R

1.,
8.T
.INTRODUCTION
Power system operati on consi dcrcd so far was under condi ti ons of stcady l oad.
However, both active and reactive power demands are never steady and they
continually change with the rising or falling trend. Steam input to turbo-
generators (or water input to hydro-generators) must, therefore, be continuously
regulated to match the active power demand, failing which the machine speed
will vary with consequent change in frequency whieh may be highly
undesirable* (maximum permissible change in power fiequency is t 0.5 Hz).
Also the excitation of generators must be continuously regulated to match the
reactive power demand with reuctive generation, otherwise the voltages at
various system buses may go beyond the prescribed limits. In modern large
interconnected systems, manual regulation is not feasible and therefore
automatic generation and voltage regulation equipment is installed on each
generator. Figure 8.1 gives the schematic diagram of load frequency and
excitation voltage regulators of a turbo-generator. The controllers are set for a
particular operatirrg condition and they take care of small changes in load
denrand without fiequency and voltage exceeding the prescribed limits. With
the passage of time, as the change in lcad demand becomes large, the
contrcl l ers must be reset ei ther ni anual l y or automati cal l y.
It has been shown in previous chapters that for small changes active power
is dependent on internal machine angle 6 and is inderrendent of bus voltage:
whiie bus voitage is dependent on machine excitation (therefore on reactive
- " -
Change in frequency causes change in speed of the consumers' plant affecting
production processes. Further, it is necessary to maintain network frequency constant
so that the power stations run satisfactorily in parallel, the various motors operating
on the system run at the desired speed, correct time is obtained from synchronous
clocks in the system, and the entertaining devices function properly.
caused by momentary charge in generafor speecl, tI'r.r.tnr*,-i;;?t;qffi; ;;
excitation voltage controls are non-interactive for small changes and can be
modelled and analysed independently. Furthermore, excitation voltage eontrol is
F: t c l : t c f i n r r i n r r r h i n h t h c - , r i ^ r f i r r r o n , r n . . r , r h r ^ 6 ^ , r r i - + ^ - ^ . 1 : - r L ^ e ^ $ r L - - ^ - ^ - - ^ -
r r r vvr r r vr r Lr r v r r r cr Jvr r r l t t w vr Jr r Jr - ( l r r r ur l LUul t Lt r l cu r 5 l l l a! ul ul c; ggi r t cf al or
field; while the power frequency control is slow acting with major time constant
contributed by the turbine and generator moment of inertia-this time constant
is much larger than that of the generator tield. Thus, the transients in excitation
voltage control vanish much faster and do not affect the dynamics of power
frequency control.
Fig. 8.1 schematic diagram of load frequency and excitation
voltage regulators of a turbo-generator
Change in load demand can be identified as: (i) slow varying changes in
mean demand, and (ii) fast random variations around the mean. The regulators
must be dusigned to be insensitive to thst random changes, otherwise the system
will be prone to hunting resulting in excessive wear and tear of rotatins
machines and control equipment.
8.2 LOAD FREOUENCY CONTROL (STNGLE AREA CASE)
Let us consider the problem of controlling the power output of the generators
of a closely knit electric area so as to maintz,in the scheduled frequency. All the
generators in such an area constitute a coherent group so that all the generators
^ - ^ ^ l - I ^ l - - - - . - l ^ - - - - ^ L - - - . r - - ^ . _ _ _ : - , . r .
speeo iip anci siow riowii togetiier rnarntarnrng thelr reiarrve power angies. Such
an area is defined as a control area. Tire boundaries of a coqtrol area will
generally coincide with that of an individual Electricity Board Company.
To understand the load fiequency control problem, let us consider a single
turbo-generator system supplying an isolated load.
I
P+JQ
-
www.MUSLIMENGINEER.info
W Modern
power
system Analys,s
Turbine Speed Governing System
Figure 8.2 shows schematically the speed governing system of a steam turbine.
The system consists of the following components:
Steam
Speed changer
Mai n
piston
A
I
rHydraul i c
ampl i fi er
(speed control mechanism)
Fig.8,2 Turbine speed governing system
Reprinted with permission of McGraw-Hilt Book Co., New York, from Olle l. Elgerd:
Electric Energy System Theory: An lntroduction, 1g71, p. 322.
(i) FIy ball speed governor: This is the heart of the system which senses the
change in speed (frequency). As the speed increases the fly balls move outwards
and the point B on linkage mechanism moves downwards. The reverse happens
when the speed decreases.
G)
Hydraulic amplifier: It comprises a pilot valve and main piston
alrangement. Low power level pilot valve movement is converted into high
power level piston valve movement. This is necessary in order to open or close
the steam valve against high pressure steam.
(xl) Lintcage mechanism: ABC is a rigid link pivoted at B and cDE is
another rigid link pivoted at D. This link mechanism provides a movement to
the control valve in proportion to change in speed. It also provides a feedback
,,fr9rn
the steam valve movement (link 4).
turbine. Its downward movement opens the upper pilot valve so that more steem
is admitted to the turbine under steady conditions (hence more steady power
. The reverse
Model of Speed Governing System
Assume that the system is initially operating under steady conditions-the
linkage mechanism stationary and pilot valve closed, stearn valve opened by a
definite magnitude, turbine running at constant speed with turbin" po*"r output
balancing the generator load. Let the operating conditions be characteizedby
"f"
= system frequency (speed)
P'c = generator output = turbine output (neglecting generator loss)
.IE
=
steam valve setting
We shall obtain a linear incremental model around these operating
conditions.
Let the point A on the linkage mechanism be moved downwards by a small
amount Aye.It is a command which causes the turbine power output to change
and can therefore be written as
Aye= kcAPc
--t-\
Pi l ot
val ue
oi l
Hi gh
pressure
( 8. 1)
(8.2)
where APc is the commanded increase in power.
\
The command signal AP, (i.e. Ayi sets into rnotion a bequence of events-
the pilot valve moves upwards, high pressure oil flows on to the top of the main
piston moving it downwards; the steam valve opening consequently increases,
the turbine generator speed increases, i.e. the frequency goes up. Let us model
these events mathematically.
Two factors contribute to the movement of C:
(i) Ayecontributer
-
[?J
Aya or
-
krAyo(i.e. upwards) of
-
ktKcApc
\ r l l
(ii) Increase in frequency
ff
causes the fly balls to move outwards so that
B moves downwards by a proportional amount k'z Af. The consequent
movemen t of Cwith A remaining fixed at Ayo - .
(+)
orO, - +
kAf
(i.e. downwards)
The net movement of C is therefore
AYc=- kt kcAPc+
kAf
The movement of D, Ayp, is the amount by which the pilot valve opens. It is
contributedby Ayg and AyB and can be written as
Ayo=(h) Ayc+(;h)
*,
-
www.MUSLIMENGINEER.info
=
ktayc + koAys
(g.3)
The movement ay.o-d,epending
upon its sign opens one of the ports of the pilot
valve admitting high pressure'o'
into thJ
"ynnJ.ithereby
moving the main
piston and opening the steam valve by ayr. certain justifiable
simprifying
assumptions,
which ean be rnade at this .tugl, ur",
(i) Inertial reaction forces of main pistoi and steam valve are negligible
compared to the forces exertecl on the
iirton by high pressure
oil.
(ii) Because of (i) above, the rate of oil admitted to the cylinder is
proportional
to port opening Ayo.
The volume of oil admitted to the cylinder is thus proportional
to the time
integral o,f ayo. The movement ay"i.s obtained by dividing the oil volume by
the area of the cross-section
of the-piston.
Thus
Avn= krfoeayrlat
It can be verified from the schematic diagram that a positive movemen t ayo,
causes negative (upward)
movement ayulccounting
for the n"gutiu" ,ign used
i n Eq. (8.4).
Taking the Laplace transform of Eqs. (g.2), (g.3) and (g.4), we ger
AYr(s)=- k&cApc(") + krAF(s)
Ayp(s)= kzAyd,s) + koAyug)
ayu( g=- ksl
or Un
Eliminating
Ayr(s) and Ayo(s), we can write
AY
u(s)
-
k'ktk'AP' (s) -
k,krAF(s)
(oo ' '
t
' )
\
"' tr
,/
-lor,<,r-*^or",].i#)
(8. 4)
(8. 5)
(8.6)
(8. 7)
( 8. 8)
where
n=
kl c
t _
K2
=
speed regulation of the governor
K., =
+y
-
gain of speed governor
. r . l "
,
rs
=
;-;
=
tlme constant of speed governor -
Kq k S
r - -
controt
E
1
E^ , , ^ r i ^ - / o o \ : - . r . -
t
riyLr.Lru'
\o.o., rs rcpfesenleo ln tne ronn of a block diagram in Fig. 9.3.
4Y5(s)
4F(s)
Steam valve
-=-&
Flg. 8.3
,Block
diagram representation of speed governor
system
The speed governing system of a hydro-turbine is more involved. An
additional feedback loop provides temporary droop compensation to prevent
instability. This is necessitated by the targe inertia or the penstoct gut" which
regulates the rate of water input to the turbine. Modelling of a hyjro-turbine
regulating system is beyond the scope of this book.
Turbine Model
Let us now relate the dynamic response of a steam turbine in tenns of changes
in power ouFut to changes in steam valve opening ^4yr. Figure
g.4a
shows a
two stage steam turbine with a reheat unit. The dynamic *ponr" is targely
influenced by two factors, (i) entrained steam betwein the inlet stbam valve and
first stage of the turbine, (ii) the storage action in the reheater which causes the
output of the low pressure stage to lag behind that of the high pressure stage. 'fttus,
the turbine transfer function is characterized by two time constants. For
ease of analysis it will be assumed here that the turbinl can be modelled to have
Ssingle
equivalent time constant. Figure 8.4b shows the transfer function model
of a sream turbine. Typicaly the time constant
{
lies'in the range o.i ro z.s
sec.
AYg(s)-FAPds)
(b) Turbine transfer function model
Fl g. 8.4
Ks9
1 + fsss
(a) Two-stage steam turbine
-
www.MUSLIMENGINEER.info
#ph-Si
rrrroarrn
po*",
s),rt"r An"ly.i,
I
Generator Load Model
The increment in power input to the generatbr-load
system is
APG
_
APD
whele AP6 =
AP,, incremental turbine
incremental loss to be negligible) and App is the load increment.
This increment in power input to the syrtem is accounted for in two ways:
(i) Rate of increase of stored kinetic energy in the generator rotor. At
scheduled frequency (fo
),
the stored energy is
Wk, =
H x
p,
kW = sec (kilojoules)
where P, is the kW rating of the turbo-generator and H is defined as its inertia
constant.
The kinetic energy being proportional to square of speed (frequency),
the
kinetic energy at a frequency of (f " + Arf
)
is given by
=nr,(r.T)
Rate of change of kinetic energy is therefore
$rr*"r
=fffrr"n
(ii) As the frequency changes, the motor load changes being sensitive to
speed, the rate of change of load with respect to frequ"n.y, i.e. arot\ycan be
regarded as nearly constant for small changes in frequency Af ard can be
expressed as
Automatlc Generation and Voltage Control
I
=tAP6g)_
aPo(,)r.[#j
(s. 13)
2H
Bf"
= powr system time constant
Kp, =
+
=power
system gain
Equation (8.13) can be represented in block diagram form as in Fig.
g.5.
laeo(s)
^Po(s) 16---ffioro,
Flg. 8.5 Block diagram representation of generator-load
model
complete Block Diagrram Representation of Load Frequenry
Control of an Isolated Power System
(8.e)
( 8. 10)
( 8. 11)
positivo for a
@PDl ?f l Af =BAf
where the constant B can be determined empirically, B is
predominantly motor load.
Writing the power balance equation, we have
APc - aP^ =THP'
d
(
, r =
- f . ] *
<of l + B Af
Dividing throughoutby
p,
and rearanging, we get
AP(s)=trPn15;
AP6(s)
Flg. 8.6 Block diagram model of load frequency control
(isolated power system)
Steady States Analysis
The model of Fig. 8.6 shows that there are two important incremental inputs to
the load frequency control system
-
APc, the change in speed changer setting;
and APo, the change in load demand. Let us consider,,.4,.simple situatiqn in
AP6$u) - AP; q; u) =
1d
/ A' ^ ' n' 7' - - - \
f
dt
(Afi + B(ptt) af (8.i2)
Taking the l,aplace transforrn, we can write AF(s) as
4Fis;
-
AP,G)
-4PoG)
B* - ' - s
-
www.MUSLIMENGINEER.info
Modern
whi ch t he sneerl . hqnrro' hoo . . g. -. ^) ^-. -.
r.han.,oo
' t::;
;-::--::^
' (rr cr rr^tr(r ucttrng
\7' e' af
c
=
o) and the load demand
:,:il?: ; 3l i: T: ::
a2
rr
e e g o,, * ;, 2
;
;; ;*;:r;;ffi;
*'
#ff:Tiil: steady change in system frequen-cy for a sudd.n .hung", ffi;ffi"ffi;ti'l;
anaount
*,
(,
e.Apog):+)is
obtained as follows:
aF@)l*,(s):o : - AP^
^f
K I ( = 1
r ^ s o r r , . I
It is also rccognized that Ko, =
in frequency).
Now
4=-(#6)o,.
7/ B, wher eB- Y^
ai
/P' (in
Pu
MWunit change
( 8. 16)
fi
roa
(J
L
8. rog
. c
102
at
l i \ d A^ t | |
\r,, ruu-lo Loao
(i i ) 60% Load
101
100
0
Percent Load
Flg. 8.7 Steady
"*-l?39-frequency
qharacteristic
of a speed
governor
system
. r L^
^L^- - ^ I
r'E .1uuy' cquauon glves tne steady state changes in frequency caused by
changes in load demand. Speed regulation R is-naturally
so adjusted that
changes in frequency are small (of the order of 5vo from no load to ruu load).
Therefore, the linear incremental relation (g.16)ican
be applied from no load to
full load' with this understanding,
Fig. 8.7 shows the linear relationship
between frequency and load for free governor operation with speed changer set
to give a scheduled frequency of r00% at full toao. The
.droop,
or slope of this
(
relationship is
-l
I
'l
-
\
B+( t / R)
)
Power system parameter B is generaily
much smalrer* than r/R (a typical
value is B =
0.01 pu Mwalz and l/R =
U3) so that B can be neglected in
comparison. Equation (8.16) then simplifies to
rhe droop
"r,,fl",
fjfli;], curve is
speed governor regulation.
( 8. 17)
thus mainly determined by R, the
MW. let the change in load
= 50 Hz). Then
ap,=_
*"r:
(r^;)o",
Decrease in system load =
BAf=
(uffi)*,
Of course, the contribution of decrease in system load is much less than the
increase in generation. For typical values of B and R quoted earlier
APo =
0.971 APo
Decrease in system load =
0.029 ApD
consider now the steady effect of changing speed changer setting
(Or"<rl-
+)with
load demand remaining fixed (i.e. Apo= 0). The sready
state change in frequency is obtained as follows.
*For
250 MW machine with an operating load of 125
be i%o for IVo change in frequency (scheduled
frequency
a-:?:r?:
:2.5
NNVtHz
af 0.s
:
#:
o'ol
Pu
Mwgz
' =(#)b
-
www.MUSLIMENGINEER.info
W
uodern Power system Analysis
I
AF@lap,{s):o:
r t v (
t t sgt t f t ^ ps
A D
x u' c ( 8. 18)
s ( 1+
T, r s) ( l *
4s)
I
4,flr*uoyro,":
I
AP',:g
xl-
_ t
( 1
K
\
I
- l
r l
p,
/ R
AP,
* zors) + KseKt
KreKrKp,
+ K. sKt Kps / R
( 8. 1e)
(8.20)
If
Kr r K, =l
Ar=
(
|
\rc"
"
\ B + l l R )
If the speed changer setting is changed by AP, while the load demand
changes by APo, the steady frequency change is obtained by superposition, i.e.
( 8. 21)
According to Eq. (8.2I) the frequency change caused by load demand can be
compensated by changing the setting of the speed changer, i.e.
APc- APo, for Af = Q
Figure 8,7 depicts two load frequency plots-one to give scheduled
frequency at I00Vo rated load and the other to give the same frequency at 6O7o
rated load.
A 100 MVA synchronous generator operates on full load at at frequency of 50
Hz. The load is suddenly reduced to 50 MW. Due to time lag in governor
system, the steam valve begins to close after 0.4 seconds. Determine the change
in frequency that occurs in this time.
Given H =
5 kW-sec/kVA of generator capacity.
Solution Kinetic energy stored in rotating parts of generator and turbine
= 5 x 100 x 1.000 = 5 x 105 kW-sec
Excess power input to generator before the steam valve
begins to close = 50 MW
Excess energy input to rotating parts in 0.4 sec
= 50 x 1,000 x 0.4 = 20,000 kW-sec
Stored kinetic energy oo (frequency)2
Frequency at the end of 0.4 sec
= 5o x I
soo,ooo + zo,ooo
)t"= 5r rfz
\ 500,000 )
Ar =
(
".
ru)
'o"
-
APo)
Autor"tic G"n"r"tion and Volt"g" Conttol
F
Two generators rated 200 MW and 400 MW are operating in parallel. The
droop characteristics of their governors are 4Vo and 5Vo, respectively from no
load to full load. Assuming that the generators are operating at 50 Hz at no
load, how would a load of 600 MW be shared between them? What will be the
system frequency at this load? Assume free governor operation.
Repeat the problem if both governors have a droop of 4Vo.
Solution Since the generators are in parallel, they will operate at the same
frequency at steady load.
Let load on generator 1 (200 MW)
= x MW
and load on generator 2 (400 MW)
= (600
-
x) MW
Reduction in frequency
= Af
Now
a f _
x
af
600- x
Equating Af in (i) and
v -
600- x =
System frequency
=
50
-
0' 0-1150
x 231
= 47 .69 Hz
'
200
It is observed here that due to difference in droop characteristics of
governors, generator I gets overloaded while generator 2 is underloaded.
It easily follows from above that if both governors have a droop of.4Vo, they
will share the load as 200 MW and 400 MW respectively, i.e. they are loaded
corresponding to their ratings. This indeed is desirable from operational
considerations.
Dynamic Response
To obtain the dynamic response giving the change in frequency as function of
the time for a step change in load, we must obtain the Laplace inverse of Eq.
(8.14). The characteristic equation being of third order, dynamic response can
r ' r | 1- ! - - - I f - , - - - ^ ^ ^ t C: ^ - - - * ^ - : ^ ^ 1 ^ ^ ^ ^ t I ^ . - , ^ , , ^ - +L^ ^ L^ - ^ ^ +^ - - i ^ +i n
Onfy Dg ODIal neU l uf A SPtrUfffU l l |' l l l l (' l l ua1' I Ua1DE. II(rwsYsIr LfIs r,Il <l l 4i vLsl l Drl v
equation can be approximated as first order by examining the relative
magnitudes of the time constants involved. Typical values of the time constants
of load frequency control system are rdlated as
0.04 x 50
200
0.05 x 50
400
(ii), we get
231 MW (load on generator
- / A t r l t f / 1 ^ - l ^ -
JOy lvlw (IUau ull
Btrrltrriltur
(i)
(ii)
r)
L )
-
www.MUSLIMENGINEER.info
T r r 4 T ,
< T o ,
Typically* t, =
0.4 sec, Tt =
0.5 sec and
Flg' 8.8 First order approximate brock diagram of road
frequency controt of an isolated area
Irning Tro =
T, =
reduced to thlt of F'ig.
AF(s)l*r(s):o =
Ar (,)= -ft{' -
*,[-,,a[n#)]]
*, g
22)
Taking R =
3, Kp, =
llB =
100,
e,
=
20, Apo = 0.01 pu
Af (t) = -
0.029 (I -
,-t:tt' ,
Aflrt"udystare = -
0.029 Hz
0:
Iuld
K*\ =1),
the block diagram of Fig. 8.6 is
8.8, from which we can write
-
to,
.-.
APo
(1+ Kps l R)+ Zp.s
"
s
-
- "o{1:- =xaP,
, l , +^+r o' 1
L
R 4 , J
Dynamic response_of change in frequency for a step change in load
(APo= 0.01 pu,
4s
= 0.4 sec,
|
= 0.5 sLc, Io. = 2b sec, (" = 100,
R= 3)
The plot of change in frequency versus time for first order approximadon
given above and the exact response are shown in Fig. a.g.
^rirst
order
approximation is obviously a poor approximation.
Gontrol Area Concept
So far we have considered the simplified case of a single turbo-generator
supplying an isolated load. Consider now a practical system with e number of
generating stations aird loads. It is possible to divide an extended power system
(say, national grid) into subareas (may be, State Electricity Boards) in which
the generators are tightly coupled together so as to form a coherent group, i.e.
all the generators respond in unison to changes in load o, ,p"rJ changer
settings. Such a coherent area is called a control area in which the frequency
is assumed to be the same throughout in static as well as dynamic conditions.
For purposes of developing a suitable control strategy, a control area can be
reduced to a single speed governor, turbo-generator and load system. All the
control strategies discussed so far are, therefore, applibable to an independent
control area.
Proportional Plus fntegral Control
It is seen from the above discussion that with the speed governing sysrem
installed on each machine, the steady load frequency charartitirti" fi agiven
speed changer setting has considerable droop, e.g. for the system being used for
the illustration above, the steady state- droop in fieo=ueney will be 2.9 Hz
[see
Eq. (8.23b)l from no load to tull load (l pu load). System frequency
specifications are rather stringent and, therefore, so much change in frequency
cannot be tolerated. In fact, it is expected that the steady change in frequency
will be zero. While steadystate frequency can be brought back io the scheduled
Time (sec)------->
- 1
t
I
I
o
First order approximatiorl
(8.23a)
(8.23b)
"For
a 250 MW machine quoted earlier, inertia constanr
, = 4 : . 2 * 5 =
= 2 o s e c
'
Bf
o
0.01x 50
Il =
SkW-seclkVA
-
www.MUSLIMENGINEER.info
ffil
Modern Power system Analys
I
vaiue by adjus'ring speed changer setting, the system could under go intolerable
dynamic frequency changes with changes in load. It leads to the natural
suggestion that the speed changer setting be adjusted automatically by
monitoring the frequency changes. For this purpose, a signal from Af is fed
througfan integrator to the s diagram
configuration shown in Fig. 8.10. The system now modifies to a proportional
plus integral controller, which, as is well known from control theory, gives zero
steady state error, i.e. Af lrt""d"
,,ut,
= 0.
Integral
controller
APp(s)
AP6(s)
Frequency sensor
Fig. 8.10 Proportional plus integral load frequency control
The signal APr(s) generated by the integral control must be of opposite sign
to /F(s) which accounts for negative sign in the block for integral controller.
Now
( l * f , r s) ( l +4s)
RKo,s(l +
{rs)(l + 4s)
obviousry
+
{' s)(1
+
4sXl
f zo' s)R * Ko' (KiR f s)
Af l"t"^dy
state
=
,
so/F(s)
: o
In contrast to Eq. (8.16) we find that the steady state change in frequency
has been reduced to zero by the additio4 of the integral controller. This can be
argued out physically as well. Af reaches steady state (a constant value) only
rr.,lrsrr Ap^
-
Ap- = .ons-fant Becarrs-e of fhe intes!'atins actiOn Of the
w l M l u r c - H r D
- v v u u l q r ! .
controller, this is only possible if Af
= 0.
In central load frequency control of a given control area, the change (error)
in frequency is known as Area Contol Error (ACE). The additional signal fed
back in the modified control scheme presented above is the integral of ACE.
l +t-r8-
I I t - +
t l
Kn,
AF(s1 =
(r
+
%"s). (*
*
+).
Ko,
"+
APe(s)
-1
+
I
I
t
o
r
x
AF(s)
Automatic Generation and Voltage Control
t-
in ihe above scheme ACE being zero uncier steaciy conditions*, 4 logical
design criterion is the minimization of II,CZ dr for a step disturbance. This
integral is indeed the time error of a synchronous electric clock run from the
power supply. Infact, modern powersystems keep Eaekofintegra+e4tinae errsr
all the time. A corrective action (manual adjustment apc, the speed changer
setting) is taken by a large (preassigned) station in the area as soon as the time
error exceeds a prescribed value.
The dynamics of the proportional plus integral controller can be studied
numerically only, the system being of fourth order-the order of the system has
increased by one with the addition of the integral loop. The dynamic response
of the proportional plus integral controller with Ki = 0.09 for a step load
disturbance of 0.01 pu obtained through digital computer are plotted in Fig.
8.11. For the sake of comparison the dynamic response without integral control
action is also plotted on the same figure.
Flg. 8.11 Dynamic response of load frequency controller with and without
integral control action (APo = 0.01 pu,
4s
= 0.4 sec, Ir =
0.5
sec, Ips = 20 sec, Kp. = 100, B
-
B, Ki= 0.-09)
8.3 IOAD FREOUENCY CONTROL AND ECONOMIC
DESPATCH CONTROL
Load f reouencv cont rol wi t h i nt eorel eonf rnl l er qnhi errAe ?a?^ craol rr ora+o
I
_ _ J _ _ _ _ _ _ _ _ , . _ _ _ _ - - - _ O ' v u l v r v o t v s u J D l 4 l g
frequency elTor and a fast dynamic response, but it exercises no control over the
relative loadings of various generating stations (i.e. economic despatch) of the
control area. For example, if a sudden small increase in load (say, 17o) occurs
'Such
a control is known as isochronous control, but it has its time (integral of
frequency) error though steady frequency error is zero.
(8.24)
(8.25)
-
www.MUSLIMENGINEER.info
1i..l1r_::ltrol
area, the road
-frequency
conrior
,changes
the speed changer
Dcrurgs or tne governors of all generating units of the area so that, together,
these units match the load and the frequenry returns tp the scheduled value (this
action takes place in a few seconds). However, in the,process of this change the
Ioadings of u@units
change in a manner independent of
economi@
In fact, some units in the pro""r, may even
get overloaded. Some control over loading of individual units cafi be Lxercised
by adjusting the gain factors (K,) includeJin the signal representing integral of
the area cogtrol error as fed to individual unitr. However, this is not
satisfactory.
l r.
EDC
-
Economi c despatch control l er
CEDC
-
Central economi c despatch computer
Fl g. 8-12 Control area l oad frequency and economi c despatch control
Reprinted (with modification) with permission of McGraw-Hill Book Company,
New York from Olle I. Elgerd: Electric Energy Systems Theory: An Introd.uction,
I 971, p. 345.
"fnce
ot
Speed
Automatic
f
_T---
command signai generated'oy
the centrai economic despatch computer. Figure
8'12 gives the schematic diagram of both these controlsior two typi.ut units of
a control area. The signal to change the speed chan3er setting is lonstructed in
accordance with economic despatch error,
[po
(desired) - pJactual)].
suitabry
modified by the signal representing integral ncg at that instant of time. The
signal P6 (desired)
is computed by the central economic despatch computer
(CEDC) and is transmitted to the local econornic despatch controller (EDC)
installed at each station. The system thus operates with economic desfatch error
only for very short periods of time beforJ it is readjusted.
8.4 TWO-AREA LOAD FREOUENCY
CONTROL
An extended power system can be divided into a number of load frequency
control areas interconnected by means of tie lines. Without loss of generality
we
shall consider a two-area case connected by a single tie line as lilusnated in
Fi g. 8. 13.
Fi g. B.i 3 Two i nterconnected contror areas (si ngre ti e ri ne)
The control objective now is to regulate the frequency of each area and to
srnnultaneously regulate the tie line power as per inter-area power contracts. As
in the case of frequency, proportional plus integral controller will be installed
so as to give zero steady state error in tie line power flow as compared to the
contracted power,
It is conveniently hssumed that each control area canbe represented by an
equivalent turbine, generator and governor system. Symbols used with suffix I
refer to area 7 and those with suffi x 2 refer to area 2.
In an isolated control area case the incremental power (apc _
apo) was
accounted for by the rate of increase of stored kinetic energy and increase in
area load caused by increase in fregueircy. since a tie line t *rport, power in
or out of an area, this fact must be accounted for in the incremental power
balance equation of each area.
Power transported out of area 1 is .eiven bv
Ptie,
r
=
''rrl''l
sin ({ -
q
X,,
where
q'q -
power angles of equivalent machines of the two areas.
(8.26)
-
www.MUSLIMENGINEER.info
I
308 |
Modefn Power System Analysis
I
For incremental changes in
{
and 6r, the incre.mental tie line power can be
expressed as
AP,i ,,
r(pu)
= Tp(Afi
-
462)
where
T, =
' Y:t' -Yf
cos (f
-
E)
-
synchronizing coefficient
PrrXrz
Since incremental power angles are integrals of incremental frequencies, we
can wri te Eq. (8.27) as
AP,i,,
r
= 2*.(l Afrdt
-
I
Urat)
where Afi nd Af,, arc incremental frequency changes of areas 1
respectively.
Similarly the incremental tie line power out of area 2 is given by
aPt;",
z
= 2ilzr([ yrat -
[
ayrat)
where
rzr
=
t Yr: J
cos ({
-
E): [ S] t i z:
ar2rrz (s. 30)
L L
Pr z x z r
L
"
\ Pr r )
With reference to Eq. (8.12), the incremental power balance equation for
area 1 can be written as
APo,
-
APor =
+ *w)+
nrz| r* AP,,",t
Jr"
or
(8. 27)
( 8. 28)
and 2,
(8.2e)
It rnay be noted that al l quanti ti es other than fi equency are i n per uni t i n
Eq. ( 8. 3l ) .
Taki ng the l -apl ace transf' orm of Eq. (8.31) and reorgani zi ng, we get
AF( s) =
I AP61G)
-
APr , ( s)
-
APt i ", , 1r ; ]
"t $-
$. 32)
I +
4, , t , !
where as defined earlier [see
Eq. (8.13)]
Kp31 = I / 81
Tpi l = LHr/BJ"
Compared to Eq. (8.13) of the isolated control area case, the only change is
the appearance ol the signal APri"J (s) as shown in Fig. 8.14.
' - l ' ^ Li - -
f ho T - ^ l - ^ a f * onof nr m ^ f E^ / a t a\ t ha ci onol , 4P / " \ i c nl r f oi ner l
I 4Ar r r S r r r w l sl / l 4vv Ll 4l l Dr ur r r r ur LY.
\ v. L9) ,
l l r v or 6r r s^
" , t i e. I \ . r /
AS
AP,i .,1(s) =
ffroor(s)
-
/4 (s)l
( 8. 31)
( 8. 33)
(8. 34)
Automatic Generation and Vortage contror
Fil
I
APti ".r(s)
Fi g. 8. 14
The correspondi ng bl ock di agram i s shown i n Fi g.
g.15.
+
APti ",r(s)
AF1(s)
-iE=
--n7ri"l
Fi g. 8. 15
For the control area 2, Ap6",
r(s)
is given by
tEq.
(g.Zg)l
apt i ", z( s) =
- : gr r r ,
[ AFr ( s)
-
4F, ( s) ]
( g: 35)
which is also indicated Uy ,i. block diagram of Fig. 8.15.
\
Let us now turn our attention to ACE (area control error; in the presence
of
a tie line. In the case of an isolated control area, ACE is the change in area
frequency which when used in integral control loop forced the steady state
frequency elror to zero. In order that the steady state tie line power error in a
two-area control be made zero another integral control loop (one for each area)
must be introduced to integrate the incremental tie line power signal and feed
i t back to the speed changer. Thi s i s aeeompl i shed by a si ngl e i ntegrati ng bl oek
by redef ining ACE as a linear combination of incremental frequenry and tie line
power. Thus, fbr control area I
ACEI = APu" . r + br Af ,
where the constant b, is called area
frequency bias.
Equation (8.36) can be expressed in the Laplace transform as
ACEl (s) = APo.,
r(s)
+ b1AF1g)
Similarly, for the control are a 2, ACE2 is expressed as
ACEr(s) = APti".z(s) + b2AF,(s)
Combining the basic block diagrams of the two control areas corresponding
to Fig. 8.6, with AP5rg) and Apr2(s) generated by integrals of respective
ACEs (obtained through signals representing changes in tie line power and local
frequency bias) and employing the block diagrams of Figs.
g.t+
to
g.15,
we
easily obtain the composite block diagram of Fig.
g.16.
( 8. 36)
(8.37)
( 8. 38)
-
www.MUSLIMENGINEER.info
WIU&| Modern Power svstem Analvsis
Let the step changes in loads APo, and APrrbe simultaneously applied in
control areas 1 and 2, respectively. When steady conditions are reached, the
output signals of all integrating blocks will become constant and in order for
this to be so, their input signals must become zero. We have, therefore, from
Fi e. 8. 16
APu",
,
+ b
rAfr=
O
finput
of integrating block
-
KtL)
-
\ , r )
APti",
,
+ brAfr= o
finpot
of integrating block
-
K'z)
-
\ r l
Afr
-
Afz =o
finpurot
integrating block
-'4'\
- \ s )
From Eqs. (8.28) and (8.29)
APn" , ,
=- Tr ,
- .
I . =const ant
AP. i " , z , T z t ; a r 2
Hence Eqs. (8.39)
-
(8.41) are simultaneously satisfied only for
(8.39a)
(8.3eb)
(8.40)
(8. 41)
(8.42)
and
Thus, under steady condition change in the tie line power and frequency of
each area is zero. This has been achieved by integration of ACEs in the
feedback loops of each area.
Dynamic response is difficult to obtain by the transfer function approach (as
used in the single area case) because of the complexity of blocks-and multi-
input (APop APor) and multi-output (APri",1, Ap6",2, Afr Afr) situation. A
more organized and more conveniently carried out analysis is through the state
space approach (a tirne domain approach). Formulation of the state space model
for the two-area system will be illustrated in Sec. 8.5.
The results of the two-area system (APri", change in tie line power and, Af,
change in frequency) obtained through digital computer study are shown in the
form of a dotted line in Figs. 8.18 and 8.19. The two areas are assumed to be
identical with system parameters given by
Trs= 0.4 sec, 7r = 0.5 sec, ?r, = 20 sec
Ko r = 1 0 0 , R= 3 , b = 0 . 4 2 5 ,
&= 0 . 0 9 , 2 f l r 2 = 0 . 0 5
8.5 OPTTMAL (TWO-AREA) LOAD FREOUENCY CONTROL
Modern control theory is applied in this section to design an optimal load
frequency controller for a two-a3ea system. In accordance with modern control
terminology APcr arrd AP62 will be referred to as control inputs q and u2.ln
the conventional approach ul and uzwere provided by the integral of ACEs. In
.Y
o
(U
-o
!t
o
g
o y ,
E F
8 6
:pE
o 6
=a c)
- ( d
E 9
g t u
Q c l
e . >
( g ( )
o o
F A
O E
5 *
* o
b 6
t r ( ' )
E g
H' .s
* 3
5 u
= o
a ' 5
' 8 9
o o .
* , n
E O -
o o
o o
@
d
<;
l!
E
o
o
o
o
o
G'
6
o
o
E
o
E
(D
()
(U
CL
.n
o
o
o
r\
ai
cit
lr
tt)
.9
d
APr i " , r = AP, : " , 2=0
Af i = Af z =0
a
(\
.g
q
<.1
t
l +
5 l
- l d
trJ
o
oi
ra
I
il
S A
N
o
o-
u I f|:-
* l i
o l . a
v'it
l r
-
www.MUSLIMENGINEER.info
itZ I
rrrrodern Power System Analysis
-
modern control theory approach ur and u2 wtll be created by a linear
combination of all the system states (full state feedback). For formt'lating the
state variable rnodel for this purpose the conventional feedback loops are
resented bv a se block as shown in
Fig. 8.17. State variables are defined as the outputs of all blocks baving either
an integrator or ar tirne, constanf.. We immediately notice that the systern has nine
state variables.
-1-+-r+--i,
-f-.'f--
\
Opti mal case (ful l state feedback)
'
Wi th i ntegral control acti on
change i n ti e l i ne power due to step l oad (0.01 pu) change i n area 1
Automatic Generation and Voltage controt
M&
f*
Compari ng Fi gs. 8.16 and 8.17,
xt = Aft
.r2 - AP,;1
xq = Af.
x5 = AP52
XS =
JACE
it
t, =
JACE, dt
I
t
t-
,I
1 8
- l
t'-2
-3
1
8. '
+
I
I
I
o
o
X
(
Fi g.
A
L
I
N
I
o
x
IL
r;+-.1 I
-21.--+-_'--';-;;7-1=a.-1-1=--1
/ ' 8
' -
- - ' 12 14 16 18 20
/
Ti me (sec)-----
t t 1= APg,
w1= AP"
For block 1
x1 + T. r r i , =
L P
. 1
h l
-
4 l
' p s l
For block 2
x. 2+ Ti l i z= xt
or *z=
- +- r **n
For block 3
t r +
{ , sr i :
= - Lr r +r ,
' R , r r
or *t =
-
^h
r , -
t *, **, ,
For block 4
X n *
+
or i q=
For block 5
x s t
or i s=
For block 6
x s *
u2 = /)Pa
w2 = APp,
K^t ( xz-
h
-
w)
,
Kpr t
, -
Kpr t
- ,
Kpr t
* f x z - ; - x t
- ; - wt
( 8. 43)
t
psl
t
pt l
t
pt l
( 8. 445
(8. 45)
( 8. 46)
( 8. 47)
wi th i ntegral control acti on
Opti mal case (ful l state feedback)
Fi g. 8. 19 Change i n f requency of area 1 due t o st ep l oad (0. 01 pu)
change i n a. rea 1
Before presenting the optimal design, we must formulate the state model.
This is achieved below by writing the differential equations ciescribing each
individual block of Fig. 8.17 in terms of state variables (note that differential
equations are written by replacing s UV
*1. '
d t '
Torz*+= Krrz(xs + ar2x7
-
wz)
I
Knr z at ?K
or 2
Ko*2
' \ A ' 1 - - . { <
- T - - - y ' - a - _ - W ^
Tprz
-
Tps2
''
Tps2
'
Tpsz
z
7, 2i 5 - x6
l 1
Y I - V
4 < t 4 l
Ttz
r
T,z
u
. l
I
,szx6
- -;
x4 + u2
I \ 2
or i o=
- #* o- * * u
' 2t
sg2
t
sg2
( 8. 48)
-
www.MUSLIMENGINEER.info
' 3i4' ,"1 Modern Power System Anatysis
T
For block 7
i t =2i Tt z x t - 2i Tr 2x a
For block 8
i s= brx, + x.i
For block 9
(8. 4e)
(8.5O-)
( 8. s 1)
(8.s2)
i 9= b2xa- anxt
The nine equations (8.43) to (8.51) can be organized in the following vector
matrix form
A _
where
x _ l x r
u = f u t
w = l w t
while the matrices
* = A x + B u + F w
x2 ... xg)r = state vector
u2fT
= control vector
w2fT = clisturbance vector
A, B and F are defined below:
2 3 4 5
Y
' t Psl
0 0 0
Tprt
- 1
1
o o
Tt Ttr
o
- 1
o o
Trst
0 0 - 1 K p ' z
Tprz Tprz
o o i o
--1-
Ttz
I
I
Tpst,
0
1
Rr4er
0
0
0
2 irrz
bL
0
7 8 9
- b L o o
Tprt
0 0 0
0 0 0
at zKprz
0 0
Tprz
0 0 0
6
7
8
9
o o - 1 0
RzTrsz
0 0
-2i l r2
0
0 0 0 0
0 0 b 2 0
[ o o
I
o o o
I T
Br =
|
-ss1
l o o
o o o +
I
a c o )
L ' O-
1
7,,
I
TreZ
0
0
0
0
0
1
- at z
0 0
0 0
0 0
0 0
-
Kpr t
0
Tprt
0 0 . l
I
I
0 0 l
I
J
,;T
(8.s7b)
"
'--
co","".t""
constructed as under from the state variables x, and -rn only.
ut =- Ki r x s =- Ki r I eCn, Ar
uz=- Ki { s=- Ki z l a. Cer ar
ln the optimal control scheme the control inputs u, and uz are generated by
means of feedbacks from all the nine states with feedback constants to be
determined in accordance with an optimality criterion.
Examination of Eq. (8.52) reverals that our model is not in the standard form
employed in optimal control theory. The standard form is
i = A x + B u
which does not contain the disturbance term Fw present in Eq. (g.52).
Furthermore, a constant disturbance vector
p
would drive some of the system
states and the control vector z to constant steady values; while the cost function
employed in optimal control requires that the system state and control vectors
have zero steady state values for the cost function to have a minimum.
For a constant disturbance vector w, the steady state is reached when
* = 0
i n Eq. (8.52); whi ch then gi ves
0=A. r r " + Bur r + Fw (8.s3)
Defining x and z as the sum of transient and steady state terms, we can write
,
x = x' * Ir" (8.54)
n = ut * z' ,
(8.55)
Substituting r and z from Eqs. (8.54) and (8.55) in Eq. (8.52), we have
i' = A (r/ + x"r) + B(at + usr) + Fw
By virtue of relationship (8.53), we get
*' = Axt + But ( g. 56)
This represents system model in terms of excursion of state and conhol
vectors fiom their respective steady state values.
For full state feedback, the control vector z is constructed by a linear
combination of all states. i.e.
u=- Kx ( 8. 57a)
where K is the feedback matrix.
Now
t t t + I t r r =- l ( ( r / + r r " )
For a stable system both r/ and ut go to zero, therefore
ur, = _
Kx*
Hence
tt
/=
-
Ikl
-
www.MUSLIMENGINEER.info
Modern Power System Analysis
Examination of Fig. 8.17 easily reveals the steady state values of state and
control variables for constant values of disturbance inputs w, andwr. These are
I l r r =X4" r = / 7r " =
0
Automatic Generation
b ? o
0 0
0 0
0 0
0 0
0 0
0 0 0 0 0 0
0 0 0 0 0 0
4 0 0 - a n b z 0 0
0 0 0 0 0 0
0 0 0 0 0 0
4
0 0
0 0 0
0 0 0
-arzbz
0 0
0 0 0
0 0 0
Q+a?) o o
0 1 0
0 0 1
(8. 63)
for which the system remains stable.
the system dynamics with foedback is
0
0
0
0
0
0
ul r , = wl
r 5r r = x6r r = l v2
uzr, = wz
(8. s8)
Igr, = COnstant
I9r, = Constant
The values of xr* and xe* depend upon the feedback constants and can be
determined from the following steady state equations:
ut rr= kt t xt r, + . . . + f t t 8r8", * kt *sr, = wl
r,t2ss = k2txl r, + ... + kzgxgr., * kz*gr, =
wz (8.se)
The feedback rnatrix K in Eq. (8.57b) is to be determined so that a certain
performance index (PI) is minimized in transferring the system from an
arbitrary initial state x' (0) to origin in infinitie tirne (i.e. x' (-) =
0). A
convenient PI has the quadratic form
'
Pr =
;ll
' .' ' Qx'
+ u' r Ru' dt
The manices
Q
arrd R are defined for the problem in hand through the
fol l owi ng desi gn consi cl erati ons:
(i) Excursions of ACEs about the steady values (r,t + brx\;
-
arrxt, + bzx,q)
are minimized. The steady values of ACEs are of course zero.
(ii) Excursions of JnCg dr about the steady values (xts,
xte) are nrinimized.
The steacly values of JeCg dt are, of course, constants.
(i i i ) Excursi ons o1' the contt' ol vector (ut1, ut2) about the steady val ue are
rninirnized. The steady value of the control vector is, of course, a constant.
'
Thi s nri ni mi zati on i s i ntended to i ndi rectl y l i mi t the control effbrt wi thi n
the physical capability of components. For example, the steam valve
catmot be opened more than a certain value without causing the boiler
presi sure to drop severel y.
With the above reasoning, we can write the PI as
= symmetric matrix
R - kI =
symmetric matrix
K =
R-rBrS
The acceptable solution of K is that
Substituting Eq. (8.57b) in Eq. (8.56),
defined bv
i' = (A -
BIgx,
(g.64)
Fol stability all thc cigenvalues of the matrix (A
-
Bn should have negative
real parts.
For illustration we consider two identical control areas with the following
syste|ll parameters:
4r*
= 0'4 scc; T'r = 0.5 sec; 7'r* = 20 sec
/l =
3:
(n* = l /l J =
100
b = O. 425; Ki =
0. 09; up =
I ; 2i l n =
0. 05
f, [
0.52tt6 l .l 4l 9 0.68 l 3
-
0.0046
-0.021
|
-0.0100 -0.743
7 0.gggg0.00001
^ =
L-o.tl 046-0.o2tl -0.0100 0.5286 t.t4rg 0.6813 0.74370.0000 0.gggsl
(8.60)
pr=
* fU-+
+ h,.r,,)2 + (- tt,2xt, + brxta)z + (.r,? + ,,])
2 J t t '
+ kfu' l + u,|11 at
From the PI of Eq. (8.51),
Q
md R can be recognized as
( 8. 61)
'*Refer
Nagrath and Gopal
[5].
-
www.MUSLIMENGINEER.info
iiii'f:l Modern
power
rystem in4gs
As the control areas extend over vast geographical regions, there are two
ways of obtaining full state information in each area for control purposes.
(i) Transport the state information of the distant area over communication
channels. This is, of course, expensive.
8.6 AUTOMATIC VOLTAGE CONTROL
Figure 8.20 gives the schematic diagram of an automatic voltage regulator of
a generator. It basically consists of a main exciter which excites the alternator
field to control the output voltage. The exciter field is automatically controlled
through error e =
vr"r
-
vr, suitably amplified through voltage and power
amplifiers. It is a type-0 system which requires a constant error e for aspecified
voltage at generator terminals. The block diagram of the system is given in
Fig. 8.20 Schematic diagram of alternator voltage regulator scheme
Fig. 8.21. The function of important components and their transfer functions is
given below:
Potential transformer: It gives a sample of terminal voltage v..
Dffirencing device; It gives the actuating error
c= vR.f
-
vr
'_-
The error initiates the corrective action of adjusting the alternator excitation.
Error wave form is suppressed carrier modulated, tt" carrier frequency being
the system frequency of 50 Hz.
Change in voltage
caused by l oad
Load change
Fig. 8.21 Brock diagram of arternator vortage regurator scheme
Error amplifier: It demodulates and amplifies the error signal. Its gain is Kr.
scR power amplffier and exciter
fierd: It provides the n"."rriry power
amplification to the signal for controlling thl exciter n"ro.- arr*;"g ,rr"
amplifier time constant to be small enoughio be neglected, the ovelail fansfer
function of these two is
K,
l * T"rs
where T"y is the exciter field time constant.
Alternator; Its field is excited by the main exciter voltage vu. Under no road
it produces a voltage proportional
to field current. The no load transfer function
i s
Ks
7 * T* s
where
T*= generator field time constant.
The load causes a voltage drop which is a complex function of direct and
quadrature axis currents. The effect is only schematically reBresented hv hlock
G.. The exact load model of the alternator is beyond ,t" ,iop" ;rhtJ;;:
stabitizing transformer: T4*d
-lq
are large enough time constants to impair
the system's dynamic response. Itjs weil known that the dynami. r"rpoor" of
a control system can be improved by the internal derivative feedback loop. The
derivative feedback in this system is provided by means of a stabi
yzing
transformer excited by the exciter output voltage vE. The output of the
tG
1+Iers
skrt
L
o
A
D
Potential
-
www.MUSLIMENGINEER.info
I
320'
l
Modern Power System Analysis
I
stabiliz,ing transformer is fccl ncgativcly at the input terminals of thc SCR power
amplifier. The transfer function of the stabilizing transfo"mer is derived below.
Since the secondary is connected at the input ternfnals of an amplifier, it can
be assumed to draw zero current. Now
dt
vr = Rr i ., + LrJi l L
' d t
' r r = MY
dt
Taking the Laplace transform, we get
%,
(s)
_
VuG) R, * s,Lt
sK",
sMlRt
l * I r s
sM
1+
{ , s
Accurate state rrariable models of loaded alternator around an operating point
are available in literature using which optimal voltage regulation schemes can
be devised. This is, of course, beyond the scope of this book.
8.7 LOAD FREOUENCY CONTROL WITH GENERATION
RATE CONSTRAINTS (GRCs)
The l<-racl frcquency control problcm discussed so far does not consicler the effect
of the restrictions on the rate of change of power generation. In power systems
havi ng steam pl ants, power generati on can change onl y at a speci fi ed maxi mum
rate. The generation rate
(fiom saf' ety considerations o1 the equipment) for
reheat units is quit low. Most of the reheat units have a generatiol rate around
3%olmin. Some have a generation rate between 5 to 7jo/o/min. If these
constrai nts arc not consi rl crcd, systertt i s l i kel y to c:ha.sc l argc tttottrcl ttrry
disturbances, Thrs results in undue wear and tear of the controller. Several
methocl s have been proposecl to consi der the effect of GRCs for the cl esi gn of
automatic generation controllers. When GRC is considered, the systeln dynamic
rnodel becomes non-l i near and l i near control techni ques cannot be appl i ed for
the opti mi zati on of the control l er setti ng.
If the generation rates denoted by P", are included in the state vec:tor, the
systerm order will be altered. Instead of augntenting them, while solving the
stare equations, it may be verified at each step if the GRCs are viclated.
Another way of consiciering GRCs for both areas is to arjri iinriiers io ihe
governors
[15,
17] as shown i n Fi g. 8.22, r.e., the maxi mum rate of val ve
opening or closing speed is restricted by the limiters. Here 2",
tr,r,,
iS the power
rate limit irnposed by valve or gate control. In this model
l AYEl . - - gu, nr
( 8. 6s)
Automatic Generation and Voltage Control
Jffif ------_-----l
E
The banded values imposed hy the limiters are selected to resffict the generation
rate by l}Vo per minute.
I
I.g
9t",
u' t +/
_+(
-t*9r"'--l
Fi g.8.22 Governor model wi th GRC
The GRCs result in larger deviations in ACEs as the rate at which generation
can cha-nge in the area is constrained by the limits imposed. Therefore, the
duration for which the power needs to be imported increases considerably as
cornpared to the case where generation rate is not constrained. With GRCs, R
should be selected with care so as to give the best dynamic response. In hydro-
thennal system, the generation rate in the hydro area norrnally remains below
the safe limit and therefore GRCs for all the hydro plants can be.ignored.
8.8 SPEED GOVERNOR DEAD-BAND AND ITS EFFECT
ON AGC
The eff'ect of the speed governor dead-band is that for a given position of the
governor control valves, an increase/decrease in speed can occur before the
position of the valve changes. The governor dead-band can materially affect the
system response. ln AGC studies, the dead-band eff' ect indeed can be
significant, since relativcly small signals are under considerations.
Tl Le speed governor characterri sti c. though non-l i rrear, has been approxi nraaed
by linear characteristics in earlier analysis. Further, there is another non-
iinearity introduced by the dead-band in the governor operation. Mechanical
f' riction and backlash and also valve overlaps in hydraulic relays cause the
governor dead-band. Dur to this, though the input signal increases, the speed
governor may not irnmediately react until the input reaches a particular value.
Similar a.ction takes place when the input signal decreases. Thus the governor
dead-band is defined as the total rnagnitude of sustained speed change within
which there is no change in valve position. The limiting value of dead-band is
specified as 0.06Vo. It was shown by Concordia et. al
[18]
that one of the
effects of governor dead-band is to increase the apparent steady-state speed
regulation R.
A
l -
-
www.MUSLIMENGINEER.info
lFFf
Modrrn Po*., svrt.t Analuri,
The effect of the dead-band may be included in the speed governor control
loop block diagram as shown in Fig. 8.23.Considering the worst case forthe
dead-band, (i.e., the system starts responding after the whole dead-band is
traversed) and examining the dead-band block in Fig. 8.23,the following set of
ly define the behaviourolthe dead.band
[9]-
Speed governor
Dead-band
Fl g. 8.23 Dead-band i n speed-governor control l oop
u(r
+1)
= 7(r) 1:
"(r+1)
_
x, 1 dead-band
-
"(r+l)
_
dead-band; if x('+l)
-
,(r) I
g
-
"(r+1).
tf Xr*l
_
xt < 0
(r is the step in the computation)
Reference [20]
considers the effect of governor dead-band nonlinearity by using
the describing function approach
[11]
and including the linearised equations in
the state space model.
The presence of governor dead-band makes the dynamic response oscillatory.
It has been seen
[9J
that the governor dead-band does not intluence the
selection of integral controller gain settings in the presence of GRCs. In the
presence of GRC and dead band even for small load perturbation, the system
becomes highly non-linear and hence the optimization problem becomes rather
complex.
8.9 DIGITAL LF CONTROLLERS
In recent years, increasingly more attention is being paid to the question of
digital implementation of the automatic generation control algorithrns. This is
mainly due to the facts that digital control turns out to be more accurate and
r cl i qhl c nnr nnaef i n qi ze l ess censi f i ve t o nni se end dr i f t nnd mor e f l exi hl e Tt
r v ^ r E v ^ v t
may also be implemented in a time shared fashion by using the computer
systems in load despatch centre, if so desired. The ACE, a signal which is used
for AGC is available in the discrete form, i.e., there occurs sampling operation
;
between the system and the controller. Unlike the continuous-time system, the
control vector in the discrete mode is constrained to remain constant between
(8.66)
Discrete-Time Control Model
The continuous-time dynamic system is described by a set of linear differential
equations
x = A x + B u + f p
(8.67)
where
f u,
P
are state, conhol and disturbance vectors respectively and A,B
and f are constant matrices associated with the above vectors.
The discrete-time behaviour of the continuous-time system is modelled by the
system of first order linear difference equations:
x( k+1) =
Qx( k) +
Vu( k) + j p&)
(8.68)
where x(k), u(k) and p(k) are the state, control and disturbance vectors and are
speci fi ed at t= kr, ft =
0, 1,2,... etc. and ri s the sampl i ng peri od.
6,
tl ,nd
7
Te the state, control and disturbance transition matrices and they are
evaluated using the following relations.
d=
eAT
{ = ( { r _ l n - t r
j = ( e A r - D A - t f
where A, B and, I are the constant matrices associated with r, ,,LO p vectors
in the conesponding continuous-time dynamic system. The matri x
y'r
can be
evaluated using various well-documented approaches like Sylvestor's expansion
theorem, series expansion technique etc. The optimal digital load frequency
controller design problem is discussed in detail in Ref
[7].
8.10 DECENTRALIZED CONTROL
In view of the large size of a modern power system, it is virtually impossible
to implement either the classical or the modern LFC algorithm in a centralized
manner. ln Fig. 8.24, a decentralized control scheme is shown. x, is used to find
out the vector u, while x, alone is employed to find out u". Thus.
Flg. 8.24 Decentralized control
-
www.MUSLIMENGINEER.info
i,i2[,:*.1 Modern Power System Analysis
-
4
x
-
(x1 x2)'
u t = - k t x t
u 2 - - k z x z
been shown possible using the modal control principle. Decentralized or
hierarchical implementation of the optimal LFC algorithms seems to have been
studied more widely for the stochastic case since the real load disturbances are
truely stochastic. A simple approach is discussed in Ref.
[7].
It may by noted that other techniques of model simplification are available
in the literature on alternative tools to decentralized control. These include the
method of "aggregation", "singular perturbation", "moment matching" and
other techniques
[9]
for finding lower order models of a given large scale
system.
PROB IE I/I S
Two generators rated 200 MW and 400 MW are operating in parallel.
The droop characteristics of their governors are 47o and 5Vo respectively
from no load to full load. The speed changers are so set that the generators
operate at 50 Hz sharing the full load of 600 MW in the ratio of their
ratings. If the load reduces to 400 MW, how will it be shared among the
generators and what will the s)/stem frequency be? Assume free governor
operatlon.
The speed changers of the governors are reset so that the load of 400 MW
is shared among the generators at 50 Hz in the ratio of their ratings. What
are the no l oad frequenci es of the generators?
Consider the block diagram model of lcad frequency control given in Fig.
8.6. Make the following approximatron.
(1 + Z.rs) (1 + Z,s)
=-
t + (7rg + T,),s = 1 + Z"c.r
Solve for Af (l) with parameters giveu below. Given AP,
-
0.01 pu
T"q= 0.4 + 0.5
= 0.9 sec; 70, = 20 sec
K r r K , = 1 ; K p r = 1 0 0 ; R = 3
Coi npare wi th the exact response gi ven i n Fi g. 8.9.
For the load frequency control with proportional plus integral controller
oc ol r n' r n. i - Ti i c e 1n nhf ai n en Asnr Acci nn f nr t ha cf enr l r r cf r f p er r nr i n
c l J o r l v Y Y l l l l ( L L 6 . v . r v t v u L a r r r
cycles, i.".
f' 41t)d
r; for a urrit step APr. What is the corresponding time
t ^ "
, 1 ,
l i r nl * m
error in seconds (with respect to 50 Hz).lComment on the dependence of
error in cycles upon the integral controller gain K,.
Automatic Generation and voltage Control
ffi
I
n,n,, tf^461dv
-
aF(s)
'
1' af (t)dr : liq, *
' 4F(s)
: hm/F(")]
L
JO
, t
JO s- 0 S s+0
8.4 For the two area load frequency control of Fie. 8.16 assume that inte
controller blocks are replaced by gain blocks, i.e. ACEI and ACE are fed
to the respective speed changers through gains
-
K, and
-
Ko. Derive an
expression for the steady values of change in frequency and tie line power
for simultaneously applied unit step load disturbance inputs in the two
areas.
8.5 For the two area load frequency control employing integral of area control
error in each area (Fig. 8.16), obtain an expression for AP6"$) for unit
step disturbance in one of the areas. Assume both areas to be identical.
Comment upon the stability of the system for parameter values given
below:
4e
= 0' 4 sec; Z, =
0' 5 sec;
Kp r = 1 0 0 ; R= 3 ; Ki = l ; b
ar 2= I ; 2t Tr ,
= 0. 05
lHint:
Apply Routh's stability criterion
the system.l
Zp. = 20 sec
= 0.425
to the characteristic equation of
8 . 1
8. 2
REFERE N CES
Books
l . El gcrd, O.1., El ccu' i c Energv.Sv,s/cl r T' l rcorv: An l tttnxl ut' l i on. 2nd cdn. McCraw-
Hi l l , New York, 1982.
2. Weedy, B.M. and B.J. Cory El ectri c Pow' er Systems,4th edn, Wi l ey, New York,
I 998.
Cohn, N., Control of Generation and Power Flou, on Interconnected Systents,
Wi l ey, New Yor k, i 971.
Wood, A.J., and B.F. Wool enberg, Power Generati on, Operati on and Control ,2nd
edn Wi l ey, New Yor k, 1996.
Nagarth, I.J. and M. Gopal , Control Systems Engi neeri ng, 3rd edn. New Del hi ,
200 l .
Handschin, E. (Ed.), Real Time Control of Electric Power Systems, Elsevier, New
York 1972.
Mahalanabis, A.K., D.P. Kothari and S.I Ahson, Computer Aided Power Systent
Analysis and Control, Tata McGraw-Hill, New Delhi, 1988.
Ki rcl rrnayer, L.K., Economi c Control of l nterconnected Systems, Wi l ey, New York,
t 959.
Jamshidi, M., Inrge Scale System.s: Modelling and Control, North Holland, N.Y.,
1983.
a
1
4.
5 .
6.
7.
8.
9.
8. 3
-
www.MUSLIMENGINEER.info