Lecture 2 & 3 Unit Commitment and ELD

Unit Commitment and Economic Load
Dispatch
1
Outline
Characteristics of thermal units
Characteristics of hydro units
Unit commitment
Economic Load Dispatch (ELD)
ELD using Lambda Iteration Technique
ELD Considering Transmission Loss
Penalty Factor
Operating (I/O) Characteristics of thermal

Units
2
fuel Input (h)

or fuel cost (
f)
Output power, Pg (MW)

Pg min
Pg max
The curve can be represented by a polynomial of nth order i. e.
h=a0+a1p+a2p2+a3p3++anpn Mbtu/hr (i)
For mathematical simplicity, it is represented by a quadratic equation (i.e second
order polynomial) as:
h=a0+a1p+a2p2 Mbtu/hr (ii)
And the fuel cost, F = H x fuel cost
f =(a0+a1p+a2p2)x fuel cost Rs./hr (iii)
Incremental Characteristics of thermal Units

3
dH/dP in
Mbtu/MWhr or
dF/dP in
Rs/MWhr
Output power, Pg (MW)
The incremental heat rate curve for 2 nd order i-o characteristics is

represented by a straight line as:
dh/dp=a1+a2p Mbtu/MWhr
(iv) :derivative of equation (ii)
And the incremental fuel cost, dF/dp = Hx fuel cost
df/dP =(a1+a2p)xfuel cost MRs./MWhr (v) :derivative of equation
(iii)
I O Characteristics of hydro plant

4
The power output varies from minimum to

rated load proportionally and above rated
load discharge increases rapidly due to fall
in efficiency.
3
Q in m /s
Pg=9.81Q.H.
H3
H2
H1
Pg min
Pg rated
H3<H2<H
1
Pg in MW
Incremental Water Characteristics

H3 <H2<H1
5
H3
Discharge
rate, dQ/dP
H2
H1
Pg
Pg min
Pg
Pg rated
Incremental water rate is constant for different output power up to rated power and
increases beyond rated power due to fall in efficiency.
Thus, for hydropower plant, theoretically it has no difference to run at full load or
intermediate load economically.
Unit Commitment (UC)

6
UC is solve to answer this question: Given

that there are number of subsets of the
complete set of N generating units that
would satisfy the expected demand, which
of these subsets should be used in order
to provide the minimum operating cost?
The problem of UC is to decide which units
to interconnect over the next T hours,
where T is more commonly 24 or 48 hours,
or even 1 week.
UC example of 3 unit thermal plant

7
Data for UC problem for a 3 unit

thermal system are given below.
Uni Min Ma I-O
t
x
characteristics
MW M
( H in Mbtu/hr)
W
1
150
60
0
H1=510+7.2P1+0.001
42P12
100
40
0
H2=310+7.85P2+0.00
142P22
50
20 H3=78+7.9P3+0.0014
2P32
0
Fuel costs for unit 1,2 and 3 are
MU 1.1, 1.0 and 1.2 respectively.
Total
load
120
0
MW
500
MW
4 pm
4 am
Time o
4 pm day
Typical peak valley load

curve
Possible unit combinations and dispatches

for a load of 550 MW
8
Unit Unit Unit Max Min P1

P2
P3
F1
1
2
3
Gen, Gen, from from from MU
MW MW ELD ELD ELD
F2
MU
F3
MU
Total
cost
MU
Off
Off
Off
Infeasible
Off
Off
On
200
50
Infeasible
Off
On
Off
400
100
Infeasible
Off
On
On
600
150
400
150
On
Off
Off
600
150
550
5389 0
5389
On
Off
On
800
200
500
50
4911 0
586
5497
On
On
Off
1000
250
295
255
3030 2440 0
5471
On
On
On
1200
300
267
233
50
2787 2244 586
5617
3760 1658 5418
The most favorable combination is to turn unit 1 and to turn units 2 and 3 off based on
the least cost operation.
Load,
MW
Optimum combination
Unit 1
Unit 2
Unit 3
1200
On
On
On
1150
On
On
On
1100
On
On
On
1050
On
On
On
1000
On
On
Off
950
On
On
Off
900
On
On
Off
850
On
On
Off
800
On
On
Off
750
On
On
Off
700
On
On
Off
650
On
On
Off
600
On
Off
Off
550
On
Off
Off
Shut down rule for

load ranging fro 500
to 1200 MW:
1.For load more than
1000 MW, run all the
units.
2.For load more than
600 MW to 1000 MW,
run unit 1 and 2.
3.For load less than 600
MW, run unit 1 only.
The unit commitment
schedule with optimum
combination of units for
various loads is as
10
Constraints in UC
problem(1)
Power Balance (we already considered)
Total generation=total load plus losses
Spinning Reserve
Total generation available from units already

synchronized minus present load and losses
Should be arranged such that loss of one or more unit
does not cause too far a drop in system frequency
Must be certain percentage of forecasted peak demand or
equal to capacity of the largest unit (US reliability council)
Must be spread over geographically to avoid bottling of
reserve and to allow various regions to run as islands if
they are to be isolated
11
Constraints in UC
problem(2)
Thermal unit constraints
Hydro constraints
Minimum up time
Minimum down time
Fuel
crew
Hydro thermal coordination
Must run constraints
Certain units are given Must run status for

voltage support or steam supply outside the plant
Economic Load Dispatch (1)

12
A constrained optimization problem

Objective is to minimize the total operating cost ( in
case of thermal plants)
Constraints are the power balance equation and
other operational restrictions as stated
Decision variables are the power dispatches from
each generator
Determines the load to be shared by the to be
committed generators so that the total load plus
losses are met (power balance constraint) at
minimum operating cost and also satisfies other
operational constraints as specified
Economic Load Dispatch (ELD) (2)

13
Assumptions:
The generators to be committed are known before the

solution of ELD.
Considers only the real power generations.
Mathematical Statement of ELD problem:
Minimize total generation cost; F= fi
Subject to a number of constraints.
These constrains are dealt with one by one.
Economic Load Dispatch (ELD) (3)

14
The first constraint could be:

Pi = D, where network loss is neglected
Considering the ELD problem in Lagrangian function is given by:

L = fi + (D-Pi )
L / Pi f i / Pi (0 1) 0
Applying the condition as
L/ pi = 0 & L/ = 0
We get
f i / Pi Incremental cos t
and
NG
L / 0 1( D Pi )
i 1
NG
or , D Pi 0
i 1
It means
fi must be available in terms of Pi
for the minimum generation cost, the incremental cost of individual generators
should be equal
Solution of ELD problem
Assuming all generation cost characteristics can be expressed in

Quadratic equation
fi = ai P
+ bi Pi +ci
Applying ELD conditions as
fi/ Pi = ai Pi + bi = : N equations for N generating units

Pgi D = 0
: one equality constraint
N+1 linear equations and same number of unknowns can be

solved (i.e. P1, P2,,PN and )
Example-1
2
The cost characteristics
of20the
two
are given by;
C1
Pg1
0.01generators
Pg1
C 2 25 Pg 2 0.02 Pg 2
Determines the most economical generation schedules if the

15
demand is 300 MW. Neglect the transmission losses.
Solution:
C1
20 0.02 Pg1
Pg1
C2
25 0.04 Pg 2
Pg 2
Pg1 50 1000 &
Pg 2 25 625
When load is 300 MW
Pg1 Pg 2 300
16
or , 50 1000 25 625 300

or , 75 1925
or , 25.66
Pg1 283.33 MW , Pg 2 16.667 MW
General Expression for Generator Power Output (1)
Assumptions:
all the cost characteristics are represented
by quadratic equations
there are no other constraints apart from
the load balance constraint.
The procedures are as follows;

Let us assume the cost function of ith
generators are given as;
2
fi 0.5ai Pi bi Pi ci
differentiating above equation with

f i to
/ PPi
i ai .Pi bi
respect
General Expression for Generator Power Output

(2)
or , Pi ( bi ) / ai
NG
or , D ( bi ) / ai 0
i 1
bi
or ,
D
ai
i 1
NG
bi
or , D
i 1 a i
i 1 a i
NG
NG
NG
bi
1
or , D
i 1 a i
i 1 a i
NG
bi
D
i 1 a i
or , NG
1
i 1 a i
NG
and ,
Pi
bi
D
i 1 a i
NG
1
i 1 a i
NG
b /a
i
Consideration of generator maximum and

minimum power limit
19
Generators have limits on the minimum and maximum

amount of power they can produce
Typically the minimum limit is not zero.
Because of varying system economics usually many
generators in a system are operated at their maximum
MW limits: Base load generators are at their maximum
limits except during the off-peak.
The necessary condition can be modified by taking
generator max
dFi & min power limit as
dPi
............ forPi min Pi Pi max
Set Pi = Plimit (i.e. Pi = Pimax or Pi =Pimin) for those generators

which crosses their max and min limits.
Example-2
20
a.
b.
Determine the most economic

operating points for these three
units to meet a load of 850 MW
without considering
transmission losses.
Repeat the problem when the
fuel cost for unit 1 decreases to
49.5 MRs/Mbtu
Uni Min
t
MW
Max
MW
I-O
Fuel
Characteristic cost
s
(H in
(MRs)
Mbtu/hr)
600
150
510+7.2P1+
0.00142P22
60.5
100
400
310+7.8P2+
0.00194P22
55
50
200
78+7.97P3+
0.00482P32
55
Lambda iteration method
Suppose
the
cost
characteristics
is
expressed other than quadratic, for example
a cubic or higher order polynomial

exponential function
combination of exponential & polynomial
In this case some iterative method is used

One such method is lambda iteration
method
The idea behind the method is;
If generation is greater than demand , is to be

reduced.
If generation is smaller than demand , is to be
Steps to be followed for iteration

method.
2.
Assume a suitable value of

Calculate P gi
3.
Calculate = P
1.
4.
5.
6.
gi
PD
If is within the tolerance, stop

If > 0 ( P gi > PD) , decrease & go to
step 2
If < 0 ( P gi < PD) , increase & go to
step 2
Lambda Iteration Method- Alternative

Algorithm
23
Pick L and H such that

m
L
P
(
Gi ) PD 0
i 1
H
P
(
Gi ) PD 0
i 1
H L Do
While
M ( H L ) / 2
If
M
H
M
P
(
0
Then
Gi
D
i 1
Else L M
End While
Lambda-Iteration Example
24
Consider a three generator system with

IC1 ( PG1 ) 15 0.02 PG1
$/MWh
IC2 ( PG 2 ) 20 0.01PG 2
$/MWh
IC3 ( PG 3 ) 18 0.025PG 3
$/MWh
and with constraint PG1 PG 2 PG 3 1000 MW

Rewriting generation as a function of , PGi ( ),
we have
15
PG1 ( )
0.02
18
PG 3 ( )
0.025
20
PG 2 ( )
0.01
25
Lambda-Iteration Example,
contd
Pick L so
L
P
(
Gi ) 1000 0 and
i=1
m
H
P
(
Gi ) 1000 0
i=1
Try
20 then
PGi (20) 1000
i 1
15 20 18
1000 670 MW
0.02
0.01 0.025
Try H 30 then
PGi (30) 1000

i 1
1230 MW
26
contd
Pick convergence tolerance 0.05 $/MWh
Then iterate since H L 0.05
M ( H L ) / 2 25
Then since
H
P
(25)
1000
280
we
set
25
Gi
i 1
Since 25 20 0.05
M (25 20) / 2 22.5

m
L
P
(22.5)
1000
195
we
set
22.5
Gi
i 1
27
contd
Continue iterating until H L 0.05
*
The solution value of , , is 23.53 $/MWh

Once * is known we can calculate the PGi
23.53 15
PG1 (23.5)
426 MW
0.02
23.53 20
PG 2 (23.5)
353 MW
0.01
23.53 18
PG 3 (23.5)
221 MW
0.025
28
Lambda-Iteration with Gen

Limits
In the lambda-iteration method the limits are taken
into account when calculating PGi ( ) :
if calculated production for PGi PGi ,max
then set PGi ( ) PGi ,max
if calculated production for PGi PGi ,min
then set PGi ( ) PGi ,min
Lambda-Iteration Gen Limit

Example
29
In the previous three generator example assume

the same cost characteristics but also with limits
0 PG1 300 MW
100 PG 2 500 MW
200 PG 3 600 MW
With limits we get:
m
PGi (20) 1000

i 1
PG1 (20) PG 2 (20) PG 3 (20) 1000
250 100 200 1000
450 MW (compared to 670MW)

m
PGi (30) 1000

i 1
300 500 480 1000 280 MW
Lambda-Iteration Limit
Example,contd
30
Again we continue iterating until the convergence

condition is satisfied.
With limits the final solution of , is 24.43 $/MWh
(compared to 23.53 $/MWh without limits).
Maximum limits will always cause to either increase
or remain the same.
Final solution is:
PG1 (24.43) 300 MW (at maximum limit)
PG 2 (24.43) 443 MW
PG 3 (24.43) 257 MW
Economic Load Dispatch

considering network losses
31
If the network loss is included ,Minimize total cost

Subject to
Pgi= Pd+ Plosses
Where,
Pd
= power demand
Plosses = power loss in the system
Ng
Lagrangian function
for ELD problem Ng
L fi Pgi Pd Ploss Pgi

i 1
i 1
For minimum cost of power generation the partial

derivatives of above equation w.r.t. power generated
by individual generators and partial derivatives w.r.t.
Ploss
L
fi

1 0
i.e.
Pgi Pgi
Pgi
fi
0
Pgi
Henc
e,
Where,
32
Ploss
1 Pgi
fi
Pgi
1
fi
i
Pgi
Ploss
1 Pgi
1
i is arbitrary penalty function for loss
Ploss
1
Pgi
If ith generator contribute more loss than jth

generator then
Ploss
Ploss
Pgi
Or,
Hence,
>
Pgj
1
1
Ploss
Ploss
1
1
Pgi
Pgj
i j
is greater for the generator which contribute

more losses.
The generator which is responsible for more

losses contribute less power than without
33 considering losses.
34
Nature of Loss Penalty

Factor
If the increase in power generation of unit i

results in increase in loss, than incremental
transmission loss is said to be positive. The
penalty factor in this case is greater than unity.
If the increase in power generation of unit i

results in decrease in loss, than incremental
transmission loss is said to be negative. The
penalty factor in this case is less than unity.
ELD Procedure Considering Loss (1)

35
When Cost function is quadratic

Steps
1. Assume suitable value of power generation P gi
2. Calculate
3. Compute
Cost function,
and power loss P loss
and Pgi
ai 2
f i Pgi bi Pg i ci
2
fi
ai Pgi bi
Pgi
Ploss
1
Pgi

fi
i
i ai Pgi bi
Pgi
fi
i
ai Pgi bi
Pgi
where,
ai ai i
bi bi i
Pd Pd Ploss
Pd i

i 1 a
i
Ng
Ng
i 1 ai
bi
Pgi
ai
36

37
4. Apply suitable convergence criteria

fi
i
for i 1 to N g is nearly equal
Pgi
or
Pi
k 1
Pi
or
k 1 k
or
Ploss
k 1
Pl oss
5. If the above condition is satisfied stop otherwise go to

step two
Cost function quadratic considering generator limit
1.
Assume suitable value of generation power Pgi
2.
Calculate
i and power loss P
3.
Compute and Pgi
Cost function,
loss
1
i =
Ploss
1 Pgi
fi
ai
2
Pgi bi Pg i ci
2
fi
ai Pgi bi
Pgi
38

39
fi
i
i ai Pgi bi
Pgi
fi
i
ai Pgi bi
Pgi
where,
ai ai i
bi bi i
Pd Pd Ploss
Pd i

i 1 a
i
Ng
Ng
i 1 ai
bi
Pgi
ai

4. If k generators crosses their minimum
and maximum limit then set the
generation
schedule
of
these
k
generators to their respective minimum
and maximum limiting value,
Now, effective number of generator
K
j
=Ng=Ng-K
Pd Pd Ploss Plim it
j 1
And new demand

40

41
5. Apply suitable convergence criteria

i
fi
for i 1 to N g is nearly equal
Pgi
or
Pi
k 1
Pi
or
k 1 k
or
Ploss
k 1
Pl oss
6. If the above condition is satisfied stop

otherwise go to step two
ELD Procedure Considering Loss using Lambda

Iteration Method (When Cost function is nonquadratic)
1. Assume suitable value of

2. Assumesuitable
value of power generation
i
Pgi
1
i = power
fi
Ploss
3. Calculate and
loss
P
loss
1
i
Pgi
Pgi
Ng
Pgi Pd
4.IfCalculate
Pgi the
by using
the relation
tolerable
range stop
is within
go to step six
5.otherwise
Check convergence
6.If is positive decrease or if
is
42 negative increase and go to step 3.
11

Lecture 2 & 3 Unit Commitment and ELD

Uploaded by

Copyright:

Available Formats

Lecture 2 & 3 Unit Commitment and ELD

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Lecture 2 & 3 Unit Commitment and ELD

Uploaded by

Copyright:

Available Formats

Unit Commitment and Economic Load

Operating (I/O) Characteristics of thermal

fuel Input (h)

Output power, Pg (MW)

Incremental Characteristics of thermal Units

The incremental heat rate curve for 2 nd order i-o characteristics is

I O Characteristics of hydro plant

The power output varies from minimum to

Incremental Water Characteristics

Unit Commitment (UC)

UC is solve to answer this question: Given

UC example of 3 unit thermal plant

Data for UC problem for a 3 unit

MU 1.1, 1.0 and 1.2 respectively.

Typical peak valley load

Possible unit combinations and dispatches

Unit Unit Unit Max Min P1

2787 2244 586

3760 1658 5418

Shut down rule for

Power Balance (we already considered)

Total generation=total load plus losses

Total generation available from units already

Thermal unit constraints

Must run constraints

Certain units are given Must run status for

Economic Load Dispatch (1)

A constrained optimization problem

Economic Load Dispatch (ELD) (2)

The generators to be committed are known before the

Economic Load Dispatch (ELD) (3)

The first constraint could be:

Considering the ELD problem in Lagrangian function is given by:

Solution of ELD problem

Assuming all generation cost characteristics can be expressed in

Applying ELD conditions as

fi/ Pi = ai Pi + bi = : N equations for N generating units

N+1 linear equations and same number of unknowns can be

Determines the most economical generation schedules if the

or , 50 1000 25 625 300

General Expression for Generator Power Output (1)

The procedures are as follows;

differentiating above equation with

General Expression for Generator Power Output

Consideration of generator maximum and

Generators have limits on the minimum and maximum

............ forPi min Pi Pi max

Set Pi = Plimit (i.e. Pi = Pimax or Pi =Pimin) for those generators

Determine the most economic

Lambda iteration method

a cubic or higher order polynomial

In this case some iterative method is used

If generation is greater than demand , is to be

Steps to be followed for iteration

Assume a suitable value of

If is within the tolerance, stop

Lambda Iteration Method- Alternative

Pick L and H such that

Consider a three generator system with