Load Frequency Control
Load Frequency Control
By proper choice of K J' K~ and K3 all the specifications for the system performance can
be satisfied.
Load Frequency Control 213
E 6.1 A \00 MW generator has a regulation parameter R of 5%. By how much will the
turbine power increase if the frequency drops by 0.1 Hz with the reference unchanged.
Solution:
Actual change in frequency = 5% of 50Hz = 0.05 x 50 = 2.5Hz
R = 2.5Hzl100Mw = 0.025 HzlMw
If M = -0.1 Hz, the increase in turbine power
I I
~P= -- M= ---x(-O.I) =4 MW
2 0.025
The turbine power increase = 4MW.
E 6.2 A 100 MW generator with R = 0.02 HzlMW has its frequency fallen by 0.1 Hz. If the
turbine power remains unchanged by how much the reference power setting be changed.
Solution:
The signal to increase the generation is blocked. Thus at the input summing point the reference
power setting must be changed. Such that
~ Pref - ~R M= 0
I 1
(i.e .. ) ~ Pref - R M= 0.025 xO.l = 4 Mw
E 6.3 Two generators with ratings 100 MW and 300 MW operate at 50Hz frequency. The
-s.ystem load increases by 100 MW when both the generators are operating at about half
of their capacity. The frequency then falls to 49.5Hz.
If the generators are to share the increased load in proportion to their ratings
what should be the individual regulations? What should be regulations if expressed in
, per unit Hertz/per unit megawatt?
1
Solution: AP= --M
R
I
~P = --M
2 R2
J
M=0.5 Hz
Power is shared in proportional to their ratings
214 Operation and Control in Power Systems
100
~P = 100 x - = 25 MW
I 400
300
~P = 100 x - = 75 MW
2 400
0.00667 300
R2 = - x- = 0.04 pu HzlPu MW
50 I
Both have the same value, even though based on their individual ratings, they have
different regulation.
E 6.4 Determine the primary load frequency control loop parameters for a control area having
the following data:
Total rated area capacity Pr = 1000 Mw
Normal operating load = 500 MW
Inertia constant H = 4.0 sec
Regulation R = 2.5 Hzlpu MW
Solution:
OP 500MW
Load damping D = Of = 50Hz = 10 MW/Hz
[Here, the load damping is assumed linear and percentage change is assumed to be
the same]
10
In per unit D =- = 0.01 pu MW/Hz
1000
2H 2x4.0
Tp = fsD = 50x 0.01 = 16 sec
1 1
K
p
= -
D
= -O.oI~
- = 100 Hzlpu MW
Load Frequency Control 215
E 6.5 Determine the area frequency response characteristic and the static frequency error for
a system with the following data, when I % load change occurs?
D = 0.0 I pu Mw/Hz
R = 2.5Hzlpu MW
Tp = 16 sec
Kp = 100 Hzlpu MW
Solution:
1
Area frequency response characteristic ~ =D+ R
1
= 0.01 + 2.5 = 0.41 MW/Hz
M=- M =- =0.02439Hz
~ 100 x 0.41
E 6.6 In the example E6.5, the governor is blocked so that it does not change the generation.
In that case what would be the steady state frequency error?
Solution:
When the governor is not acting, the feedback loop is not existing. In such a case R is infinite.
I
~=D+ R =D=O.OI puMW/Hz
M
Hence, ~f = - - =- -0.01 = -I Hz
~ 0.01
Frequency falls by 1 Hz (i.e.) f = 50 - I = 49 Hz.
It may be noted that with the generator acting the frequency from E6.5 is
50 - 0.02439 = 49.9756 Hz
The importance of feedback through governor mechanism can be understood from the
above.
E 6.7 A 100 MVA synchronous generator operates initially at 3000 rpm, 50Hz. A 25MW load
is suddenly applied to the machine and the steam valve to the turbine opens only after
0.5 sec due to the time lag in the generator action. Calculate the frequency to which the
generated voltage drops before the steam flow commences to increase to meet the new
load. The value of the stored energy for the machine is 5kW-sec per KVA of generator
energy. Also calculate the value of H constant for the generator.
216 Operation and Control in Power Systems
Solution:
Stored energy = 5KW per/kVA
(i.e.,) = 500 MW sec pr 100 MVA
Load increase =25 MW
Energy required to supply this load for 0.5 sec = 25 MW sec
Frequency at 500 MW sec stored energy = 50Hz
Frequency fall = M
M i1f 25xO.5 MW
-=-= sec
f 50 500
H constant = stored kinetic energy at rated frequency = 5MW - sec x I OOMV A = 5 sec
machine rating MV A I OOMV A
E 6.S Given the following parameters, obtain the frequency error. Plot it when a step load
disturbance of (i) I % and (ii) 2% occur in the system.
Tp = 22 sec
R = 2.5
Kp= 100
Solution:
Gp
The expression for i1F(s) = --I-~--i1PD(S)
l+-GsGToG p
R
Neglecting the turbine dynamics and governor action (G s GTO :::: 1.0)
k
i1F(S) =
I +STp I +~_~ s
R 1+ STp
Fig. E6.8 Step load frequency error characteristic without supplementary control 1. k = 0.01
=
2. k 0.02 (governor action and turbine dynamics are neglected
E 6.9 Show the effect of governor action and turbine dynamics, if they are not to be neglected
in E6.8 given Ts = 100 msec and TTG = 0.5 sec.
Solution:
For this the exact frequency error is used
_ Ks
LlPD(S) = O.Olpu; G s --'"""-- G TG -- K TG
1 +STs I+STTG
Ks KTG ::::: 1
The response is shown in Fig. E(6.9).
It can be seen that greater the step load change, larger the error and the governor action
and turbine dynamics does not cause any change in the response in the steady state, except for
a transient deviation at the beginning of the distribution.
218 Operation and Control in Power Systems
Fig. E6.9 Step load frequency error characteristic without supplementary control 1. k =0.01
= = =
2. k 0.02 Ts 100 msec and TTG 0.5 sec
Solution:
To create a SIMULINK block diagram presentation select new (model) file from FILE menu.
This provides an untitled blank window for designing and simulating a dynamic system. Copy
different blocks from the simqIink libraries or other previously opened windows into the new
window by depressing the mouse button and dragging.
I. Open the continuous library and drag the transfer function block to the window.
Double click on transfer function to open the dialog box. Enter the numerator and
denominator values (the coefficients in the descending powers of s, if any power of
s is missing, enter zero) of the transfer function .
2. Open the math library and drag the sum block in to the window. Open the sum dialog
box and enter + - under list of signs.
3. From the math library drag the Gain block into the model file right click on the gain
block and click on the Flip option to rotate the gain block by 180degrees.
4. Open the source library and drag the step input block to window. Double click on it
to open its dialog box and set up the step time(step duration), initial and final
values(which will be same) to represent the step input.
5. Open the sink library and drag scope to window to observe the response.
By using the left mouse button, connect all the blocks.
Before starting simulation, set the simulation parameters. Pull down the simulation
dialog box and select parameters. Set the start time, stop time and for a more accurate integration,
set the maximum step size.
In this example the parameters for all the blocks for the system in the fig. are initialized.
Open m-fiIe and enter the parameter values. The following m-fiIe has to be run prior to the
simulation.( of model file)
Open new m-fiIe and enter the parameter values as shown below.
Tg = 0.2;
T t = 0.5;
H= 5;
D = 0.6;
R = 0.05;
%for integral control
Kj = 7;
Save the m-fiIe under parameters and run the file. SIMULINK block diagram and results
for LFC
By using above procedure construct the simulink block diagram for the load frequency
control of isolated power system as shown in Fig. E 6.1O(a).
220 Operation and Control in Power Systems
Step
l/R
'-----------<-K-I4-----------'
Fig. E 6.1 O(a) Block diagram Model of Load Frequency Control (Isolated Power System)
Pull down the file menu and use save as to save the model under AFC. Start the
simulation. Double click on the scope, click on the auto scale, the result is displayed as shown
in Fig. E 6.1 O(b).
Add integral controIIer block for above system and save the file under LFC 1.
Start the simulation. Double click on the scope, click on the auto scale, the result is
displayed as shown in figure below.
E 6.11 A single control area system with the following data experiences a sudden load change
of3%.
Kp= 100
Tp= 25
R = 2 Hzlp.u. MW
with integral control using a gain of 10 obtain the frequency plot with time and show that the
frequency deviation is reduced to zero.
Solution:
Using MATLAB, the solution is obtained and shown in Fig. (E 6.11).
Kj=IO Kp = IOO Tp = 25 R =2 num = Kp/Tp
d l = (I1Tp) + (K/ (R * Tp)) d2 = Kj * K/Tp den = [I d l d 2]
Transfer Fan
Fig. (E 6.11)
222 Operation and Control i!!.!Qwer Systems
0= au ..... (6.87)
au
Where u = L (x, u, t) + At [f(x, U, t)] ..... (6.88)
and A. denotes Lagrange multipliers. Simultaneous solution to eqn. (6.85, 6.86, 6.87)
subject to boundary conditions gives optimal control Uo. But, the equations are non linear and
time varying and hence even for simple systems the solution will be difficult to obtain.
Where Hand Q are real symmetric positive semi definite matries and P is a real symmetric
positive definite matrix. Again t2 is fixed, but X(t 2) is free. For simplicity, omitting the arguments .
..... (6.91 )
Load Frequency Control 223
..... (6.93)
..... (6.94)
..... (6.97)
or ..... (6.98)
where A.B and F are the system, input distribution and disturbance distribution matrices
respectively and X, U and d are the state, control and disturbance vectors respectively.
By a suitable transformation, eqn. (6.24) is transformed into the form.
X' =A'X' +B'U' ..... (6.110)
The LFC problem is then treated as a linear regulator problem and the solution is obtained
by solving algebraic Riccatic equation.
J= f 2"1 ('T~
00
Qx + UT
)
Pu dt ..... (6.111)
o
Where Q and P are the weighting matrices for the state and control vectors and X and u.
Consider the block diagram for the single area system shown in Fig. 6.26.
r---II/R~--------------.....,
~...-+t.F(s)
S ~ F(s) 1
= __ ~F(s) + Kp ~PTG (s) - Kp M>D (s) ..... (6.115)
Tp Tp Tp
In the time domain, using the state variables
. () 1 Kp Kp
x2 t =--x2+-x3-~' d ..... (6.116)
Tp Tp. Tp
Also, from the block diagram
1
i.e., STs ~(s) = -~ ~ (s) + U(s) - - M(s) ..... (6.121)
R
In the time domain
..... (6.122)
Load Frequency Control 227
Putting eqn. (6.113, 6.116, 6.119 and 6.122) in matrix from, the state variable model is
obtained as
0 0 0
XI (t) 0 0
-I Kp
0 0 0 kp
x2 (t) Tp 1p [ x2
Xl
-T
= __1_ _1_ + 0 + p d
x3 (t) 0 0 x3 0 ..... (6.123)
TTG TTG I
x 4 (t) __1_ I
x4 Ts
0 0 0
RTs -Ts
"2 = x 2
,
x3 = X3 + LlPD
x4 = X4 + LlPD
U + u' + LlPD ..... (6.124)
Eqn. (6.112) can be reduced to
0 1 0 0
.' 1 Kp 0
[~}
Xl 0 0
Tp 1Tp 0
X2
= 0 I u' ..... (6.125)
0 0
X3 Tr Tr I
x4
.' 1
0 0 Ts
RTs Ts
The optimal control to minimize the performance index given in equation (6.111) can be
determined using the solution technique described for the linear regulator problem. It is required
to solve the algebric matrix Ricoati equation.
Q + ATR + RA- RBp-l BT R =0 ..... (6.126)
for the elements of the R matrix which is positive defmite and symmetric. Kleinman's method
may be used to solve the matrix Riccati equation. An initial feedback gain vector KI is selected
such that the matrix ~ = [A - B K I] has eigenvalues with negative real parts. Then the matrix
equations.
Ai RT + ~ Al + Q + ~T P KI = 0
are to be solved for the elements of RI" The new gains are computed using
K(l) = p-l BT R
I I
The procedure is repeated till convergence is obtained for the elements of R I .
228 Operation and Control in Power Systems
Assume Q[~ ~1
Solution:
Substituting the parameters in the algebraic matrix Riccati equation and solving the equation
using, Kleinman's method with an initial vector of K j = [1 1 1 1]
The values of Ki converge to
K = [1.0000 1.1368 1.7092 0.2976]
The optimal control is shown in Fig. E 6.12.
8
7
r
';'
4
::: 2
~
0
::s
0
-1
6.27 Optimal Control for Tandem Compound Single Reheat Turbine- Generator
System
The model for tandem compound single reheat turbine is discussed in
Chapter 3 : The system is shown in Fig. 6.27.
x = AX+Bu+ Fd
with the initial conditions X(O) = 0
where XT = [Xl X2 X3 X4 Xs X6]
= (f~f M ~PG ~PRH ~PCH ~XV ] .... (6.128)
0 1 0 0 0 0
1 Kp
0 0 0 0
Tp Tp
1 FHP
0 0
Tco TJ T2 TCH
A= 0 1 1
0 0 0
TRH TRH
1
0 0 0 0
TCH TCH
1
0 0 0 0
Ts
BT = [0 0 0 0 0 ;s 1
FT = [0 - ~: 0 0 0 0 1
u=APc and d=AP D
The above equations are transformed into the form
Xl =AX I +Bu l
using the transformation
0
0
d
X=X'+
d
d
d
and u~1 + d
A quadratic cost function
= r ±(X'T.QX· +u'TRu'~t
is selected and the algebraic matrix Riccati equation is solved as before to obtain the
optimal gains using the equations.
Q-RB p-l BTR + RA +ATR= 0
and u opt = _p-l BT RX
=_LTX ..... (6.129)
The results are plotted in Fig. E.6.
E 6.13 Compute the optimal load frequency control for a single thermal power system with
tandem compound single-reheat steam turbine with the following data
R = 20s TRH = 7.5s Kp = 2 p.u. Hz / p.u. MW
Ts = O.ls Tco = 0.45s Tp = 20s
TCH = 0.25s FHP = 0.3 D = 0.01 p.u.
FJP = 0.4 FLP = 0.3
Load Frequency Control 231
1 0 0 0 0 0
0 1 0 0 0 0
0 0 0 0 0 0
Q= ; P=1
0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0
The solution is obtained using a computer program for the algorithm explained
in 6.26. The feedback gains, starting from initial values of unity for each, converge to the
following solution.
LT = [1.0 0.6617 0.0269 0.3037 0.0047 0.0177]
Plots of fM and ,M, the two optimally controlled states are shown in Fig. E. 6.13 (a)
and (b)
o 5 10 15
Time (5)
I x 10-3
....
..::: -2 x I0- 3
1 -3xI0-3
-4xI0-4 ,.-------------------
....~t 5
Or------+-~---_r-------+_--~-
10
+
-4xI0-4
15
-8x 10-4
-12x 10-4
BT =[0 0 0 0 0
TpT~TG 1
Tp~Gl
T
C =[0 0 0 0 0
pT = [0
_ Kps
Tps
0 0 0 Kps
Tp TR Tps
1
u = ~ Pc = control input
and d = ~ PD = step-load disturbance.
The constants Kl' K:z, K3 K4 Ks are given by
- Tps -TR
K1-
TpTGTRTpS
_ Kps
K2 -
Tp TG Tps
cr
K 3 =---
TpTGTR
K4 = TR(cr+o)+TG
. TpTGTR
1 1
Ks =-+-
Tp TR
Load Frequency Control 233
The state and control variable are transfonned using the relation,
Xl
0
Xl
X 0
X2 2
X d
X3 3
= + d ..... (6.131)
X4 X
4
0
X5 X5 u'
X6 X6 TpTG
U = u' + d (j
J= ~ r(X'TQX' +u'Tpu')dt
Selecting Q and R matrices are :
1 0 0 0 0 0
0 1 0 0 0 0
0 0 0 0 0 0
Q=
0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0
andP=l
the algebraic matrix Ricccati equation
Q = R B' p-l B'T R + R A' +A'T R = 0
is solved for the elements of R.
The optimal control is obtained using
u = _p-l B'T R X'
opt
=_LTX ..... (6.120)
234 Operation and Control in Power Systems
..... (6.133)
E 6.14 For the hydro system shown in Fig. 6.14.1; using the following data, compute the
optimal control
~ -0.004
~
5 (i) Optimal response
..,l: -0.008
.., (ii) Uncontrolled response
E
E=
-0.012
1 -0.016
(a)
Fig. ES.14.2 Uncontrolled and optimal response of hydro speed governing system
(a) uncontrolled and optimal response of time. error
Load Frequency Control 235
8 = 0.2
The 100 % load conditions on the plant.
.,
0.0004
t 0.0008
E 0.0012
:::- (i) Optimal response
'- (ii) Uncontrolled response
<l 0.0016
+ 0.0020
0.0024
Solution:
Following the procedure outlined in Sec. 6.20, the optimal feed-back gain components are
computed as
LT =[1.0 3.4061 0.078 0.5393 0.1656 0.0063]
The uncontrolled and optimal response are shown in figure E6.14.2 (a) and (b).