ’’ IEEE Transactions on Power
Systems,Vol.24, No.3 Year Aug. 2009, pp. 1478-1488.
[4] Yare Y., Venayagamoorthy G. K., and Saber A. Y.,
“Economic Dispatch of a Differential Evolution Based
Generator Maintenance Schedulingofa Power System”, IEEE
Transactions on Power Systems,Vol. 12 July 2009, pp. 1-8.
[5] Chung, C.Y.,Han Yu and Kit Po Wong‘‘An Advanced
Quantum-Inspired Evolutionary Algorithm for Unit
Commitment’’ IEEE Transactions on PowerSystems,Vol.26,
Year 2011, pp. 847 – 854
[6] Manisha Govardhan , Ranjit Roy “An application of
Differential Evolution technique on Unit Commitment
Problem using Priority List Approach”IEEE Transactions on
Power Systems,Vol 14 .Dec.2012, pp. 858 – 863
[6] Nicola Pedroni, and Yan-Fu Li, ‘A Memetic Evolutionary
Multi-Objective Optimization Method for Environmental
Power Unit Commitment ’’ IEEE Transactions on Power
Systems,Vol 28, Year 2013 ,pp. 2660 – 2669
[7] Ranjit Roy et.al. ‘‘An application of Differential Evolution
technique on Unit Commitment Problem using Priority List
Approach’’ IEEE Transactions on Power Systems, Dec.2013
pp.858-863
[8] Divya Mathur ‘‘New Methodology BBO for Solving
DifferentEconomicDispatch Problems’’International Journal
of Engineering Science and Innovative Technology (IJESIT)
Volume 2, Issue 1, January 2013
[9]Vikram Kumar Kamboj and S.K. Bath ‘‘Mathematical
Formulation of Scalar and Multi-ObjectiveUnitCommitment
Problem Considering System and Physical
Constraints’’NCACCNES, February 2014, pp 245-250
[10] tao ding and rui bo, ‘‘big-m based miqp method for
economic dispatch with disjoint prohibited zonesIEEE
Transactions on power systems, vol. 29, no. 2, march 2014
pp 976-977
[11] M. S. P. Subathra et. al. ‘‘A Hybrid With Cross-Entropy
Method and Sequential Quadratic Programming to Solve
Economic Load Dispatch Problem’’ Year 2014, pp 1 – 14.
[12] David Naso and Ali Davoudi ‘‘A Distributed Auction-
Based Algorithm for the Nonconvex Economic Dispatch
problem’’ IEEE Transactions on industrial informatics, vol.
10, no. 2, may 2014 pp 1124-1132

Optimization of Unit Commitment Problem using Classical Soft Computing Technique (PSO)

Recommended

More Related Content

What's hot (16)

Similar to Optimization of Unit Commitment Problem using Classical Soft Computing Technique (PSO) (20)

More from IRJET Journal (20)

Recently uploaded (20)

Optimization of Unit Commitment Problem using Classical Soft Computing Technique (PSO)