Académique Documents
Professionnel Documents
Culture Documents
Pimin unit i minimum generation output limit Pit is calculated by the running units at time-
step t according to equal loss incremental rate
Mi start up and down times limitation principle and met
TOi minimum uptime of ith unit dF1 t dF 2 t dF it
= =. .. . .= =
TSi minimum uptime of ith unit dP1 t dP2 t dPit λ (5)
t=1, 2…24, i=1,2,..G
TOi duration during which unit is
continuously on
2) Spinning Reserve
TSi duration during which unit is
continuously off If spinning reserve needs to be more
than 7% of the total load at each time interval,
B.Objective Function it must be met
G
In this paper it is assumed that the
schedule periods are 24 hours and divide into ∑ ¿¿U P it imax ] ≥ 1.07 PDt , t=1,2,..,24 (6)
i=1
24 time-steps. The total cost is the sum of the
running cost and start up cost for all of the
units over the whole scheduling periods. 3) Unit Generation Output Limitation
Accordingly, overall objective function of the
UC problem is Pimin ≤ Pit ≤ Pimax ,
t=1,2,…,24; i=1,2,…,G (7)
MinF (Uit ,P it ) =
24 G 4) Start Up- and down times Limitation
∑ ∑ ¿ ¿¿ U F (P ) + U (1- U
it i it it it-1 )Si ] (1)
24
t −1 i=1
∑|U it −Uit−1|
Generally, the running cost, per unit
t=1 ≤ Mi i=1,2,…,G (8)
in any given time interval is a function of the
generator power output. The cost function is 5) Minimum Up-and down-time constraints
usually in the form of
Toi ≥ Toi (9)
2
Fi(Pit)=ai Pit +bi Pit+ci (2)
Tsi ≥ Tsi (10)
The generator start up cost depends on
the time the unit has been off prior to start up.
In this paper, time-dependent start up cost is III.PARTICLE SWARM BASED UC
represented as follows: ALGORITHM
Si = Soi + S1i (1- e-T/τi) (3)
As explained, Unit commitment (UC)
is a nonlinear mixed integer optimization
The shut down cost is usually given
problem to schedule the operation of the
a constant value for each unit. In this paper,
generating units at minimum operating cost
while satisfying the demand and other equality Where Ui,t is the binary variable to
and inequality constrains. The UC problem has indicate the on/off state of the unit i at time t.
to determine the on/off state of the generating Ui,t =1 if unit i is committed at time t,
units at each hour of the planning period and otherwise Ui,t=0.
optimally dispatch the load and reserve among
the committed units. UC is the most The overall objective is to minimize
significant optimization task in the operation OCT subject to a number of system and unit
of the power systems. Solving the UC problem constraints. All the generators are assumed to
for large power systems is computationally be connected to the same bus supplying the
expensive. The complexity of the UC total system demand. Therefore, the network
problems grows exponentially to the number constraints are as follows
of generating units.
a. Power Balance Constraint
There are many methods to solve the N
unit commitment but here we are choosing
particle swarm optimization. Here the
∑ pi , t U i ,t =P D (14)
i=1
generating units are behaving like dimension
of the system, in PSO. The position or the b. Spinning Reserve Constraint
value of the generating unit generated
N
randomly between the upper and lower limit.
After that, these values are updating till the ∑ pi , t U i ,t ≥ PD + Rt (15)
i=1
optimization level reached.
A. Problem Formulation
T off
i ,t ≥ MDT i (18)
The objective of the UC problem is to
The initial unit states at the start of the
minimize the total operating costs subjected to
scheduling period must be taken into account.
a set of system and unit constraints.
Where pmin
i & pmaxi are the minimum and
PC i=ai +bi Pi+ c i Pi (11) maximum generation limit of the ith unit, T on
i ,t
& T off
i ,t represents the duration during which
Where ai, bi, ci are the unit cost coefficients.
the ith unit is continuously on and off
respectively and MUTi& MDTi are the
SC =σ +δ {1−e
( )}(12)−T off i
τi minimum up-time and down-time respectively.
i i i
The Particle swarm optimization
Where σ i is the hot start-up cost, δi the cold (PSO) is a population based searching
start-up cost and τi is the cooling time algorithm. This approach simulates the
constant. simplified social system such as fish schooling
and birds flocking. PSO is initialized by a
The total operating costs, OCT for the population of potential solutions called
scheduling period T is the sum of the particles. Each particle flies in the search
production costs and the start-up costs. space with a certain velocity. The particle’s
T N flight is influenced by cognitive and social
OC T =∑ ∑ PC i , t U i ,t + SC i ,t ( 1−U i ,t−1 ) U i ,t (13) information attained during its exploration. It
t=1 i=1 has very few tunable parameters and the
evolutionary process is very simple. It is
capable of providing quality solutions to many
complex power system problems. One such
problem is the unit commitment of thermal 4. Compare each individual’s fitness value
units in the power system. PSO is used to with its pbest. The best fitness value among
minimize the total operating cost by pbest is denoted as gbest.
committing those optimal combinations of the
units which satisfy the constraints and gives 5. Modify the individual velocity vid of each
the minimum cost corresponding to that individual pi as
combination. Our main aim is to minimize the 6. Modify the individual position pi as
operating cost, so we are using the ALM
method for handling equality and in equality v(t) t =1
+C 1∗rand ( )∗( p pbesti− pi(t )) +C2∗rand ( )∗( p gbesti− p
i =v i
constraints. In this problem the up and down
time of the units are not taken into
consideration. p(ti )= p(t−1)
i + p(t)
i (25)
HRS /
1 2 3 4 5 6
UNITS
1 1 1 1 1 1 1
2 1 1 1 1 1 1
3 1 1 1 1 1 1
4 1 1 1 1 1 1
5 0 1 0 1 1 1
6 0 0 1 0 1 0
7 0 0 0 0 0 0
8 0 0 0 0 0 0
9 0 0 0 0 0 0
10 0 0 0 0 0 0
HRS /
7 8 9 10 11 12
UNITS
1 1 1 1 1 1 1
2 1 1 1 1 1 1
3 1 1 1 1 1 1
4 1 1 1 1 1 1
5 0 1 1 0 1 1
6 0 0 1 0 0 0
7 0 0 0 0 0 0
8 0 0 0 0 0 0
Fig 1. Flow chart for solving Unit 9 0 0 0 0 0 0
Commitment using PSO 10 0 0 0 0 0 0
HRS /
19 20 21 22 23 24
UNITS
1 1 1 1 1 1 1
2 1 1 1 1 1 1
3 1 1 1 1 1 1
4 1 1 1 1 1 1
5 1 1 1 0 1 1
6 1 0 0 1 0 0
7 0 0 0 0 0 0
8 0 0 0 0 0 0
9 0 0 0 0 0 0
10 0 0 0 0 0 0
V.CONCLUSION
This paper proposed a Particle Swarm
algorithm. The algorithm considered various