Y+Y+*+m45'

+*+v+T'
+T N *< < < < .*< Y v y v * t * e v * *'*'m< t ' *'7'e *'* *'*'*'*'v F % *'Y *'v *'*'*'r't t *'> t t * 5'T r e ' ' # *'Y r r r 1 v Y F v t F Y v *'F v v T +'t ' t t * *'*

. . . . . . . .
T'Y r'Y T *'e d'*'f e'>'e Y T < Y Y *'*'t Y'r'*'' l A'F v m Yi' + m + + e v z *'r'e

4.4.*.**4::y*+*.4:*+#:*y***+*#*.4*.+#. .%. + v ..+.+....-..+...w*v..++ .w+..*+*.%.+.4*+*. . .s6.....+.+'v%.*..*+*.*.*.w*.++.v+..*.v*.v*.**..*v'v'.+..*.v*...we.'w.v..v . i,;.,z .-.4ke4-*A+.*.*1***.....4*+.. .'w..%..v.v.4 *.# :...*.# # *...+ :.:-.+....*..**4:*.y*.*s. * . . .+ .++*.+%+*.++..+.+*..*.j
. - . .
C hapter 1 7 9
?
@
#
+
4
*
.
@
*
.
5
+.
Y
A
%
:
x
v
*
#
.*
1
y
+
#
*
.
@
z
+
:
T.
*
=
w
1
T
+
.

*
w
v
5
+
.-
*
#
.
5
+
#
*
=
v
M
1
.
+
x
*
@.

+@
***
z
k
+
*
x
2
@
4
+
*+
5
f
*
.
+
*@
*
+
*p
#*
#
*
9*
@

@
u
v
4
*
Y
+
.
vY
*
1
x
*
4
+ + ..4.+.4 * * *.* *.+ 1
.

M e th o d s o f Im P rtlh/irl9 S ta b ility

T he preceding 5 ve chapters have described the physical aspects of the various

categories of pow er system stability, and m ethods of their analysis. A s w ell, key
factors irlfluencing stability problem s and m ethods of m itigation have been covered .
T his chapter discusses special m ethods used for enhancing transient and sm all-signal
stability . T hese tw o categories of system stability have received considerable attention
since the 1960s, and m ethods for their im provem ent are further evolved than those for
other categories.
F or @ given system , any One m ethod of im proving stability m ay not be
adequate. T he best approach I @S likely to be a com bination of several m ethods
J'udiciously chosen so as to m ost effectively assist in m aintaining stbility for different
contingencies and system conditions. ln applying these m ethods to th e solution of -
z

specis c stability problem s, it is im portant to keep in m ind the overall p erform ance of
the P OW CC system . Solutions to the stability PrOblena of one category should nOt be
effected at the expense of another category .
s4any of the m ethods for stability enhancem ent describd 1 @l1 this chapter are
*

options norm ally available for econom lc design of the system . W ith proper design and
application they should greatly contribute to the i exibility of system operation w ithout
com prom ising other aspects of system perform ance.
Som e of the m ethods described are, how ever, som ew hat (theroic''in nature and
@

Can be justifed only 111 speci al situations. n ile im proving system stability, they
im pose duty O n So m e of the equipm ent. T heir application, therefore, has to be based
on a careful assessnlent of the benef ts and costs.

1 1O 3
 1 10 4 M etho d s o f Im P rov in 9 S tab ility C h a P . 17

1 7 .1 T R A N S IE N T S T A B ILIT Y E N H A N C E M E N T

M ethods of im proving transient stability try to achieve One Or nAore of the

follow ing effects'
.

(a) Reduction 1
@11 the disturbing iniuence by m inim izing the fault severity and
duration .

(b) lncrease of the restoring synchronizing forces.

(c) R eduction of the accelerating torque through control of prim e-m over
m echanical POW er.

(d) Reduction of the accelerating torque by applying artifcial load.

The follow ing are various m ethods of achieving these objectives.

1 7 .1 .1 H ig h -s p e ed Fau lt C lea rin 9

T he am ount of kinetic energy gained by the generators during a fault I @S

directly prop ortional to the fault duration ; the quicker the fault I@S cleared, the less
disturbance it causes.
T w o-cycle breakers, together w ith high-speed relays and com m unication, arC
@ *

nOW w idely used 111 locations w here rapid fault clearing IS im portant.
*

ln special circum stances, Cven faster clearlng m ay be desirable. R eference 1

describes the developm ent and application of a one-cycle circuit breaker by
B olm eville Pow er A dm inistration (BPA ).Com bined w ith a rapid response Ov ercu rrent
type sensor, w hich anticipates fault m agnitude, nearly one-cycle total fault duration
is attained. O ne-cycle breakers are not yet in w idespread use. R eference 2 describes
an ultra-high-speed relaying system for E H V lines based on travelling w ave detection.

1 7 .1 .2 R ed u ctio n o f T ra n sm issio n S y ste m R ea cta n ce

T he series inductive reactances of transm ission netw orks are prim ary
determ inants of stability lim its.T he reduction of reactances of various elem ents of the
transm ission netw ork im proves transient stability by l
'ncreasing postfault synchronizing
PoW er transfers. O bviously, the m ost direct w ay of achieving this is by reducing the
reactances of transm ission circuits, w hich are determ ined by the voltage rating, line
and conductor conf gurations, and num ber of parallel circuits. T he follow ing a rC
additional m ethods of reducing the netw ork reactances:

(a) U se of transform ers Ahritlllosver leakage reactances.

(b) Series capacitor com pensation of transm ission lines*

 S ec . 1 7 .1 T ran sie nt S tab ility E n h an cem e nt 1 1O 5

T yp ically , the p er unit tran sform er leakage reactance ranges betw een 0 .1 and 0 .15. F or
n eW er tran sform ers, the m inim um acceptable leakage reactance that can be achieved
Asritllill the norm al transform er design practices has to be established in consultation
sAritll the m anufacturer g3j.ln m any situations, there m ay be a signifcant econom ic
advantage 1 @11 opting for a transform er w ith the low est possible reactance.
@

Series capacltors directly offset the line series reactance. W e discussed the
* *

application of serles capacitors for reactive PoW er and voltage control 111 C hapter 1 1.
A s show n in Section 1 1.2 .8, the m ax im um p ow er tran sfer capability of a transm ission
line m ay be signis cantly increased by the use of series capacitor banks. T his directly
@

translates into enhancem ent of transient stability, dependlng on the facilities provided
for bypassing the capacitor during faults and for reinsertion after fault clearing (see
@

Section 11.2.5). Speed of reinsertion I

*S an im portant factor 1f1 m aintain ing tran sient
stabilit)r (41.Early designs of protective gaps and bypass sw itches lim ited the benet
its
achievable by series capacitor com pensation . H ow ever, w ith the present trend of using
nonlinear resistors of zinc oxide, the reinsertion is practically instantaneous.
A s discussed in Chapter 15 (Section 15.3),one problem w ith series capacitor
*

com p en sation IS the possibility of subsynchronous resonan ce w ith the nearby turbo
alternators. T his aspect m ust be analyzed carefully and appropriate preventative
nAeasures m ust be taken .
T raditionally, series capacitors have been used to com pensate for Very long
overhead lines* R ecently, there has been an increasing recognition of the advantages
of com pensating shorter, but heavily loaded, lines by using series capacitors.
F or transient stability applications, the use of sw itched series cap acitors offers
som e advantages. U pon detection of a fault or pow er sw ing, a series capacitor balzk
can be sw itched in and then rem oved abqut 0.5 second later. Such a sw itched bank
can be located in a substation w here it can serve several lines.
ln reference 5 it I *S show n that, for a given transient stabilit)r ll @m l@t, the
aggregate rating of series capacitors required is less if som e are sw itched th an if all
are unsw itched. T he schem e w ith a portion of the capacitors sw itched reduces the
angular sw ings of the m achines, and this 1 *11 turn reduces P uctuation of loads,
particularly those near the electrical centre.
P rotective relaying is m ade m ore com p lex w hen series com pensation I
*S used,
@

p articularly if the Se rle s capacitors are sw itch ed .

1 7 .1 .3 R e g u lated S h u nt C o m P e nsatio n

Shunt com pensation capable of m aintaining voltages at selected points of the

@ *

tran sm ission system 1m PrOV e Can system stabllity by increasing the tlow of
synchronizing pow er am ong interconnected gen erators. SynchTonous condensers Or
static var com pensators described in C hapter 11 (Sections 11.2.6 to 11.2.8) Can be
used for this PurPOSe.
*11 Section 1 1.2 .8, regulated shunt com p en sation increases th e
A s illustrated 1
m axim um POW er transfer capability of a long transm ission line. T his clearly enhances
transient stability .
 1 10 6 M eth o d s o f Im p ro v in g S tab ility C h a p . 17

17 .1 .4 D ynam ic B raking 16 -10 1

D ynam ic braking uses the concept of applying an arti cial electrical load
during a transient disturbance to increase the electrical pow er Output of generators and
thereby reduce rotor acceleration .
O ne form of dynam ic braking inyolves the sw itching 1 @11 of shunt resistors for
about 0.5 second follow ing a fault to reduce the accelerating POW er of nearby
gen erators and rem ove the kinetic en ergy gained during the fault.B PA has used such
@

a schem e for enhancing transient stability for faults 111the Pacifc northw est g81;the
brake consists of a 1400 M W , 240 kv resistor naade UP of 4 5,000 ft of 1/2 inch
stainless steel w ire strung On three tow ers. R eference 9 provides brief reports of
*

braking resistor applications 11I Japan, C hina, R ussia, and A u stralia.

T o date, braking resistors have been applied only to hydraulic generating
* @ *

stations rem ote from load centres. H ydraulic unlts, ln com p arison to th erm al unlts, are
quite rugged; therefore, they C a n w ithstand the sudden shock from the sw itching in
of resistors w ithout any adverse effect on the units.
lf braking resistors arC applied to therm al units,the effect On shaft fatigue life
@

m ust be carefully exam ined (see Chapter 15, Section 15.4).If the sw itchlng duty I@S
found unacceptable, the sw itching in of the resistors m ay have to be perform ed 1 *11
three Or four step s spread OVer One fullcycle of the low est torsional m ode E101.
T he braking resistors used ttl date are all shunt devices. A lternatively , series
resistors m ay b used to provide the braking effect.ln this CaSC, the Cnergy dissipated
@

IS proportional to the generator current rather than to the voltage. O ne W ay of

@

inserting the resistors 111 series I

@S to install a star-colm ected three-phase resistor
arrangem ent w ith a bypass sw itch in the neutral of the generator step -up transform er
@

to reduce resistor insulation and sw itch requirem ents (101. T he resistor IS inserted
during a transient disturbance by opening the bypass sw itch .
A nother form of braking resistor application that enhances system stability for
unbalanced ground faults only, consists of a resistor connected perm anently betw een
the ground and the neutral of the Y -connected high-voltage w inding of the generator
step-up transform er (301. U nder balanced conditions nO current tlow s through the
@

neutral resistor. W hen llne-to-ground or double line-to-ground faults OCCUF, current

i ow s through the neutral connection and the resistive losses act RS a dynam ic brake.
@

sTitll the sw itched form of braking resistors, the sw itching tlm es should be
based On detailed sim ulations. lf the resistors rem ain connected too long, there is a

possibility of instability On the ttbacksw ing.''

1 7 .1 .5 R ea cto r S w itch in g

Shunt reactors near generators provide a sim ple and convenient DReans of
im proving transient stability . T he reactor norm ally rem ains connected to the netw ork.
T he resulting reactive load increases the generator internal voltage, and this is
r '
..

benef cial to stability. Follow ing a fault, sw itching out the reactof further im proves
stability.
 S e c . 1 7 -1 T ra n sie nt S ta b ility E n h an ce m e nt 1 10 7

1 7 .1 .6 Ind e p en d ent-p o le O p e ratio n o f C ircu it B rea kers

Indep endent-p ole op eration refers to the use of separate m echanism s for each
ph ase of the circuit-beaker so that the three phases are closed and opened
independently of each other. A s a result, the failure of one pole w ill not restrict the
operation of the other tw o poles. A lthough the breaker poles operate independently
of each other, the relaying system is norm ally arranged to trip all three poles for any
typ e of fault.
lndependent-pole operation can be used advantageously at locations w here the
system design criteria l 'nclude a three-phase fault com pounded by breaker failure.
M aintaining system stability for the contingency of a tk ee-phase fault w ith all tlzree
poles of a prim ary circuit-breaker failing to open is extrem ely diff cult. W ith breakers
designed for independent-pole operation, a failure of al1 three poles is highly
im probable. T he use of duplicate relay system s, circuit-breaker trip coils, and
operating m echanism s practically guarantees that at least tw o poles w ill open .
T herefore, the independent operation of the failed breaker w ill reduce a three-phase
fault to a single line-to-ground fault w hen tw o of the poles OP en . T hus, the severity
of a three-phase fault w ith a stuck breaker is signi cantly reduced .

1 7 .1 .7 S in g le -p o le S w itch in g

Single-pole sw itching USWS separate operating m echanism s on each phase; for

single line-to-ground faults, the relaying I*S designed to trip only the faulted phase,
follow ed by Q jt reclosure w ithin 0.5 to 1.5 seconds. F or m ultiphase faults, al1 three
phases are tripped.
*

D uring the period w hen OnC phase IS Open, POW er I @S tran sferred OVCC the
rem aining tAs?tl phases.
A s m ost faults On transm ission lines are of the single line-to-ground type,
*

op ening and reclosing of only the faulted phase results in an im provem ent in translent
@ @

stability Over tk ee-phase trlpplng and reclosing.

Single-pole sw itching is particularly attractive for situations w here a single
@

m ajor line connects tAArtl sy stem s Or w here a single m aJOr line connects a generating
station to the rest of the sy stem . lt m ay also be used On system s w ith m ultiple lines
to im prove system security against m ultiple contingency disturbances g11).
T here are three potential problem s that need to be considered 1 @I1 applying
single-pole sw itching '.

@ Secondary-arc extinction .

@ F atigue duty On turbine-generator shafts and turbine blades.

@ T herm al duty on nearby generators due to negative-sequence currents.

 1 1O 8 M etho d s o f Im p rov in g S tab ility C h ap . 17

S econ daw arc ex tin ction

W hen one phase of a three-phase line is open at both ends, the faulted phase
@

IS cap acitively and inductively coupled to the tw o unfaulted phases, w hich are still
energized (see Figure 17.1).A voltage w ill be induced in the isolated phase because
* @ @

of the capacltlve coupling, and to a lesser extent because of the inductlve coupling .
@

T he m agnitude of the induced voltage IS a direct function of the phase-to-phase and

*

phase-to-ground CaP aC1tances. T he induced voltage m ay su stain th e fault arC for an

extended period follow in g the opening of the phase. T he arC On the faulted phase
conductor after it has been sw itched off is called the secondary Jrc. T he secondary-
arC current I
@S prim arily determ ined by the circuit voltage and the length of the line
section that I*S sw itched out. lt I
*S also ini uenced by the fault location, any shunt
reactors connected directly to the Open phase, and line transposition .

C losed O pen
a

A N x
AA l w
N w
zy j x.x x

M b , r A Q abv+
z
xx1
a h I I I
U
- - - - -
- I I I
I N x 1

M l I --> l
tkkf,c ac trac)C
' tr,c-n,
. > <
.
-) . . I
j
,
I
.

C Cg lC b I
IC ag l
l
Secondary I ?I 1
i , ,
a rC I I
1 l
I

F igure 17.1 Secondary-arc current

T he capacitive com ponent of the secondary-arc current I

@S essentially
independent of the fault location and the load current. T he inductive com ponent of the
current is very dependent on the load currents on the tw o energized phases and on the
location of the fault; it is nearly zero for a fault at the m iddle of the line and is
m axim um for a fault at either end.R eference 12 gives typical values of the secondary-
arc current for 765 kv and 345 kv lines: the capacitive com ponent of the current per
100 m i(160 km ) of line length is on the order of 50 A for 765 kv lines and 30 A for
345 kv lines; the m axim um inductive com ponent of the current is in the range of 10
to 15 A for 765 kv lines w hen the pow er transfer on the tw o energized phases
corresponds to the surge im pedance loading.
In reference 13, the results of arc extinction tests carried out on the A m erican
E lectric P ow er system are com pared Asritll the published results of tests On other
 S ec . 1 7 .1 T ran sie nt S tab ility E n h an cem ent 1 10 9

sy stem s apd laboratory tests. T he general conclusion I

@S that the secondary arC should
self extinguish AAritllill 500 m S if the arC current I
-
@S On the order of 40 A Or less On
lines w ith shunt reactor com p en sation and 20 A or less On uncom pensated lines.
*

F or application s w here the secondary-arc current IS higher than the above

values, it I
*S n ecessary to neutralize the capacitive reactance. O ne m ethod I *S to uSC
shunt inductive reactors. T he four-legged reactor schem e described 1 *1l references 14
and 15 has been used in a num ber of single-pole sw itching application s for reduction
*

of secondary-arc current g11j.ln this schem e, a phase reactor IS connected betw een
* @

each phase and a neutral polnt, and a neutral reactor ls connected betw een ground and
the neutral point. T his m ethod is particularly attractive if shunt reactors are required
to com pensate the norm al line-charging current for voltage control (see Chapter 11,
Section 11.2.3). By appropriate connection, these reactors can be m ade to serve the
additional PurPOSe of reducing secondary-arc current.
A nother m ethod of secondary-arc suppression is to u SC high-speed grounding
sw itches as described in reference 16.T his m ethod is attractive for lines w ithout shunt
reactors.

F atigue #v/y 0n turbine-generator turbine blades and shafts

A s discussed 1
*1lChapter 9 (Section 9.2.3),One of the design considerations of
*

long losv-pressure turbine blades ISthat the frequencies of the fundam ental vibratory
m odes m ust nOt coincide Arritll m ultiples of running speed . Single-pole sw itchin g
during the reclosing dead tim e w hen one phase is open can excite 120 H z torqu e
oscillations. T hese oscillation s m ay resovate w ith the blade natural-vibration m odes
and thus dam age the blades.
A single-pole sw itching sequence im poses successive disturbances On the
turbine-generator shaft system . T he initial disturbance is due to the fault itself, w hich
for a line-to-ground fault is not very severe. W hen the fault is cleared by opening the
faulted phase, a second disturbance is im posed. R eclosing of the opened phase, either
successful or unsuccessful, im poses an additional disturbance. A s discussed in C hapter
15 (Section 15.4), the shaft torsional oscillations resulting from successive im pacts
m ay am plify the tran sien t shaft torques. E ach such incident of single-pole sw itching
op eration m y contribute to a sm all loss of shaft system fatigue life. lf the turbine-
*

generator ls subjected to a large num ber of such incidents, their cum ulative effect m ay
CaUSC shaft failure.
T he effects of single-pole sw itching on the fatigue life of shafts and blades
have been investigated for specil c units in reference 17. T he results show ed no
signif cant loss of life to either the blades or the shaft sections. ln addition, experience
to date has not l
'ndicated any turbine-generator blade or shaft problenAs resulting from
single-pole sw itching (111.H ow ever, it w ould be prudent to study each application
individually before im plem entation .
 1 1 1O M eth o d s o f Im p ro v in g S ta b ility C h ap . 17

Therm al #v@ nearby generators

System unbalance during the period w hen One phase I @S OPCn results 1 *l1
@

negative-sequence currents 111 the stators of nearby generators and synchronous

condensers. T he negative-sequence stator currents induce 120 H z rotor currents that
Cause heating. Standards have been established def ning generator continuous and
*

short-tlm e unbalanced current capability g18j.T he short-tim e cap ability I

*S expressed
in term s ofhlt,the integrated productofm achine negative-sequence phase current (fz)
and tim e (f).The perm issible h2tdepends on the type of m achine as show n in T able
17 .1.

T ab le 17.1 Short-tim e unbalanced current cap ability

T ype of M achine P erm issible 122t

Salient pole generator 40

Synchronous conden ser 30
R ound rotor generators
Indirectly cooled
-
30
D irectly cooled : up to 800 M V A
-
10
-

D irectly cooled: 800-1600 M V A IO-gIM V A -800)x0.00625)

f2 expressed ill Per unit of rated arnAature current;
@

t expressed 111 seconds.

Studies of specif c applications of single-pole sw itching applied to lines near

therm al units have show n the therm al duty to be not lim iting (for exam ple, see
reference 12).H ow ever, each application should be exam ined individually to be sure
that the negative-sequence heating duty I @S nOt excessive.

1 7 .1 .8 S tea m T u rb in e Fast-v a lv in g

Fast-valving (or early valving, aS it I

*S som etim es referred to) IS a technique
applicable to therm al units to assist 1
@11 m aintaining P OW er system tran sient stability .
@

lt involves rapid closing and opening of Steam valves 111 a prescribed m anner to
reduce the generator accelerating PoW er follow ing the recognition of a SCVCrC
transm ission system fault.
A lthough the principle of fast-valving as a stability aid w as recognized in the
early 1930s g19j,the procedure has not been very w idely applied for several reason s.
A m ong them arC the Concerns for any possible adverse effects On the turbine and
energy supply system .
Since the m id-1960s, utilities have realized that fast-valving could be an
effective m ethod of im proving system stability in som e situation s. A num ber of
technical papers have been published describing the basic concepts and effects of fast-
S ec . 1 7 .1 T ran sient S tab ility E p h an cem e nt 1111

valving (29-231. Several utilities have tested and im plem ented fast-valving On SOCnC
of their units (24-271.

F ast-valving p rocedures

C hapter 9 discussed steam turbine cons gurations for fossil-fuel and nuclear-
generating units. F or illustration of fast-valving applications, 1et us consider a fossil-
fuel generating unit w ith a tandem -com pound single reheat turbine and a nuclear unit.
T he turbine conf gurations are show n in F igure 17.2 . F or thes units, the m ain inlt
controlvalves (CV s) and the reheat interceptvalves (lV s)provide a convenient m eans
of controlling the turbine m echanical pow er. D epending on how these valves are used
to control the steam i ow , a variety of possibilities exists for the im plem entation of
fast-valving schem es.

F rom
boiler C rossover
M SV
CV
J4P ----- JP jp ------ ).P I.P ------ ).P J.P ----- (;

RS
RH IV c ondenser

(a) Tandem -com pound single reheat steam turbine of a fossil-fuel unit

F rom steam M SR M SR M SR
generator
A4SV R SV R SV R SV
CV lV lV 1V

C ondenser

(b) Tgrbine of a nuclear unit

HP high-pressure (turbine) IP interm ediate-pressure (turbine)

=

LP loxv-pressure (turbine) M SV = m ain (inlet) safety valve

R SV reheat safety valve lAJ intercept valve
M SR m oisture separator reheater CV ontrol valve = reheater

F igu re 17 @2 T ypical steam turbine conf gurations

 1 1 12 M eth o d s o f Im P ro v in 9 S ta b ility C h a P . 17

In one com m only used schem e,only the intercept valves are rapidly closed and
then fully reopened after a short tim e delay . Since the intercept valves control n early
*

70% of the total unit POW CC, this m ethod results 1 @11 a fairly signis cant reduction 111
turbine POW eC. A DAOCC pronounced tem porary reduction 1 @11 turbine POW er Can be
achieved through actuation of both control and intercept valves. T he procedure of
rapid closing and subsequent full opening of the valves I@S called m om entary /J1'/-
valving .
111 Son;e situation s w here the postfault tran sm ission system I
*S m uch w eaker
than the prefault one, it is desirable to have the prim e m over pow er, after being
reduced rapidly, return to a level low er than the initial pow er. O ne approach is to
provide for rapid closure of control and intercept valves, follow ed by partial opening
of control valves and full opening of intercept valves. A n alternative approach is to
provide for rapid closure and full opening of intercept valves, coupled w ith partial
closure of control valves. T his procedure w hereby, in addition to a rapid tem porary
reduction , a sustained reduction in turbine p ow er is achieved is referred to as
sustained fast-valving.

Improvem ent of system stability

F ast-valving assists in m aintaining system stability follow ing a Severe fault by

reducing the turbine m echanical pow er. R esults of studies on the effectiveness of fast-
valving in enhancing the stability of individual system s have been reported in
references 27 to 32. slany of these studies also consider other DRCanS of enhancing
sy stem stability such aS high-response exciters, series cap acitor com pensation,
independent-pole sw itching, braking resistors, and generator tripp ing .
G enerally, fast-valving has been found to be an effective and econom ical
m ethod of m eeting the perform ance rCquirem ents of POW CC system s w hose design and
operating criteria require stability to be m aintained for a three-phase fault w ith
delayed clearing because of a stuck breaker. T he transient stability PrOblenA ilz such
@ *

C a SC S IS generally associated w ith a sw ing m ode local to a generatlng station and

having a period of about 0.5 to 1.2 seconds.
A lthough nOt Very w idely recognized, fast-valving could be v ery effective 1 *11
situation s w here instability O C C U CS through a slow interarea sw ing having a period of
@

about 2.0 to 4.0 seconds. T he PrOblena I *S usually caused by generators 111 O n C a re a ,

because of a fault 1 @1l the area, advancing w ith respect to the generators 1 *lI other
interconnected areas.T his causes heavy p ow er i ow s through w eak ties connecting the
a re a s , w hich in turn leads to separation of the areas. F ast-valving of one or m ore units

shritllill the area w here the fault occurs m inim izes the accelerating pow er and hence
reduces the transient POW CC sw ings through the w eak ties.F ast-valving is partl 'cularly
@ @

effective 1 *1 such situatlons, since the periods of the P OW er SW 1n g S are long; tlziS,
therefore, allow s nAore tim e for the reduction in P OW er to be achieved .
T he cost of im plem enting fast-valving I*S usually sm all. H ow ever, 1
@11 vie:v of
the concerns about possible adverse effects on the turbine and boiler/steam generator,
fast-valving should be used only in situations w here other less tGheroic'' DACaSUFCS arC
 S ec . 1 7 .1 T ran sie nt S tab ility En h a nce m ent 1 1 13

not able to m aintain system stability . ln such situations, often the only other effective
*

DACanS of m aintaining system stability IS to u SC generator tripp ing . C om p ared to

generator tripping, fast-valving has the advantage that the unit rem ains colm ected to
the system . A s a result, the total system inertia is not reduced and full or partial pow er
output is restpred w ithin a few seconds. ln addition , the resulting stress on the prim e
m over I*S believed to be signil cantly less Severe. H ow ever, fast-valving I *S nOt aS
effective 1@11 aiding stability aS generator tripping.
T he fast-valving logic Can take several form s, depending on the supplier and
@ *

the intended use of fast-valving. G enerally, such logic AArill contain tw o m aln clrcuits:
One for generating the valve control Sequence and the other for generating an
unloading signal if fast-valving is to be follow ed by a reduced generator Output. T he
valve control Sequence Asrill generate a signal to close the valves at a preset rate, hold
them closed for an adjustable p eriod of tim e and reOPCn them at a preset rate. F igure
@

17.3 show s a typical valve closing- and OPCnIng-SCQUCnCC. T he allow able valve
actuation tim es are ini uenced by equipm ent aS w ell aS system considerations.

V alve
position
F ully open
j i
.
f
- - - - - - - - - - -

I?artially op en
I
l p
I F u lly or

! t
I partially closed
T im e

FI Tz F? L

F1 delay betw een the tim e of initi*

ation and the tim e
w hen the valve begins to close
Fz valve-closing tim e
F3 tim e during w hich the valve rem ain s closed
F4 valve-opening tim e

F igu re 17.3 T ypical valve closing- and opening- SCQUCnCC

T he allilit)r of steam valves to rapidly close and rCOPCn depends On the type
of governor system used. T he electrohydraulic turbine governor system that USCS
@

solid-state electronics and a high-pressure hydraulic actuator system IS capable of

rap id control. A Com m on practice, particularly of N orth A m erican m anufacturers, IS
to u se fast-acting hydraulic valves arranged to dum p oi1 from the spring-loaded
actuating cylinders of the intercept and control valves to facilitate rapid valve closure.
 1 1 14 M eth o d s o f Im p ro v in g S ta b ility C hap . 17

T his perm its com plete valve closure 1@11 0.08 to 0.4 second . R eopening after a fast
closure I @S inherently delayed for approxim ately 0.3 to 1.0 second to ftll()A&? for
restoration of oil to the hydraulic cylinder. T he valve-opening tim e depends On the
size of the hydraulic operating cylinder but is usually in the range of 3 to 10 seconds.
F or SonAe applications, faster valve-reopening m ay be needed for fast-valving than is
* #

used in norm al electrohydraulic governlng Operatlons. T his can be achieved AAritlz the
aid of accum ulators on the hydraulic system (22,331.
ln contrast, som e E uropean m anufacturers tend to use large SCFVO valves w ith
facilities for fast operation ; this achieves reopening tim es of less than 1 second.
F ast-valving can also be applied to units w ith m echanical-hydraulic turbine
governors, but it is less i exible and som ew hat m ore diff cult to im plem ent.
F or fossil-ful-f red units, reheater m etal protection considerations dictate the
intercept valve-opening tim e. T he PrOblenA I @S OnC of reheater tube heating w hen the
steam flow is shut off and subsequently restored. T he tim e w ithin w hich reheater
@

steam ; ow should be reestablished IS typically on the order of 10 to 12 seconds.

F rom a sy stem stability view point, the valves should be closed as rap idly aS
possible. shritll m om entary fast-valving, it m ay also be desirable to initiate valve
reop ening slijlltl)r before $he f rst peak of rotor angle I @S reached. Slosv reopening
@

ln c re a se s the backsw ing, and this could lead to second-sw ing instability (22,291.Ifthe
delay 1
*1I valve op ening is signif cant, it could also accentuate the sw inging of
@

gen erators 111 one area against the rest of the system , and this could lead to instability
(221. F or any specif c system , acceptable valve actuation tim es should therefore be
based on detailed stability studies.
F ast-valving action Can be initiated by POW er load unbalance relays,
acceleration detectors, or relays that recognize severe transm ission faults E6).
R eferences 23 and 33 describe typical fast-valving initiating logics used in the
standard circuitry provided by N orth A m erican turbine-generator m anufacturers. T he
logic is usually a sim ple extension of the schem e used to lim it overspeed and depends
@

on the com parlson of m echanical input pow er and electrical output pow er. T he form er
is m easured by m oniforing the reheat steam bow l pressure. ln order to avoid
unnecessary initiation of fast-valving action, it is usually necess/ry to introduce
additional logic to discrim inate against faults that do not require fast-valving . ln m ost
cases, since rapid closing of the valves is critical, it is im portant to ensure that the
delay introduced by the selectivity of the logic is m inim al.
T he Sim e delay introduced by the feedback control logic, such as detection
based on pow er-load im balances, is on the order of 0.1 second and this m ay be
@

unacceptable 111 SonAe CaSeS. A feedforw ard control logic derived directly from
protective relay sy stem s could reduce the initiating tim e to OnC Or tAh?tl cycles g22j.
T he effectiveness of fast-valving 1 @1I im proving transient stabilit)r I @S
dem onstrated in F igure 17.4. T he f gure show s the rCSPOnSCS Asritlz and w ithout fast-
valving of a fossil-fuel-f red P OW er p lant consisting of tw o 500 M W unl*ts. T he fast-
valving Sequence assum ed closing and full reopening of intercept valves only, w ith
F j=0.1 s, F2 =0.25 s, F3=0.1 S, and F4 =0.85 S. T he units have a thyristor excitation
sy stem w ith pow er system stabilizel-s. T he disturbance considered is a three-phase
S ec . 1 7 .1 T ran sie nt S ta b ility E n h an cem e nt 1 1 15
@

fault on a m aJOf transm ission line close to the station, cleared by prim ary relaying 1 @l1
60 nIS. W ithout fast-valving, the generating units exhibit a dom inant interarea SW l @ng
having a period of about 4 .5 seconds. C on sequently, the fast-valving sequence l *s very
effective in lim iting the peak of the s rst sw ing in rotor angle. H ere, very fast valve
closing and opening tim es have been assum ed. W ith slow er valve-opening tim es, the
fast-valving sequence, as show n in reference 29, results in second-sw ing instability .

l30

1 10 / N -w W ithout fast-valving
I y
O
* I e
O : l
w'
= ; l
o 90 / l
= , l
Q ; l v w
*- y' j z# xx
2 70 'z I ,
; xxx
x I ! , x z

- '--3
(::) , %s
= l , $.
l ; w ith j-ast-valving
30 A1 #z
N .Z

10
0 1 2 3 4 5 6 7 8 9 10

T im e 1
@11 seconds

12
1

% 11 ! echanical pow er
m l
! n wl
.th fast-valv l
.ng
l l l /
V
> 4g !j j
l !j l
s . o o
. . o

u a yy m m . m m m m m
l l l ,: l 1: h% %...de*
N p................-
m m m m m

x l l # l . # s ... ,... .
I ! J l *zx ; . . ..
.....
c
c 9 l %# ; V N**V
I # N.
: J ''x
Q l # E jegtyj
*gaj gow gr
l . .

> 8 l w lthout fast-valv lng

%
=
1 #
: # . .
*-
7 1l E lectrlcal pow er
t l
., w1 .tj, s st-valv ing
/
m 6

5
0 1 2 3 4 5 6 7 8 9 10

T im e 1
*11 seconds

F igu re 17.4 E ffect of fast-valving on the stability of a fossil-fuel-

f red station exhibiting a slow interarea sw ing
 1 1 16 M eth o d s o f Im P ro v in 9 S ta b ility C h a9 . 17

P lan t con sideration s

F ast-valving im poses a relatively SCVCrC tran sient On the turbine and steam
generator sy stem . T here are several potential problem areas that m ust be considered
in the application of fast-valving. T he follow ing is a brief discussion of the equipm ent
considerations that m ust be addressed.

Fossil-fuel-/ rc# unitsteam generator considerations:

C losing the intercept valves results in a tem porary increase in reheater Pressure
that could cause the reheat safety relief valves to lift.T he safety valves m ay not reseat
@

properly follow ing this operatlon . A lso, any noise im pact on the environm ent resulting
from the escap ing steam m ust be addressed. C losing both the control and intercept
valves reduces the possibility of a lifting of the reheat relief valves but this m erely
transfers the problem to the m ain relief valves.
A s noted earlier, reheater tube protection I @S a lilzAitillj factor that determ ines
how SOOn the steam flow should be estored to the reh eater. F or a typical reheater
design, tlle lillzitill!j tilzze is on the order of 1 1 seconds. If pow er-op erated relief valves
arC utilized On the reheater outlet and autom atically reop ened, the problem I *S
@

m inim ized 1*11 any Occurance w here only the lntercept valves are closed.
Sustained fast-valving appears to the boiler as a partialload rejection.W ithout
a turbine Or superheater bypass system , the sudden closure of turbine inlet valves
requires venting Of the throttled steam to the atm osphere. If this I @S done by D3eans of
spring-loaded safety valves, it m ay be follow ed by an Outage required to repair these
valves. T herefore, pow er-assisted pressure relief valves and/or a turbine bypass should
be used. U nder optim um conditions, how ever, a once-through boiler can be run back
quickly enough to w ithstand sustained fast-valving w ithout lifting the safety valves;
this W aS dem onstrated by TV A tests On the Cum berland units (261.
W ith a drum -type boiler, if a bypass system is not used, sustained fast-valving
C a u se s a reduction of drum level. T his could result in tripping of the unit by low

drum -level protection that is provided to prevent overheating of the furnace f rew alls.
lm provem ents 1*1I control of feedw ater, druna level and rate of reduction of fuel input
to the furnace and the use of pow er-operated relief valves should enhance the ability
of drum -type boilers to cope w ith sustained fast-valving .

N uclear unit steam g enerator considerations.'

C om pared to fossil-fuel-f red units, nuclear units have sm aller reheater tim e
co n stan ts, and hence closing of only the intercept valves results 1 @11 a faster rise 1
@11
reheater PreSSurC. R eheat safety valve lifting is therefore m ore likely in a nuclear unit.
In addition, the probability of a dam aged safety valve seat is higher because less
*

superheat 111 the steam Can lead to w ater-cutting during lifting of the safety valves.
T he reheater pressure rise can be of greater concern if large bursting disk s, rather than
safety relief valves, have been used in the reheat system . T he bursting of a disk
 S ec . 1 7 .1 T ran sie nt S tab ility En h an ce m ent 1 1 17

essentially takes the unit out of service and adds to the system up set.
T he closing of both the control and intercept valves has the SanAC effect aS for
the fossil-fuel unl
@ts; that 1
@S, it PaSSCS the PrOblenA On to the m ain safety-relief valves.
@

T he c1osing of the control valves results 111 a rise 1 @11 steam generator Pressure, and
steam m ust be bypassed either to the condenser Or to the atm osphere. T he boiler
pressure-control sy stem USCS the turbine Or the bypass system to control the boiler
PFCSSUrC.

Turbine considerations..

The integrity of control-valve (CV ) Seat and stem I*S of som e Concern. T he
@

dum ping of the actuator tluid results an accelerated closing of the valve that m ay
111

lead to valve-bouncing or stem -bending. O n existing units, it m ay be n ecessary to add

nAore solenold dum p valves to achieve the fast-closing tim e for the C V s.
T he com bined action of the control and intercept valves necessitates a review
of the extraction steam system to determ ine if there is a likelihood of actuating any
start-up bypass valve from the H P turbine to the condenser. T he sustained fast-valving
cycle m ay therefore require the addition of a tim e delay 1
*11 the control logic for this
bypass valve.
T he decay of pressure dow nstream of the intercept valves m eans that the valve
actuators m ust be capable of reopening the valves against a high-differential pressure.
A s in the case of the control valves, the control system m ust have fast-acting dum p
solenoids and an adequate-control fluid supply for resetting the valves w ithout
illitilttillj a low -control oil pressure trip .
T he intercept valves are larger than the control valves; thus the effect of higher
impactloading on the devices used for stopping the valves atthe end of their stroke
m ust be evaluated. A voiding a full closure w old be preferable.
ln the Case of a nuclear unit Arritll a m oisture sep arator reheater, higher-than-
*

norm al steam velocities arC created in part of the tubing; the result IS higher stres7 On
the tubes. T he tubing response to this excitation m ust be evaluated.
T he rapid decay of the pressure in the L P turbine raises the possibility of steam
S ashing in the feed-w ater heaters and hence a w ater induction incident that w ould
*

result 111 the feed train xbeing closed by the protection system .
T he turbine-generator rotors m u st n Ot suffer a significant loss of life from the
torsional stresses im posed on them by a fast-valving event. R esults of sim ulations of
specis c applications have show n that the resulting shaft stresses are acceptable (241.
R eference 34 describes fast-valving tests On tw O units at the P ickering-B
nuclear station 1 @11 O ntario. T he tests revealed nO obvious adverse effects On
equipm ent. F rona the plant viesvpoint, fast-valving appeared to be a nonevent. F rona
the system view point, the fast-valving sequence w as found to be effective in causing
a rapid and substantial reduction of turbine pow er.
 1 1 18 M eth o d s o f Im P ro v in g S ta b ility C h aP . 17

1 7 .1 .9 G e n e rato r T rip p in g

T he selective tripping of generating units for Severe transm ission system

contingencies has been used RS a m ethod of im proving sy stem stability for nAany
* *

years. T he reJ'ection of generation at an approprlate location 1l1 the system reduces

POW CC to be tran sferred over the critical transm ission interfaces. Since generating units
can be tripped rapidly, this is a very effective m eans of im proving transient stability .

(See Chapter 13, Section 13.6 for illustration.)

H istorically, the practice of generator tripping aS a stability aid W aS coni ned
to hydro plants. Such plants are usually renlote from load Centres and there I @S little
risk of dam age to the unit from a sudden trip . Since the 1970s, this practice has been
gradually extended to fossil-fuel-f red and nuclear units aS a m eans of solving Severe
stability problenAs.
T he schem e used for detection of system conditions requiring unit tripping is
@

often an extension of trip circuits from Varlous renAote and local line protections. lf
the faulted line is restored Asritllilzm inutes,the rejected units are brought back on-line
quickly.
U nless special facilities are provided, the rejected unitgoes through a standard
shut-dow n and start-up cycle; consequently, full PoW er m ay n ot be available for
*

several hours. A practlce used by DRany tltilities is to design therm al units SO that,
after tripping, they continue ttl run, supplying unit auxiliaries. T his perm its the units
to be resynchronized to the system and restored to full load in about 15 to 30 m inutes

(351.
T he Sequence of events that follow generator tripping I
@S aS follow s: the boiler
@

@S not tripped) and the sp eed

IS tripped ; turbine controls lim it oversp eed (the turbine I
is returned to near-rated speed; the unit then Operates either at no load or supplies the
unit auxiliary load; the boiler is purged and ref red 1
*11 preparation for reloading; the
@

unit IS resynchronized to the POW Cr system and reloaded at a predeterm ined rate.
E vidently, the tripping of a unit subjects it to sudden changes in m echanical
and electrical loading, AAritlz the associated im pact on the gen erator, prim e m over, and
en ergy supp ly system . W hi le therm al uni ts and their cont rols arC designed to
w ithstand such shocks, there I *S som e possibility that controls w ill not function
correctly . A lso, therm al units are not designed for frequent full-load relect @ l @on;
*

signil cantly increasing the num ber of full-load reJ'ections m ay increase unlt
m aintenance requirem ents and reduce unit availability . T herefore, this type of control
@ *

nAeasure for lm provlng system stability should nOt be used indiscrim inately .
T he follow ing arC the m aJ*Of turbine-generator COnCCrnS:

(a) The overspeed resulting from tripping the generator

(b) Therm al stresses caused by the rapid load changes

(c) H igh levels of shafttorques aS the result of successive disturbances

 S ec . 1 7 .1 T ran sient S tab ility E n h a nce m ent 1 1 19

O versp eed con siderations

T he turbine controls should be capable of full-load rejection w ithout a turbine

trip . A s discussed 1 *11 C hapter 9 (Section 9.2.2), the controls of steam turbines a r e
generally designed to lilznit the overspeed follow ing a full-load rejection to about 1%$
below the overspeed trip level of 110 to 115% of rated speed . P revention of turbine
@

trip IS essential if the unit h@s to Carry its auxiliary load until it is resynck onized to
the p ow er sy stem .
F ollow ing the illitiftl overspeed, the unit Alrill reach a steady-state speed
determ ined by the governor droop characteristic. T he standard 5% regulation w ill
result 1@11 a steady-state speed slightly below 195% of rated speed (corresponding to
a frequency of 63 H z).Continuous operation at this frequency is forbidden as it m ay
dam ag to the low -pressure turbine blades (see Section 9.2.3). O n older units
C a u se

w ith m echanical-hydraulic control (M H C), the operator has to adjust th e sp eed-load

changer to return the unit to rated speed . O n m odern units w ith electrohydraulic
control (EH C), this speed reduction to rated speed is autom atic.
lf frequent generator tripping I
@S required to m aintain system stability, there I
*S
a greater risk of turbine-generator runaw ay because of equipm ent m alfunction . T o
ensure safe and reliable operation of speed controls, greater em phasis should be placed
on preventive m aintenance and on periodic testing of the turbine valves and protecfive
sy stem .

T h erm al stress con sideration s

G enerator tripping im poses a high therm al stress on the turbine because of the
variation of unit outpuf over a w ide range. Prior to the disturbance, the unit Output I @S
relatively high and the turbine m etal tem peratures in the critical parts w ould be near
rated conditions. svhen the unit I*S tripped, the unit Operates for about 20 m inutes at
*

very light load Or nO load;this results ln rapid cooling of the turbine m etals, follow ed
by resynchronization and reloading of the unit. T he turbine m etals continue to cool
during the initial reloading until the steam tem peratures m atch turbine m etal
tem peratures in the critical areas. A n increase in load beyond this point begins to heat
the m etals, the rate of heating being a function of the loading rate and the degree of
cooling that occurred during lijht-load operation . T he parts of the turbine that are
critical from therm al stress considerations are high-tem perature region s of the H P
turbine section .
T he therm al stresses resulting from
generator tripping C/n be lim ited to
@

acceptable levels by appropriate operatlng procedures developed in consultation Nsritll

the turbine-generator m anufacturer.

Shaftfatigue-lfe considerations

W hen a generating unit 111 a llltlltillllit PoW er plant is tripped, al1 units 1
*11 the
*

plant experience SUCCCSSIVC disturbances that can result in high lvels of shaft torques.
 1 12O M eth o d s o f Im P ro v in 9 S ta b ilitY C h aP . 17

A s discussed 1
*11 Chapter 15 (Section 15.4), the torsional oscillations caused by
successive im pacts m ay reinforce the initial oscillations. D etailed sim ulations should
be carried Out to ensure that there are no adverse effects on the shaft fatigue-life from
the com pounding effects of a netw ork fault, clearing of the fault, and unit tripping.
F or the unit that I#S tripped, the electric torque I @S suddenly reduced to a Very low
value. T he rem aining units initially experience a sudden proportionate increase in the
electric torque. lf tw o or m ore units are tripp ed, it is very un likely that they w ill be
tripped at precisely the San;e in stant. T herefore, a m ultiunit tripp ing schem e m ay
im pose a nAore SCVCrC duty .
R eferences 36 and 37 describe a generation and load reJ'ection schem e
im plem ented by O ntario H ydro at the B ruce N uclear P ow er D evelopm ent C om plex
having a total generation cap acity of about 6000 M W . T he schem e, im plem ented aS
for
a reject
stop gapion
nAeasure because of delays 1
*l1acquiring transm ission right-of-w ay provided
of up to four 750 M W units and 1500 M W of custom er load, depending
on total generation at the com plex, the num ber of transm ission circuits in service, and
the type of fault. T he decision to im plem ent the reJ'ection schem e W aS based On a
detailed evaluation of the benel ts and im pacts. T he follow ing factors Nvere given
consideration in the selection of the schem e'
.

(a) The costof displacing locked-in nuclear energy w ith fossil-fuelled generation

(b) The expected frequency of transm ission system

@
faults and the estim ated
num ber of units trlp s

(c) The im pact On generating units w ith regard to risk of turbine-generator

*

run aW ay , lncreased forced and m aintenance outage rates, increased cost

because of additional W ear and tear, and cum ulative loss of com ponent life,
including shaft fatigue-life

1 7 .1 .1 0 C o ntro lled S y ste m S ep a ratio n a n d Lo ad S h ed d in g

C ontrolled sep aration m ay be used to prevent a m aJ0f disturban ce 1

@1I One P art
@

O f an l
'nterconnected sy stem from propagating into the rest of the Sy stem and Catlslng
a Severe system breakup . T he initiating disturbance m ay be the loss of a m aJ@Of
transm ission line (ac or dc) carrying a large am ount of pow er or loss of a signif cant
am ount of generation . T he incipient instability in such cases is usually characterized
@

by sudden changes 111 tie line POW er. If this I @S detected 1@l1 tim e and the inform ation
@

is used to initiate corrective actlo n s, sev ere sy stem up sets Can be averted.
T he im pending system instability is detected by m onitoring OnC Or m OrC of the
follow ing system quantities' . sudden change 1 *11 POW CC ; ow through specif c
* @

tran sm ission circuits, change of bus vo1tage angle, rate of POW CC change, and clrcult-
breaker auxiliary contacts.
U pon detection of the im peding instability, controlled system separation is
initiated by opening the appropriate tie lines before cascading Outages Can OCCLIr. In
 S e c . 1 7 .1 T ra n sie nt S tab ility E n h a n c e m e nt 1 12 1

Son;e instances it m ay be necessary to shed selected loads to balance generation and

load i1l the separated Sy stem s.
ln other instances, only load shedding can be initiated so that transient stability
*

IS m aintained w ithout resorting to system separation .

A n exam ple of relaying used for controlled separation is the A # /A 0 relay On
the tie lines betw een O ntario H ydro and M anitoba H ydro. T he relay m onitors sudden
@

changes 1T1POW CC tlow (2:37)OVerthe tie lines and changes in phase angle (A0) of the
bus voltage at the O ntario end of the tie lines@ T he A 0 elem ent supervises the A #
elem ent. T he settings of the relay are chosen SO that the ties are tripped for
disturbances 1 *11 M anitoba that cau se in stability in O ntario . F or SOEnC sy stem
conditions, tie line tripping I @S accom panied by load shedding to en sure satisfactory
post-sep aration system conditions.
R eference 38 describes a local independent detection schem e developed to
isolate the U .S . southw est PoW er system from the U .S .northw est system for Outages
of the P acif c ac intertie. T he schem e USCS the POW CF rate of change at tw O location s
*

for detection of conditions requiring northw est/southw est separatlon . T his involves
separation of the U tah and C olorado PoW er System s from the A rizona system by
*

Op enlng 5 ve ties, and separation of the N evada system from the C alifornia system
north of San F rarlcisco by opening three ties.
R eference 39 describes a load-shedding schem e used by the U .S . N orthw est
P ow er P ool to m aintain tran sient stability of the W estern System C oordinating
Council (W SCC) system s.

1 7 .1 .1 1 H ig h -s p eed Ex citatio n S y ste m s

Signis cant im provem ents in t/nsient stability Can be achieved through rapid
tem p orary increase of generator excitation. T he increase of gen erator f eld voltage
during a transient disturbance has the effect of increasing the internal voltage of the
@ *

m achine; this in turn increases the synchronlzlng POW er.

D uring a transient disturbance follow ing a tran sm ission sy stem fault and
clearing of the fault by isolating the faulted elem ent, the generator term inal voltage
@

ISlow . T he autom atic vo1tage regulator responds to this condition by increasing the
generator f eld voltage, and thl
t
.
-;s has a benef cial effect On the transient stability. T he
effectiveness of this type of control depends on the ability of the excitation system to
quickly increase the field voltage to the highest possible value. H igh -in itial-resp on se
* *

excltatlon system s w ith high-ceiling voltages are m ost effective in this regard . C eiling
voltages are, how ever, lim ited by generator rotor insulation considerations. F or
therm al units, the ceiling voltages are lim ited to about 2.5 to 3.0 tim es the rated-load
f eld voltage.
F ast excitation reSPOn Se to term inal voltage variations required for
im provem ent of transient stability often leads to degrading the dam p in g of local p lant
m ode oscillations (see Chapter 12,Sections 12.4 and 12.5).Supplem entary excitation
control,com m only referred to aspower system stabilizer (PSS),provides a convenient
m eans of dam ping system oscillation s that enables the high-response excitation system
 1 12 2 M eth o d s o f lm P ro v in g S tab ility C h p . 17

to be used. T he use of high-initial-response excitation system s supplem ented w ith P S S

is by far the m ost effective and econom ical m ethod of enhancing the overall system
stabllity g40j.A s an exam ple of such a system , Figure 17.5 show s a general block
diagram representation of a thyristor exciter w ith an autom atic voltage regulator
(A V R),a PSS,and a term inal voltage lim iter.The function ofthe lim iter is to prevent
the term inal voltage from exceeding a set level of typ ically 1.15 pu . T he effectiveness
of the excitation system in im proving the overall system stability depends on proper
control design and tuning procedures. T his is covered in Section l7.2 .

Vref
V oltage regulator E xciter TG R Ey'm.
.

1 . jo swz
Et 1 Z KA fyd
+s% + 1+XFs
Efmin
Qs

1 + - 1 + xF C
yt 1 v v x
+sF RL = - = 1+ xr
o +
Vs
0
T erm inal voltage lim iter

VSm .

sF . 1+ sF 1 1+ sFa
AOr K svAa l
+ xw v l+ swa 1 + sv4

W ashout P hase lead l)' i,,

P ow er system stabilizer

F igu re 17*5 B lock diagrana of thyristor excitation system Nsritlz P S S

The ini uence of excitation system response on transient stability IS illustrated

@

111 F igure 17.6. T he f gure com p ares the resp on ses of the fossil-fuel-f red plant
considered ilz Section 17.1.8 w ith tw O alternative form s of excitation system : (a) a n
aC exciter w ith diode rectif ers, having a FCSPOnSC ratio of 2.0, and (b) a bus-fed
thyristor exciter w ith a P S S . T he disturbance considered I @S a three-phase fault On a
m ajor transm ission line near the pow er plant, cleared in 60 m s.T he system is unstable
w ith the rotating exciter and I @S stable Ahritll the high-initial-response thyristor exciter.
T he critical fault-clearing tim e Arritlz the aC exciter is 47.5 m s, and 62 .5 m s w ith the
*

thyristor exclter.
1 7 .1 T ra n sie nt S ta b ility E n h a n ce m e nt

180
I
160 A#
- -
A C exciter
' w ith diodes
140 ... Z
*
O e #
O :#

l20 ,
.
#
O ,
Z
= Z
Q
100 z
a Z
80 . . - * *
.
.
Q
Z.
60 ,Z ,-,,'A
3 T hyristor exciter w ith P S S
40

20

0
0 0 .5 1.0 1.5 2 .0 2 .5 3 .0

T im e 1
*11 seconds

% A C exciter
m
X5 g,
' w ith diodes
Z '% ,
-
# l
-
x .p.** X j
&
c
3 ln l!
v

l .% -
Q &
/ 1 ' 1 #

>
1 T hy ristor exciter w ith P S S kv' %.-,'
=

t-

t
/ -
1
m

-
3
0 0.5 1.0 l.5 2 .0 2 .5 3 .0

T im e 1
*11 seconds

F igu re 17 @6 C om parison of transient stability Ahpitlz

an ac exciter and a bus-fed thyristor exciter
 1 12 4 M eth o d s o f Im p ro v in g S ta b ility C h a 9 . 17

1 7 .1 .1 2 D iscontinuous Excitation C ontrol 14 0 1

A properly applied pow er system stabilizer provides dam ping to both local and
interarea m odes of oscillations. U nder large-signal or tran sient conditions, the
stabilizer generally contributes positively to s rst-sw ing stability . ln the Presence of
both local and interarea sw ing m odes, how ever, the norm al stabilizer CCSPOnSC C an
*

allow the excitation to be reduced after the peak of the tirst local-m ode SW lng and
before the highest com posite peak of the sw ing is reached. A dditional im provem ents
*

11I transient stability can be realized by keeping the excitation at ceiling, w ithin
term inal voltage constraints, tlzltil the highest point of the sw ing is reached.
A discontinuous excitation control schem e referred to aS transient stabilip
excitation control (TSEC) has been developed by O ntario H ydro to achieve the above
g40).This control im proves transient stability by controlling the generator CXCI'tation
SO that the term inal voltage I
@S m aintained near the m axim um perm issible value of
about 1.12 to 1.15 pu over the entire positive sw ing of the rotor angle. T he schem e
@

USCS a signal proportional to the change ln angle of the generator rotor,ilz addition to
the term inal voltage and rotor speed signals. H ow ever, the angle signal is used only
during the transient period of about 2 seconds follow ing a severe disturbance, since
it results in oscillatory instability if used continuously. T he angle signal prevents
prem ature reversal of s eld voltage and hence m aintains the term inal voltage at a high
level during the positive sw ing of the rotor angle. E xcessive term inal voltage is
prevented by the term inal voltage lim iter.
F igure 17.7 show s a block diagrana of the discontinuous excitation control
schem e. T he T SE C circuitry I
@S integrated w ith the P S S circuitry . T he speed signal
(A(oJ provides continuous control to m aintain sm all-signal stability under norm al
operation . T he angle signal I
*S derived by integrating the speed signal. T he T SE C
block show n in the f gure is an integrator w ith a w ashout. T he value of FANG I*S such

T SE C circuit
V ngna
V - g,a

K ANG

1+sL xg
S 0

H lim vxpzx
.
ST w 1 + sT 3 1 + sTs .
A, K STAB 1 Y V
+ sw 1 ow 1 . ow
2 < * u'o 2 2 A ='o 2 4
(pu)
P ow er system stabilizer f 1im VSmn

F igu re 17.7 B lock diagrana of T SE C schem e

 S e c . 1 7 .1 T ran sie nt S ta b ility E n h an ce m e nt 1 12 5

that, at the frequency range of interest, the output is proportional to the angular
deviation. T he T SE C is in effect a closed-loop control based on local m easu rem en ts.
*

The relay contact (S) IS closed if the term inal voltage drop exceeds a preset value,
* @ @

f eld voltage IS at P O sltive ceiling, and the speed in crease ls above a preset value. T he
relay contact I#S opened w hen either the speed drop s below a threshold value Or the
@

exclter COCnCS O u t of saturation ; the Output of the T SE C block then decays

exponentially shritll a tim e constant FANG .
T he discontinuous excitation control I @S m ost effective 1 @1l im proving the
*

transient stability of generating units exhibiting low -frequency l'nterarea SW 1ngS.

sv h en T SE C is applied to several generating stations in an area, th e Sy stem
voltage level in the entire area I @S raised. T his increases the POW CF con sunaed by the
voltage-dependent loads 1 @11 the area, thereby contributing to further im provem ent 1@l1
tran sient stability .
T o illustrate the perform ance of T SE C ,W e Once again consider the fossil-fuel-
f red pow er plant used in Sections 17.1.8 and 17.1.1 l to dem onstrate the effectiveness
of fast-valving and high-initial-response excitation system . A s seen from the rotor
angle plots of F igure 17.6, the generators at this plant exhibit a dom inant low -
frequency interarea sw ing . F igure 17.8 show s the responses of the generators w ith and
w ithout T SE C . C learly , the sy stem transient stability is very signif cantly im proved
* *

by T SE C . T he critical clearlng tlm e, w hich is 62 .5 m s w ith a thyristor exciter and

P S S , is increased to 117.5 m S by the discontinuous excitation control.
F igure 17.9 show s an extended sim ulation of the system w ith and w ithout
@ @

T SE C . F or PUrPOSCS of com parison, the response obtalned :4,1t11 fast-valving of

intercept valves as descrl 'bed in Section 11.1.8 is also show n in F igure 17.9. T h e
discontinuous excitation control I *S Seen to be aS effective aS fast-valving for this
particular application .
C om pared to other m ethods of l
'm proving system stability, such as fast-valving
and generator-tripping, T SE C im poses Very little duty on the turbine-generator shaft
and steam supply Sy stem . H ow ever, parts of the PoW er sy stem Asrill exp erience a rise
@

11lvoltage of UP to 15% (depending on the setting of the term inal voltage lim iter) for
1 to 2 seconds. T his excitation coftrol schem e m ust be coordinated w ith other
overvoltage prtection and control functions. It m ust also be coordinated A'/itlz
transform er differential protection to en sure that the increased m agnetizing current
resulting from the elevated voltage level does not Cause this protection to operate.
T he discontinuous control described above for transient boosting of excitation
USCS local intelligence to detect a severe system disturbance condition . ln Son;e
applications, it m ay be necessary to initiate the transient excitation boosting by using
renAote telenzetered signals. R eference 4 1 describes such an application .

1 7 .1 .1 3 C o ntro l o f H V D C T ra n sm issio n Lin ks

e

A s discussed 1
@1l Chapter 10 (Section 10.4), an H V D C transm ission link I
@S
highly controllable. lt is possible to take advantage of this unique characteristic of the
H V D C lilllc to augm ent the tran sient stability of yhe ac system .
1 12 6 M eth od s o f lm p ro v in g S ta b ility C h a P . 17

12 0 # * * + N

Z h%-..- -*-----wx
' N
' x

100 &' *x
X Z -'x. .
*
O ' '-'W ithout T SE C N
Z Y
O 80 zd *N
#Z x
=
'#
.
N
Q / xx
a 60 * *N
X
Q Y

40
Y
N x ''
x s

20 'W lth T SE C

0
0 0 .5 1.0 1.5 2 .0 2 .5 3 .0

T im e in seconds

6 ,
l
1
I
.
> 4 1 zW ithout T SE C
= , ,x + -... t.
,z
1 , v x . - . .

@>A
1 # u h z x
I J # h
I l -- -# *x
; I
j,1 -- hw ..... ..xx
I 1 --
I #
= I I '
1 1
> 0 I'
1
= II
ll x
*ee 11 X
m -

2 11 -W 1
'th T SE C
tl

6
0 0.5 1.0 1.5 2 .0 2 .5 3 .0

T im e 1
@11 seconds

1.2

> W ith T SE C
=
Q Q 1.1
@-
Q
@-<
O

* ee N x .-..>.. x ,e
O 1 e x .-e > e
%
= 1.0 *xx. - # ## Nh*-**-<N > *--.--*
o .
'

> '

W lthout T SE C
0 .9
0 0.5 1.0 1.5 2 .0 2 .5 3 .0

T im e ill seconds

F igu re 17.8 E ffect of T SE C on transient stability

S ec . 1 7 .2 S m all-s ig n al S tab ility E n h an cem ent 1 12 ?

12 0
O rig inal sy stem
100 - '
--tno fast-valving or TSEC)

O

'

O 80 1th fast-
=
vajv l
'ng
Q
*e

a t'% eN .q
60 l! xAq 1! p.
; ..
x
z a N x ...... c..
...
z
Q v
%l < : l - h-.. . ..v .s.
..
.
.
...
oe .. jx
.. . .
v !< v !
.
# .....asxsswz .
,

ly
Vz
v
l ! a
...
A l ; ...a
;r ...
Y 40 t
, tuz'- ,z
l ,u'
. k'
1 #
N

20
W ith T SE C

0
0 l 2 3 4 5 6 7 8 9

T im e 1
*11 seconds

F igu re 17.9 E ffects of T SE C and fast-valving On extended system CCSPOnSC

D uring a transient disturbance, the dc PoW er Can be ram ped dow n rapidly to
reduce generatioi load unbalance of the aC system on both S1 'des. ln sonAe situations,
it m ay be necessary to ran;p UP the dc POW CC to assist system stability by taking
advantage of the short-term overload capability of the H V D C sy stem . F rona the
@ *

vlesvp olnt of the aC system P erform ance, the rapid control of dc p ow er has the sam e
@ @

benef cial effect as generator/load trlpplng. R eferences 42 and 43 provide descriptions

of special controls used in a num ber of H V D C installations for transient stability
@

augm entatlon .
T ransient stability augm entation can also be achieved by controlling the H V D C
converters SO aS to provide reactive pow er and voltage support g44j.

1 7 .2 S M A LL-S IG N A L S T A B ILITY EN H A N C EM EN T

Tihe probl
nsuf em of gm all-signal stability, aS described 1
fclent *11 C hapter 12, I
@S usually
OnC of dam ping of system o scillations. T he uSe of POW Cr sy stem
stabilizers to control generator excitation system s I
*S the m ost cost-effective m ethod
of :nhancing the sm all-signal stability of PoW er system s. A dditionally, supplem ental
stabilizing signals m ay be used to m odulate H V D C converter controls and static Var
com pensator controls to enhance dam ping of system oscillations.
T he controls used for sm all-signal stability enhancem ent should perform
satisfactorily under Severe transient disturbances. T herefore, w hile the controls are
*

designed tlslng linear techniques, their overall perform ante is assessed by considering
sm all- aS w ell aS larce-siznal CCSDOnSCS.
 1 12 8 M eth o d s o f Im p ro v in g S ta b ility C h a9 . 1 7

1 7 .2 .1 P o w e r S y ste m S ta b ilize rs

Section 12 .5 of C hapter 12 developed the th eoretical basis for a P S S . T he

@S to add dam ping to the generator rotor oscillation s. T his I
function of a P S S I *S
achieved by m odulating the generator excitation so as to develop a com ponent of
electrical torque in phase w ith rotor speed deviations. Shaft speed, integral of POW er
and term inal frequency are am ong the com m only used input signals to the P S S .

A lternative @#dN of PSS

(a) Stabilizer based on shaf speed signal (delta-om ega)..

P S S based on shaft speed signal has been used successfully on hydraulic units
since the m id-1960s. R eference 45 describes a technique developed to derive a
stabilizing signal from m easurem ent of shaft speed of a hydraulic unit. A m ong the
im portant considerations in the design of equipm ent for the m easurem ent of speed
deviation is the m inim ization of noise caused by shaft run-out and other causes
(45,46j. The allow able level of noise is dependent on its frequency. For noise
frequencies below 5 H z, the level m ust be less than 0.02% , since signif cant changes
in term inal voltage can be produced by low -frequency changes in the field voltage.
A frequency corresponding to shaft rotational speed and resulting from shaft run-out
is generally the m ost im portant noise com ponent in this range. L ateral m ovem ents of
the shaft of 0.075 cm are typical at points close to the generator guide bearing. Such
low -frequency noise calm ot be rem oved by conventional electric f lters; its elim ination
m ust be inherent to the m ethod of m easuring the speed signal. T his is achieved by
sum m ing the outputs from several pick-ups around the shaft. A t gate positions below
70% , the stabilizing signal is disconnected autom atically by an auxiliary (pallet)
sw itch to prevent excessive m odulation of the field voltage by vibrations generated
in the turbine at partial gate openings.
T he application of shaft speed-based stabilizers to therm al units requires a
careful consideration of the effects On torsional oscillations. T he stabilizer, w hile
dam ping the rotor oscillations, Can Cause instability of the torsional m odes. T his
problenA I*S discussed in C hapter 15 (Section 15.2.1).O ne approach successfully used
*

to circum vent the problem ls to sense the speed at a location on the shaft near the
nodes of the critical torsional m odes (47,481.ln addition, an electronic tilter is used
in the stabilizing path to attenuate the torsional com ponents.
W hile stabilizers based on direct m easurem ent of shaft speed have been used
On DRany therm al units, this type of stabilizer has several lim itations. T he prim ary
*

disadvantage IS the need to uSe a torsional f lter. ln attenuating the torsional

com p onents of the stabilizing signal, the s lter also introduces a phase lag at losver
frequencies. A s dem onstrated in E xam ple 15.2 of C hapter 15, this has a destabilizing
effect On the Ciexciter m o de'', thus im posing a m axim um lim it on the allow able
stabilizer gain . ln m any cases, this is too restrictive and lim its the overall effectiveness
of the stabilizer in dam ping system oscillations. ln addition , the stabilizer has to be
 S ec . 1 7 .2 S m all-s ig n al S tab ility En h an ce m ent 1 12 9

custom -designed for each typ e of generating unit dep endin g On its torsional
characteristics. T he delta-' -om ega stabilizer described next W aS developed to
OVCCCOm C these lim itations.

(b) D elta-p-om qga stabilizer (40,49,Jp7.'

T he principle of this stabilizer I

*S illustrated by the follow ing equation that
show s how a signal proportional to rotor speed deviation Can be derived from the
accelerating POW CC:

A Y eq 1J(hPm-A#e)dt
-
u
(17.l)
w here

M inertia con stant, 2//

*

change 11l m echanical POW Cr input

h P e = change in electric pow er output
A eq =
derived or equivalent speed deviation

The objective isto derive the equivalent speed

*
signal h o eq so that it does not
contain torsional m odes. A s shpw n ln C hapter 15 (Exam ple l5.2), torsional
com ponents are inherently attenuated in the integral of h P e signal. T he PrOblenA is to
nAeasure the integral O f A # free of torsional m odes.
*

I11 Diany applications, the com p onent IS neglected. T his I *S satisfactory,

except w hen changing load On the unit and other system condition s w hen the
m echanical P OW Cr changes. U nder such condition s, a spurious stabilizer Output I *S
produced if h P e alone is used as the stabilizing signal. T his in turn results in transient
oscillations in voltage and reactive pow er.
T he integral of m echanical pow er is related to shaft speed and electrical PoW er
aS follow s'
.

jhpmdt Mar+jhpedt (17.2)

T he delta-# -om ega stabilizer naakes use of the above relationship to sim ulate a signal
proportional to the integral of m echanical pow er change by adding signals
proportional to shaft-speed change and integral of electrical pow er change. T his signal
w ill contain torsional oscillations unless a f lter is used. B ecause m echanical P O W er
changes are relatively slow even for fast-valve m ovem ents, the derived integral of the
m echanical pow er signal can be conditioned w ith a sim ple low -pass f lter to rem ove
torsional frequencies.
T he overall transfer function for deriving the equivalent rotor speed deviation
signal from shaft speed and electrical PoW er nAeasurenAents is given by
 1 13 O M eth o d s o f Im P ro v in 9 S ta b ility C h a p . 17

A# @ ) 7LP e(s)
:
2Q l
s)
us+G(s)jus+AY@) (17.
3)
-

eq

w here (;(s) IS the tran sfer function of the torsional f lter. A realization of this I@S
@

show n 111 the block diagranA of Figure 17.10.

+ + h e eq
Attl G (s) E PSS VS

+ -

1
S e M
s

O 0 =
base frequency, rad/s
M =
inertia coefi cient = IH
S d ldt

F igu re 17.10 B lock diagrarn realization of delta-' -om ega stabilizer

T he delta-' -om ega stabilizer has tw o m aJ@Or advantages OVer the delta-om ega
stabilizer:

1. T he h P e signal has a high degree of torsional attenuation, and hence there I @S

generally nO need for a torsional f lter 1 @l1 the m ain stabilizing path . T his
elim in ates the exciter m ode stability problem , thereby p erm itting a h igh er
@11 better dam ping of system o scillation s.
stabilizer gain that results 1

2. A n end-of-shaft speed-sensing arrangem ent shritll a sim ple torsional s lter Can
be used Ahritll electrical PoW er to derive the m echanical P oW er signal. T his
allow s the use of a standard design for al1 units irrespective of their torsional
characteristics.

(c) Frequency-based stabilizer:

T erm inal frequency has been used aS the stabilizing signal for several P SS
applications. N orm ally, the terrninal frequency signal is used directly as the stabilizer
input signal. ln SonAe CaSCS, term inal voltage and current are u sed to derive the
frequency of a voltage behind a sim ulated m achine reactance so as to approxim ate the
nAachine rotor speed.
S e c . 1 7 .2 S m all-s ig n a l S ta b ility E n h an c e m e nt 1 13 1

ln these system s, aS in the CaSC of speed-based stabilizers, Care m u st be taken

to f lter torsional m odes w hen they are used On steam turbine unl @ts.
T he sensitivity of the frequency signal to rotor oscillations increases as the
external transm ission system becom es w eaker, tending to offset the reduction in gain
from stabilizer Output to electrical torque that results from a w eaker transm ission
sy stem g51j.H ence,the gain of a frequency-based stabilizer m ay be adjusted to obtain
the best possible perform ance under w eak aC transm ission system conditions w here
the contribution of the stabilizer I@S rCquired m ost.
T he frequency signal I *S nAore sensitive to m odes of oscillation betw een large
areas than to m odes inv olving only individual units, including those betw een units
w ithin a pow er plant. T hu s it seem s possible to obtain greater dam ping contributions
to interarea m odes of oscillation than w ould be obtainable w ith the speed input signal.
T he frequency-based stabilizer, how ever, suffers from several shortcom ings'
/
.

1. D uring a rapid transient, the term inal frequency signal Arrill undergo a sudden
*

phase shift. T his results in a spike in the f eld voltage that IS retlected ill the
generator output quantities.

2. T he frequency signal often contain s PoW er sy stem noise cau sed by large
industrial loads such as arc furnaces. ln rnany CaSCS this has prevented the uSC
of frequency aS an input signal.

3. T orsional f ltering is requl'red . H ence, a frequency-based stabilizer has the

Sanle basic lim itation RS the delta-om ega stabilizer.

(d) D igitalstabilizer.
.

D igital versions of som e of the above stabilizers have been developed and are
nOW conAnlercially available (521.
M anufacturers are producing excitation sy stem s w ith com plete digital circuitry .
ln this environm ept, ifthe appropriate inputs are provided,the stabilizer becom es just
another program in the excitation control processor.

Excitation control design (40,531

T he paranleters of the P S S and other elenlents of the excitation System arC

chosen to enhance the overall sy stem stability. Specif cally, the follow ing arC the
objectives of excitation control design '.

@ M axim ization of the dam pin g of the local plant m ode as w ell as interarea
m ode oscillation s w ithout com prom ising the stability of other m odes.

@ E nhancem ent of system transient stability.

 1 13 2 M eth o d s o f Im P rov in S tab ility C h ap . 17

@ Prevention of adverse affects On System perform ance during m aJ@Or system

upsets that Cause large frequency excursions.

@ M inim ization of the Con sequences of excitation system m alfunction because

of com ponent failures.

The procedure used for m eeting the above objectives I

*S illustrated by
considering a thyristor excitation system w hose block diagram is show n in F igure
17.5. The input to the PSS m ay be either the shaft speed deviation (A)r) or the
equivalent rotor speed deviation (h eq) .The term inal voltage transducer circuitry is
represented by tim e constants necessary for f ltering the rectif ed term inal voltage
w aveform . These can usually be reduced to a single tim e constant (Ty ) in the range
of 0.0 1 to 0.02 seconds. O ther tim e constants tk ough to the exciter output, including
any associated w ith the exciter itself, are negligible, and the m ain path can be
represented sim ply by the gain K A.
T he follow ing is a description of the considerations and procedures used for
selection of the variou s paranAeters.

E xciter Idl*n.'

A high value of K A I @S desirable from the viesvpoint of transient stability. A

suitable value of K A is about 200 AAritll nO transient gain reduction (TG R) (40,531.
*

A Com m on lndustry practice I *S to reduce the gain of the exciter at high

frequencies by use of T G R (51,521.Typicalvalues of L and Fs u sed for this purpose
are One and ten seconds, respectively . O n excitation system s w ith large values of TR,
T G R is required for satisfactory operation of the generating unit on open-circuit. W ith
TR on the order of 0.02 seconds, T G R I *S unnecessary for stable Open -cl
@rcul
@t operation
(see closure of reference 53).
T he need to uSC T G R should be based On a careful assessnAent of the overall
sy stem dynam ic perform ance and On how T G R affects selection of other param eters
of the excitation system . R eferences 40 and 53 discuss 1
@C1 dtail the im pact of using
T G R for specif c applications.

P hase-lead com p ensation..

T o dam p rotor oscillations, the P S S m ust produce a com ponent of electrical

torque in phase w ith rotor speed deviation . T his requires phase-lead circuits to be used
to com pensate for the 1ag betw een the exciter input (i.e.,PSS output) and the resulting
electrical torque (see Section 12.5). Tw o tirst-order phase-com pensation blocks are
show n in F igure 17.5. lf the degree of phase com pensation required is sm all, a single
s rst-order block m ay be used. For hydraulic generators w ith low values of J-axis
open-circuit tim e constant (TLo),the phase com pensation required is sm alland phase-
lead circuitry m ay in fact not be utilized .
T he P S S is often required to enhance dam ping of either a local plant m ode Or
 S ec . 1 7 .2 S m all-s ig n a I S tab ility E n h ance m ent 1 13 3

an interarea m ode of oscillation . W hile this m ode receives special attention,the phase
com pen sation should be designed so that the P S S contributes to dam ping over a w ide
range of frequency covering both interarea and local m odes of osci llation .
T he f rst step in determ ining the phase com pensation is to com pute the
frequency response betw een the exciter input and the generator electrical torque, using
a tool such as the M A S S program w hich w as described in C hapter 12 . In com puting
this respon se, how ever, the generator speed and rotor angle should rem ain constant.
T his is because w hen the excitation of a generator is m odulated, the resulting change
1
*11 electrical torque Cau ses variation s 1
*11 roto r speed and angle that in turn affect the
electrical torque. yts Nve are interested only in the phase characteristic betw een exciter
input and electrical torqu e, the feedback effect through rotor angle variation should
be elim inated by holding the speed constant.(See the block diagram of F igure 12 .13
1
*11 Chapter 12.) Therefore, the phase characteristic as a function of frequency is
obtained w ith a large inertia assum ed for the m achine under consideration (say 100
tim es the actualinertia).This ensures thatthe speed and angle do not change over the
frequency range of im portance for stabilizer design (0.1 to 3 H z).
T he required frequency CCSPOnSC of any m achine I *S sensitive to the T hevenin
equivalent system im pedance at its term inals but relatively independent of the
dynam ics of other m achines. It is, therefore, appropriate to assum e that al1 other
m achines act as inf nite buses. T his has the effect of elim inating their dynam ics from
the respon se calculation w hile retaining the correct T hevenin im pedance at the
term inals of the m achine under study . T he resulting phase characteristic has a
relatively sim ple form free from the effects of natural frequencies of the external
m achines.
T he phase characteristic to be com pensated varies to SODAC extent w ith system
conditions. T herefore, a characteristic acceptable for different system conditions is
selected. G enerally, slight undercom pensation is preferable to overcom pensation so
that the PSS does not contribute to the negative synchronizing torque com ponent (see
Section 12.5).A n undercom pen sation by about 10O OVer the entire frequency range
of interest provides the required degree of tolerance to allow for uncertainties in
m achine and system m odelling.

Stabilizing signal w ashout:

T he signal w ashout is a high-pass s lter that prevents steady changes in speed

from m odifying the f eld voltage. T he value of the w ashout tim e constant Fv should
be high enough to allow signals associated Asritll oscillations 1 @11 rotor speed to PaSS
unchanged.
F ronn the viexvpoint of the Gtw ashout function,'' the value of F I @S nOt critical
@

and m ay be anyw here 111 the Cange of 1 to 20 seconds. T he m ain considerations arC
that it should be long els ugh to P aSS stabilizing signals at the frequencies of interest
relatively unchanged, but not SO long that it leads to undesirable generator voltage
@

excursions aS a result of stabilizer actlo n during system -islanding conditions. ldeally,

the stabilizer should nOt respond to system -w ide frequency variations.
 1 13 4 M eth o d s o f Im p ro v in g S ta b ility C h a p . 17

F or local m ode oscillations 1

*11 the range of 0.8 to 2 .0 H z, a w ashout O L? l 5 .

seconds is satisfactory . F rom the view point of low -frequency interarea oscillations,
*

w ashout tim e constant of 10 seconds or higher is desirable, since low er-tlm e constants
result 1*11 signif cant phase lead at low frequencies. U nless this is com pensated for
elsesvh ere, it w ill reduce the synchronizing torque com ponent at interarea frequencies .

T his desynchronizing effect I @S detrim ental to interarea transient stability as it w ill

Cause the arcas to sw ing farther apart follow ing a disturbance.

Stabilizer gdl@n.'

T he stabilizer gain, K svaa, has an im portant effect On dam ping of rotor

oscillations. T he value of the gain is chosen by exam ining the effect for a w ide range
of values. T he dam ping increases w ith an increase 1 @1l stabilizer gain UP to a certain
*

p oint beyond w hich further increase in gain results in a decrease 11l dam ping. ldeally,
@

th e stabilizer gain should be set at a value corresponding to m axlm um dam ping.

@

H ow ev er, the galn is often lim ited by other considerations. hsTitll a delta-om ega
stabilizer, aS a result of the effect of the torsional s lter, the stability of the iGexciter
m o de'5 beconaes an overriding consideration. S/itll a delta-' -om ega stabilizer, exciter
m ode stability is not a problem , and a considerably higher value of gain is acceptable
provided that the phase-lead com pensation has been chosen to provide satisfactory
@

phase characteristlcs OVCC a range of frequencies that includes all dom inant m odes. ln
@

such CaSCS, the m axim um value of the stabilizer gain IS likelY to be lim ited by
@

practical considerations such aS the effect On signal n O ISC .

Stabilizer gain I *1I aS high a dam ping of
*S norm ally set to a value that results 1
the critical system m odets) aS practical w ithout com prom ising the stability of other
* *

sy stem m odes Or Causlng CXCCSSIVC am plis cation of signal noise.

Stabilizer Il.m l-ts..

T he positive Output lim it of the stabilizer is set at a relatively large value 1 *f1
the range of 0.1 to 0.2 Pu . T his allow s a high level of contribution from the P SS
@

durlng large sw ings. slTitll such a high value of stabilizer Output ll *llzl
*t, it I
@S essential
to have a m eans of lim iting the generator term inal voltage to its m axim um allow able
value, typically in the 1.12 to 1.15 pu range. T herefore, a ternainal voltage lim iter I
%S
used aS show n in F igure 17 .5. T o be effectiVe, the lim iter gain K g m ust be very high .
T he terrninal voltage signal, how ever, contain s sm all com ponents of torsional
com p onents. H ence, feedback of this signal to the excitation system through a high
@

galn m ay cau se torsional m ode instability . T herefore, F C and T D

are chosen SO aS to
@

provide high attenuation at torsional frequencies, 1 *11 addltion to ensuring an adequate

degree of lim iter loop stability .
O n the negative side, a lillxit of -0.05 to -0 .1 Pu is appropriate. T his allow s
sufl cient control range w hile providing satisfactory tran sient reSPOn Se. ln the unlikely
event of the P S S Output being held at the negative lim it because of a failure of the
stabilizer, this Ahrill not result 1
O11 a unit triP .
 S e c . 1 7 .2 S m all-s i9 n a I S ta b ility E n h a n ce m e nt 1 13 5

C heck ON selected settings:

T he l nal Stage 1
*11 stabilizer design involves the evaluation of its effect on the
overall system perform ance. F irst, the effect of the stabilizer On various m odes O t%
*

sy stem oscillation s IS determ ined over a w ide range of system conditions by using a
sm all-signal stability program . T his includes analysis of the effects of the P S S on
local plant m odes, interarea m odes, and control m odes. ln particular, it is im portant
ttl ensure that there are no adverse interactions As/itll the controls of other nearby
generating units and devices such aS H V D C converters and SV C s.
A fter checking the P S S perform ance under sm all perturbations, it is im portant
to exam ine its effect On transient stability and long-term stability.
F or system s w ith voltage problem s, the acceptability of the chosen P S S Output
lim its should be carefully assessed. ln sonae situations, it I *S possible for the m achine
term inal voltage to fall below the exciter reference level w hile the speed I *S also
falling . T his can lead to the stabilizer overriding the voltage signal to the exciter,
causing reduced transient CCCOVCCY . lt is im portant to lim it the stabilizer Output to
prevent this.
lt I
@S also im portant to coordinate the perform ance of the P S S w ith other
protection s and controls such aS V /H z lim iters and OVereXC1
'tation/underexcitation
@

protectlon .
R eference 53 provides a detailed account of the application of the above
procedure for the design of the P S S of a large nuclear pow er p lant.

G eneral com m ents ON excitation control design:

T he excitation control system s, designed as described above, provide effective

decentralized controllers for the dam ping of electrom echanical oscillations in pow er
sy stem s. G enerally, the resulting design I @S mauch DAore robust than Can be achieved
through use of other m ethods such as pole placem ent techniques and m ultivariable
state space teclm iques. T he overall approach is based on a know ledge of the physical
aspects of the pow er system stabilization problem . T he m ethod used for establishing
the phase characteristics of the P S S is sim ple and requires only the dynam ic
characteristics of the concerned m achines to be m odelled in detail. D etailed analysis
of the perform ance of the pow er system is used to establish other param eters and to
ensure adequacy of the overall perform ance of the excitation control. T he result I *S a
control that enhances the overall stability of the system under different operating
*

conditions. Since the P SS IS tuned to increase the dam ping torque com ponent for a
w ide range of frequencies, it contributes to the dam ping of a11 system m odes in w hich
the respectl *ve generator has a high participation . T his includes any new m ode that
m ay em erge aS a result of changing system conditions. lt I @S POssible to satisfy the
requirem ents for a w id> range of system conditions Ahritll fixed paranleters; hence,
there has been little incentive to date to consider an adaptive control system .
1 13 6 M etho d s o f Im rov in 9 S tab ility C h a9 . 17

Example of excitation control design

T his exam ple illustrates the application of pow er system stabilizers to a tw o-unit
*

therm al generatlng station . E ach un it has a rating of 488 M V A and is equipped w ith a
thyristor excitation sy stem . T he pow er system characteristics are such that these units exhibit
tw o dom inant rotor oscillation m odes: an interarea m ode of about 0.5 H z and a local
interm achine m ode of about 2.0 llz.The objective of excitation control design is to enhance
the transient as w ell aS sm all-signal stability of the pow er system .
F or the purpose of com parison, W e w ill also exam ine the P erform ance of the tw o
therm al units w hen equipped w ith slow rotating exciters. -

The p ow er system :

T he PoW er system consists of tw o areas, as show n in F igure 17.1 l.A rea 1 has a peak
load of 805 M W and is supplied by the tw o-unit therm al plant (G 1 and G 2) for w hl *ch W C are
designing the excitation control. A rea 2 has a peak load of 360 M W and is supplied by seven
sm aller pow er p lants.
T he param eters of each of the tw o 488 M V A units of area 1 1
*11 Per unit on its rating
are aS follow s'
.

Xd 1.8 1 Xs 1.76 & 0.16 X; 0 .3 X ' 0 .6 1

,
# #// 0 .2 17 X , 0.2 17 R 0 .002 FJo 7.80 S T qo 0.90 S
T $'
T #0
'' 0.022 S o 0.074 S H 3 .53 Ko 0

kq#!80O 0'97z 1o
- - -

> j.OZ OO
- 0 .6Z 172 .30 3 >-
G 1 G9
T herm al p lant 0 951 20
.

62 0.82Z 1
> 8
> ---0 .6Z 172 .30 4

1.0Z 0O 0 gz lo
*

>

G5
A T ie lin e p 7?g -

2o
rea 1 0* 3 lZ 9 O - '-
P eak load : 805 M W G6 G7
: 370 M V A r

slode shape

lnterarea m ode

L ocal interm achine m ode

F igu re 17.11 T xvo-area sy stem

S ec . 1 7 .2 S m all-s ig n a l S tab ility E n h ancem ent 1 13 7

Thy ristor excitation system :

T he block diagrann of the thyristor excitation System is show n l

@n F igure 17 .5 . T he
param eters of the exciter and A V R are aS follow s:

K A = 2 l2 Ta 0 .0 1 Efmax 7.77 :5 in - 6 .86

A h igh exciter gain of 2 12 (w ith no transient gain reduction) I

*S used to ensure good transient
stability perform ance.

Slow rotating excitation system ..

T he alternative excitation System considered is a self-excited dc exciter w ith the

follow ing param eters (see block diagram of Figure 8.40,Chapter 8):

KA 19.2 L 0. 1 Fs 0.65 Fc 0 Fs 0
W BX 0.46 B BX 0 .19 KF 0 .05 Fs 0.7 TR 0.02
Va m x 1.15 Va u lx - 1.15 Rc 0 Xc 0

Sm all-signal stability perform ance:

T he frequencies and dam ping ratios (()ofthe rotor angle m odes associated w ith units
G 1 and 6 2 are sum m arized in T able 17.2. T hese results w ere com puted by using the M A S S
program w ith a detailed representation of all generating units. Peak system load conditions
w ere considered and a m ajor transm ission circuit in area 1 near the generating station w as
assum ed to be out of service. T he netw ork w as represented by 83 buses, w ith the active
com ponent of loads m odelled aS 50% constant current and 50% constant im pedance, and tlle
*

reactlve com ponent of loads m odelled as constant im pedance.

W ith thyristor exciters (no PSS),the interarea m ode has a frequency of 0.55 H z and
isjuststable w ith a dam ping ratio of 0.006; the local interm achine m ode has a frequency of
1.823 H z and a dam ping ratio of 0.049.T he m ode shapes are show n in F igure l7 .l 1.
W ith the rotating exciters, the frequencies of the tw o m odes decrease slightly and the
@

dam ping ratios lncrease slightly . T he m ode shapes are essentially sim ilar to those w ith
thyristor exciters.

T able 17.2 R otor angle m odes of units G 1 and G 2

L ocal lnterm achine M ode lnterarea M ode

T ype of E xciter
Frequency ( Frequency (
(a) Thyristor (no PSS) 1.823 H z 0.049 0.550 H z 0.006
(b) R otating exciter 1.793 H z 0.075 0.498 H z 0.046
 1 13 8 M eth o d s o f Im P rov in g S ta b ility C h aP .
17

Selection of PSS param eters for the thyristor exciter:

T he basis for the selection of the phase-lead com pensation I

*S illustrated in F igure
17.12 . C urve 1 of the G gure show s the phase l1t!4 betw een the exciter input and the generator
electrical torq ue as a function of frequency . T his characteristic w as com puted using the M A SS
Program w ith G 1 and G 2 represented in detail as a single equivalent generator having a large
@ *

ln ed la, and the generators at al1 other generating stations as infin ite buses. C urve 2 of the
G gure show s the phase-lead com pensation provided by choosing the follow ing P S S param eters
(see Figure 17.5):

F1 0.06 F2=0.02 F3 1.5 L = 4.0 FW 7.5

T he effect of supp lem enting the thyristor excitation sy stem of units G 1 and G 2 w itla
the P S S on the sm all-signal stability is show n in T able 17.3 for different values of stabilizer
gain K su a. T he dam ping of both m odes of rotor oscillation increases as K su a is increased .

A gain of about 30 is considered satisfactory .

O ther param eters of the PSS (including the term inal voltage Iim iter) that affect the
large-signal perform ance are chosen aS follow s'
.

F 0.2 l7- i,, -

0.05 Vls 1.15 KL 17

Fc 0.025 TD l.2 12 T = 0.0 1

40

O 30
1
O
=

*-

2 20 -*--***
X e
Q e e

. *
e e
e e
% - #
d - *
= # #
# #

m
10 ee #
# #

##e# 2
#

N'
##
##
/#
#z

0 d
0 0 .4 0 .8 1.2 l.6 2 .0
F requency in H z

P hase lag to be com pensated

P hase com pensation prov ided

F igu re 17.12 D eterm ination of phase-lead com pensation

S ec . 1 7 .2 S m all-s i n al S tab ilily E n h ance m ent 1 13 9

T ab le 17.3 E ffect of P S S gain on rotor angle m odes

L ocal Interm achine M ode lnterarea M ode

K STAB
Frequency ( Frequency (
0 l.823 H z 0 .049 0.550 H z 0 .006
20 2 .079 H z 0 .156 0.547 H z 0 .087
30 2 .2 18 H z 0.197 0.54 8 H z 0 .124
40 2 .366 H z 0 .22 7 0 .533 H z 0 .156

Transient stability perform ance..

T he results of transient stability sim ulations for a contingency involv ing a three-phase
fault On a circuit close to the therm al station in area 1 are show n in F igures 17.13 and 17.l4 .
@

@S applied w hen tim e IS equal to 50 cycles and cleared 5 cycles later by o pen ing
T he fault I
both ends of the faulted circuit sim ultaneously . T hree types of excitation control a re
considered for units G 1 and G 2 : the thyristor exciter w ithout and w ith PSS (?t7STAB =30), and
the slow dc exciter.
T he results show n in F igure 17.13 are for the peak system load conditions. T he system
is transiently unstable w ith the rotating exciter.lt is stable w ith the thyristor exciter (110 PSS),
but poorly dam ped. T he P SS, in addition to significantly increasing the dam ping of system
oscillations, im proves transient stability perform ance by reducing the f rst rotor sw ing.
T he resu lts show n in F igure 17.14 are for sy stem cond itions w ith 94 .5% of peak Ioad
and the output of the therm al station in area l proportionately redu ced . W ith the rotating
exciters, units G 1 and G 2 are in this case G rst-sw ing stable but becom e unstab le in the second
sw ing. W ith thyristor exciters, the stability perform ance is sim ilar to that for the peak load
case.C learly,thyristor exciters w ith the P SS result 1
@11very good overallstability perform ance.

Selection of PSS location

ln large sy stem s, the selection of units On w hich to in stall the P S S to dam p

@

interarea oscillatlons m ay nOt be readily apparent. A lthough the principles of P S S

design for dam ping of local and interarea m odes are sim ilar, the m echanism s by
w hich a P SS contributes to the dam ping of the tAArtl types of oscillation are different.
A P S S adds dam ping to an interarea m ode largely by m odulating system loads,
w hereas the perform ance of the P S S Nhritll regard to a local m ode I @S only slightly
@ *

affected by the load characteristics (54,551. U nderstanding these m echanlsm s IS

essential to the effective application of the P S S .
Participation factors corresponding to speed deviations of generating units are
very useful for initial screening of generating units on w hich to add stabilizers.
H ow ever, a high participation factor is a necessary, but not a suff cient, condition for
a P S S at the unit to effectively dam p interarea oscillations. F ollow ing the initial
screening based on participation factors, a m ore rigoroup evaluation using residues and
frequency responses should be carried out to determ ine appropriate locations for the
stabilizers (541.
1 14 0 M eth o d s o f lm P ro v in g S ta b ility

*
O l80 l
l
l
O l
=
l
@e l
l2 0 Il
l
Q
1
l
: N
% ' * x
l z & # .w
60 A I zz & x zz X w e# h- -* . --- -
% # N F
O 1 #
> % #
1
'J N

00 2 4
6 8 10
T im e in seconds

1.1 rx
> z N h #.l -hx
= x > >

Q 10. .X ll !v
O
.. ;

0.9
=
>
0.8
Q
@-
Q
0.7

O @6 -

0 2 4 6 8 10
T im e ill seconds
8 l
l
I
I

I
I
>
= 6 I
Q I
*- I
I
I
1

= 4 l
I
> :g@*%k <
= # z -h h e ... . ... . . . . . . .
# N h-- ..
@-

m
2 # Nh
A :
;% #
1: % #

00 2 4
6 8 10
T im e in seconds

T hyristor exciter T hyristor ex citer R otating exciter

w ithout P S S w ith P S S eeeeeeelw ithout P S S

F igu re 17.13 R esponse of unit G 1 to a f ve-cycle

three-phase fault; peak load conditions
1 7 .2 S m all-s i9 n a I S ta b ility E n h a n ce m e nt

X
#
O
8
I
#
. .w r

O ## % #,
zr * e
= # * #
J* 1 t*
s 1 .
Q
@- < #
F *
* >
*
@ * *

# * @
# K #
2 #
*
*
%
*
#

K @
* *
Q %
*
I
@
l*
% Y l :
*
1 X 4 >* N
% x I I x .-> . -.> ...
1 J * * Z .
* -# *
1 # N *A l
Q % # R J
'
lli
.
- -
i
II
1
'' .
t
N .
.
.
*
O
> *
*
*
@
@
@

'Jd % #
* @
I J
v
O

0 2 4 6 s 10
T im e 1
@11 seconds

# *%
l.l zx .# N .
# x ->h z# x.1 x x . a* j:
* . .
> N e h . . x .. .. a ....> > g
* *
=
#d # 1 1 J :1 *# U * T U
.:y x j * l II -
. V v * M
@- #
Q 1.0 qY . ...:#..# ## w,# .* *
. tl
* Y v
.
.
oe
O
X
d 0 .9
=
o
>
0.8
@-

0.7
O
.

0.6
0 2 4 6 8 10
T im e l
@l1 seconds
8 I
1
I
I
I
I
I
I
>
= 6 l
I
I
I
Q
*- 1
I
I
I
I

= 4 I
l
l
1
I
> 1 4......... a
Fy pwk xxMq :sqv
= . ; x <% - z . p. y.w s .
z / N h '.. . - . . - a . . a: . w.. .
# N h . -
*-

m jy , .. .......,
#
.% J
: v

00 2 4 6 8 l0

T im e in seconds

T hyristor exciter e e - e
T hyristor exciter R otating exciter
w ithout P S S w ith P S S eeeeeelw ithout P S S

F igu re 17.14 R esp on se of unit G 1 for a ve-cycle three-phase

fault; 94 .5% peak load onditions N
 1 14 2 M etho d s o f Im p rov in g S ta b ility C h a9 . 17

1 7 .2 .2 S u p p le m e nta ry C o ntro l of S tatic V a r C o m p e nsato rs

ln C hapter 11, the characteristics and m odelling of static Var com pensators
(SV Cs) w ere descrl
'bed5 along w ith the basic principles of their app lication .B y rapidly
controlling the voltage and reactive POW CC,an SV C can contrl'bute to the enhancem ent
of the pow er system dynam ic P erform ance. N orm ally, voltage regulation is the
@ @

prim ary m ode of control, and thls lm proves voltage stability and transient stability .

H ow ever, the contribution of an SV C to the dam ping of system oscillations resulting

from voltage regulation alone is usually sm all; supplem entary control is necessary to
achieve signiticant dam ping. T he effectiveness of an SV C in enhancing sm all-signal
stability depends on the location of the SV C , input signals used, and controller design.

Example of S VC application

T he sim ple tw o-area system show n in F igure 17.15 is used to illustrate the application
of an SV C for the enhancem ent of system stability . T his system is identical to the one used
in E xam ple 12 .6 of C hapter 12 except for slight differences in pow er-tlow conditions because
of additional shunt capacitor com pensation at bus 8. A l1 four generators are assum ed to have
self-excited dc exciters.

System perform ance without an SVC..

F igure 17 .16 show s the system reSPOnSe to a three-phase fault near bus.9 On one of
the lines betw een buses 8 and 9. T he fault is assum ed to be cleared 1
@11 74 m s by isolating the
faulted line. W e See from the results that the system beconnes unstable through groW l @ng
oscillations of about 0.4 H z.

Stability enhancem ent w ith JD S VC :

T he f rst step is to determ ine a suitable location for the S V C .F or th is sim ple systeln ,
an obvious choice w ould be the m iddle of the interconnection betw een the tw O arcas, w here
voltage sw ings are the greatest w ithout the SV C (56,57j.

400 M W

G 1 1 5 6 7 1 10 km 8 110 km 9 10 11 3 G3
2 5 km 10 km 10 km 2 5 km rx
N F .

I -
f, - f9
2 - 4

G 2 vr vr G4
Yx ,2
k< x'
<
'
w z'

A rea 1 A rea 2

F izu re 17.15 A sim ple txvo-area svstem

1 7 .2 S m all-s i9 n aI S tab ility E n h an cem ent 1 14 3

2 50

O
N
200
O >
= O

*-
150
O
X O
Q X 100
d 'm
o
= 2
Y 41
o 50

50 0 2 4 6 8 l0

T im e in seconds

1000

Y
800
Q
@-

t 600
/
=
O 400
>
'J
o

200

00 2 4 6 8 l0

T im e in seconds

D
= 1.50
Q
@>A

O 1.25
=
>
V
.

Q 1.00

0 .75
O
X

0.50
>
X
0 .25

Q
0 .0 0 2 4 6 8 10

T im e in secon ds

F igu re 17.16 Sy stem reSPOnSe to Severe disturbance w ith no SV C

1 14 4 M eth o d s o f Im P ro v in g S ta b ility C h aP . 17

F or large com plex sy stem s, the best SV C location is not obv ious. ln such situations -
@

the PEA LS (described ln Chapter l2) or the V STA B program (described in Chapter l4)m ay
be used to identify the SV C locations. T he bus participations com puted by the V ST A B
Program and the vol tage pad icipation factors com puted by P E A L S SCrVC aS useful sensiti
v i
t y
indices for identifying SV C locations. T ables 17.4 and 17.5 give these indices for the prefault
system conditions. T he V ST A B results show bus 8 to be the best location. P E A L S results
indicate that bus 9 has the highest voltage palicipation factor (sensitivity to susceptance
change), w hile buses 10, 4, 8, and 11 also have large factors. C on sidering that the pow er
transfer m ight be in either direction beh veen the tw o areas, bus 8 w ou ld obv iously be the best
choice.
*

A n SV C com prising a G xed capacitor and a thyristor-controlled reactor (TCR ) IS

considered for enhancem ent of system stability. T he rating of the SV C is assum ed to be 0 to
200 A r (capacitive). Figure 17.17 show s the block diagrana of the SV C, including the

T ab le 17.4 B us pad icipation T ab le 17.5 B us voltage eigenvector lm agllitude

com puted by V ST A B for interarea m ode com puted by PE A L S

B us P ad icipation B V oltage V ector

us M agnitude F actor
8 0 .5 100
9 0 .17 1 1 9 1.0000

7 0.16 13 10 0 .8852
10 0.0668 4 0 .7294

6 0 .06 17 8 0 .6 54 7

l1 0 .0 153 11 0.6 147

5 0.0 138 3 0 .3645

R egulator TC R
Zref l
. 0
1.0

+ E 10 1 + 0.65 , 1
FSVC . 0
Verr 1 + 0.2 , 1 + 0.02 x

0.0
0.0 .

Fixed capacitor
Z Bsvc (pu)

l.0 Q
c

Qs
>

/ Q c = 200 M V A r
Qc = 200 M V A r

F igu re 17.17 B lock diagrana of SV C and voltaae regulator

S ec . 1 7 .2 S m all-s i9 n a I S ta b ility En h an ce m ent 1 14 5

voltage regulator. T he voltage regulator gain is set at 10 to provide a l0% slope in the control
range. T he param eters of the lead-lag block have been selected SO that the voltage regulator
enhances system transient stability (i.e.,high initialresponse).

Sm all-signal stabilip perform ance..

T able 17.6 sunnnaarizes the frequencies and dam ping ratios of the interarea m ode w ith
and w ithout the SV C , for prefault and postfault system conditions. W hile tlztl SV C has
stabilized the interarea m ode, the dam p ing is still Very 1()N5/.

T able 17.6 Effect of SV C On interarea m ode (frequency and dam ping ratio)

N o SV C W ith SV C
Sy stem C ondition
Frequency ( Frequency (
P refault 0.540 H z 0.0064 0.547 H z 0 .0096
P ostfault 0.4 17 H z -0.0228 0.476 H z 0.0 154

S upp lem entary control to im p rove dam p ing ..

T he input signal used for supplem entary control of the SV C should be responsive to
the m odes of oscillation to be dam ped . T his can be determ ined by residues and ob servability
using the M A S S program described in C hapter 12. T able 17.7 gives the residues and
observability factors for various input signals for both prefault and postfault conditions.
@ @

C learly, a good choice for the input slgnal IS the m agnitude of current in the line betw een
buses 9 and 10 . F igure 17 .18 show s the frequency I'CSPOIASC characteristics of the trallsfer
function betw een the SV C input and this current. lt has a high gain at the frequency of
interarea m ode - an indication of good selectivity of this signal.

T ab le 17.7 R esidues and observability factors

P refault P ostfault
Signal
R esidue O bservability R esidue O bservability
A of G 1 -0.2680 -/0.J156
. 0.8738 -0.9522 -70.3651 0.5471
A to of 6 2 -0.2 12 1 -j0.l04
.
0.7025 -0.7733-/0.3347
. 0.4510
A of G 3 0.4588 +70.1121 1.4140 2.6200 +/0.2945
. 1.4140
A ttl of 6 4 0.4064 +/0.0947
. 1.2510 2.4380 +70.2469 1.3140
A P, line 6-7 -0.2286 +70.4914 1.6230 -0.6551S 1.5160 0.8861
A#, line 10-9 0.2122 -70.8560 2.6400 0.5305 -/4.3570
. 2.3550
A: ,line 6-7 -0.0310 +70.2107 0.6375 0.1289 +./1.2020 0.6847
A: ,line 10-9
'
-0.0400 +70.2346 0.7126 -0.0924 +./1.6290 0.8752
.

Af, line 6-7 -0.3157 +70.8618 2.7980 -0.8919 +./3,4640 1.9120

M , line 10-9 0.26 15 -/0.8732
. 2.7290 0.6484 -./3.8810 ' 2.1110
1 14 6 M eth o d s of Im P ro v in 9 S tab ility C h a9 . 17

70

60

> 50
=
Q
*-<

O 40
=
>
t
.

30

20

10

00 0
. 2 0 .4 0.6 0.8 1.0 1.2 1.4
F requency in H z

50

0

O

O
-
50
=
Q
*-

% -
100
=
m

150

200 0 0
. 2 0 .4 0.6 0.8 l.0 1.2 1.4
F requency in H z

F igu re 17.18 F requency response of the transfer function betw een the SV C
*

ihput and the current ln line betw een buses 9 and 10

S e c . 1 7 .2 S m a ll-s ig n a l S ta b ility E n h an ce m e nt 1 14 7

C ontrol design:

Several alternative control design techniques m ay be used for determ ining the control
param eters for SV C S g55j:

@ P ole placenaent technique

* P hase and gain m argin technique

@ H -inG nity techn ique

H ere, w e w ill illustrate the use of the pole placennent technique. F irst,W e Avillprovide a brief
description of the m ethod .
T he pole placem ent technique uses root locus rules to sh ift a pair of dom inant po les
to a nexvly assigned location in the J-p lane. A fter the initial design the controller is tested for
robustness and, if necessary, appropriate m odis cations are m ade. A t this stage, the
nonlinearities m ay be included and the com pensator lim its m ay be Set.
Let the transfer function of the open-loop plant and the controller (com pensator) be
G (s) and H (s),respectively,aS show n in Figure 17.19.

u +E G (y) y

H (s4

F igu re 17.19 T ransfer function of a plant w ith feedback

T he closed-loop transfer function of the com pensated plant I

@S

(;(s)
(7c(x) 1+ (;(s)1I(s)

The zeros of 1+G (s)H (s) arC the poles of the closed-loop system .
ytssunle that the system eigenvalue l is to be
by 1. Since V m ust satisfy the characteristicshihed to a neW location in the J-plane
denoted equation of the closed-loop system ,

H (k 1
o ) G (l
c )
 1 14 8 M eth o d s o f Im P rov in g S ta b ility C hap . 17

T his Can be expressed 1

@11 term s of m agnitude and phase, aS follow s'
.

lS (l 1
o )I IG (l
o )I

arglH lkol) 1800-arg(G (ko))

Thus,the m agnitude and phase ofthe com pensator atV can be calculated from the m agnitude
and phase ofthe plant at V ;that is,from the com plex frequency respon se of the p lant at the
new pole location . T he new pole location is chosen to satisfy the specified dam p ing ratio 1ts .

im aginary part is usually chosen slightly larger than that of 1 .

T he com pensator generally consists of a w ashout and a series of lead and/or lag
functions to satisfy the argum ent equation, as show n below .

ST F 1+s T 1 1 + s T 2n - 1
K # * *

1+s T 1 + s Fa 1 + s T24

T he w ashout is intended to elim inate the dc com ponent or reduce the close-to-dc com pon ent
of the naeasured signal and has a large tim e constant. E ach lead Or l1t!j block com pensation
is lim ited to a m axim um of 60O for practical reasons. T he m ax im um angle, 0N11 that the ith .

block can provide, I

*S given by

1- ai
sin o
1+ ai

w here Jj=F2i/?-2/ 1*
-

T he frequency at w hich this m ax im um OCCUI'S I

@S

1
O

a iT 2 1 -

w hich is usually chosen to be near the frequency of 1 o. F inally, the gain K is chosen to satisfy
the m agnitude equation.
F or the system of F igure l7.l5, the m agnitude and phase of the transfer function
*

betw een the SV C lnput and the current 1 @11 the line from bus 10 to bus 9 are com puted for a
num ber of values of lo. T his IS done for the prefault as w ell as the postfault system
conditions. T he objective is to nd satisfactory values of 10 that result in approxim ately the
Sam e m agnitude and phase for both operating conditions. These values of k)are gi ven in
T able 17 .8.
 S ec . 1 7 .2 S m all-s ig n al S tab ility E n h an cem ent 1 14 9

T ab le 17.8 T he desired location of the eigenvalues and

the value of the sy stem transfer function

Prefault System Postfault System

N ew E igenvalue V alue of T ransfer N ew E igenvalue V alue of T ransfer
L ocation F unction L ocation F unction
-
0 .12 7+/3.544
. 4.262Z 137.80 -0.576+./3.498 4.177Z 132.30

F igure 17.20 show s the block diagram of the supplem entary control.T he output lim its
of the cpntroller are set to provide the desired large-signal reSPOnSe.

0 .1 Fref

I 10 N 0 1 + 0 49 N 2
. ..
. 1 y y
1 + 10 N 1 + 0 .24 N +
err

0 .1 VSVC

1 = m agnitude of current 1
*11 line betw een buses 9 and 10

F igu re 17.20 Supplem entary control block diagrann

T he frequency and dam ping ratios of the interarea m ode w ith supplem entary control
of the SV C a re aS follow s'
.

Prefault: /=0.564 H z (=0.036

Postfault: 8 0.557 H z (=0.163
.

T he transient stability perform ance of the system w ith and w ithout supplem entary
control of the SV C is depicted in F igure 17.2 1. T he disturbance considered is, as before, a 74
m S three-phase fault On OnC of the lines betw een buses 8 and 9. T he SV C w ith voltage
regulation stabilizes tll() system , but the dam ping I
@S PO0r. T he oscillations are w ell dam ped
w ith the supp lem entary control of the SV C .
1 150 M eth o d s o f Im P rov in g S tab ility

l00

Q
80
O >
= X
*

60
a
c O
Q > l
l 'J
w X 40 k. z . . . . - - - - - - -
#
oo ?
o 4>
20

00 5 10 15 20 25 30

T im e in seconds

s00
Y
Q
750 l
1
@-

t l
11 1%
Y 1 I
= 700 l / . . . . . . . . .
I l
> lI .
'J I l l
O
d 650 * l
J 'J
# y

600 0 5 10 15 20 25 30

T im e 1
@l1 seconds

>
= 1.2
Q

O
= 1.l .
1
> 1 14 i
X
. * I *
Q lI
:I l
. Av
1.0 l 1l N- -.. - . ..
# %
% #
I #

X 0.9 '
=
Q
>
X 0.8

A
m 0.7 0 5 10 15 20 25 30

T im e 1
*l1 seconds

SV C w ithout SV C w ith
supplem entary control supplem entary control

F igu re 17.2 1 Sy stem respon se to a severe disturbance w ith and

w ithout supplem entary control of an SV C at bus 8
S ec . 1 7 .2 S m all-s i9 n a I S tab ility E n h an cem e nt 115 1

1 7 .2 .3 S u p p lem e nta ry C o ntro l o f H V D C T rah sm issio n Lin ks

@ *

A s described 111 C hapter 10, the basic controlled quantity 11I an H V D C

@

tran sm ission link is the direct current. N orm ally, the direct current ls controlled by the
rectis er, and the dc line voltage is m aintained near the ratd value by inverter control.
D am ping of the RC system electrom echanical oscillations can be increased by
m odulating the current order at the rectif er. A lternatively , both the current order at
the rectif er and the voltage order at the inverter Can be m odulated.
T he H V D C link m ay either be enabedded 1 @11 an ac sy stem or form an
asynchronous link betw een tw o ac system s. Supplem entary control w ith either of these
typ es of H V D C links is effective in dam ping ac system oscillations. R eferences 42
and 43 describe dc m odulation schem es used in several H V D C transm ission links.
T he follow ing exam ple illu strates the u se of supplem entary control of H V D C
links for the eA ancem ent of system stability .

Example of N FD C tink supplem entary control

T he system used to illustrate the application of IW D C supplem entary contro l is show n

in F igure 17.22 . lt is based on the sim p le tw o-area system considered in the prev ious section
(Figure 17.15) and in Exam ple 12.6. A 200 M W bipolar dc link has been added betw een
buses 7 and 9 in parallel w ith tw o ac ties.The dc llnk is represented as a m onopolar link w ith
a voltage rating of 56 kv and a current rating of 3,600 A . T he dc line resistance is 1.5 f1, and
inductance is 100 m H . The com m utating reactance (# c) associated w ith each converter is
0.57 D . A sm oothing reactor of 50 m H is used at each end of the line.
T he ac portion of the sy stem is very sim ilar to that of E xam p le 12 .6 . T he principal
difference is the reactive support of 125 M V A r prov ided at the rectiG er and inverter buses.

M odelling of converter controls:

F igures 17.23 to 17.28 show m odels used to represent the dc link controls. T he m aster
control, show n in F igure 17.23, determ ines the current order for the rectiG er and inverter. lf

7 9
t
G 1 l 5 e l2 s 13 10 ll 3 G3

M F M

L T -
7 - f9
2 4

G 2 vr vr 64

m
.
w M

A rea 1 A rea 2

F igure 17.22 A tw o-area System w ith parallel dc and aC ties

1 15 2 M eth o d s o f Im p ro v in g S ta b ility C h a P . 17

the rectif er end dc voltage falls below 90% of rated value, the control sw itches to colnstalnt
current m ode. T he hy steresis characteristic represented by block M C 1 prevents huntin
behveen constant PoW er and constant current control m odes. T he current order I@S lim ited by
the voltage-dependent current order lim it (V D CO L) show n in Figure 17.24.

56 .0
M C l
M C2 P ordvr
K a
1.0 --
K H V + fOrder
Vdclr)
gate M C5
(1(V )
g Vdclrj M c4
> a
50.4 53.2 (kV ) M C3

V1dOr
c(de
r)r vr
at oect
co L
iGer order(r) IOrder vo coL ;
I r () at inveder orJe,()
dc
(kV ) (kV )
(r) rectifer (i) inverter

F igu re 17.23 M aster contro l

G
A

B
G F E

H
l I
I I
I
l I
l l
l l
I I

10 1
. 08 kv --- --- o

l
E l
'
F 1.08 kA - I y- Id

I = 0.36 kA

T he dow ntim e constant I@S 5 m S.

T he uptim e constant IS l00 m s.

F igu re 17.24 V oltage-dependent current order lim its

>e C . 1 7 .2 S m a ll-s ig n a l S ta b ility E n h a n c e m e nt

1800
PC R I PC R 2 PC R 3
-

55.7 1
Iorderlr) z 17.78 1+ 1.0 urectyor
s 1 + 0 .00 1N
+

(2t i;:

F igu re 17.25 R ecti er pole control

l.0
PC R 5 PC R 6 PC R 7

Zcctr) at /=0- o i
+ sin5 s n -l a
m in

1
Vaclr)
1 + 0 .005x sin 5O

PC R 4

F igu re 17.26 G min calculation

P C II P C l3

Q -

j 6Xc(i)
l + 0.02N l p c jy

P C I2 +

V 1
dolis 1
+ 0 .02,

Y X J= 0-
20 70 o cos j joo
P C 115 + P C l9 + P C 17
Ay -h -
1 1l co s 1
-

. Z 1.0 cos Z 1.0 a p,.

P C 1l6 p c lg P C 15 p c I6
.
)

00 y in cosl80O

m ax
P C I12 P C 1l 1
P C l14 P C 113
I .
. - 55.7 1
orderli) E : 17.78 1+ 1.0 ainverte
s 1+ 0 .00 1x
- + P C 110
f# Im 1100

F igu re 17.27 lnved er pole control

 1 154 M eth o d s o f Im ro v in 9 S ta b ility C h a9 . 17

1 co sl7O
P C 12 l . 0
P C 122 -> P C 124 P C l25
Vacli) at f=0- - j
+ l -cosl7o X 1.0 co s- Yn,in
1 P C 123
Vacli) 1 yoo
+ 0 .02N COS

P C l20

F igu re 17.28 Ymin calculation

@11 F igure 17 .2 5. T he norm al m ode

T he rectif er pole control representation is show n 1
of rectif er pperation ls cuq ent control. A propod ional plus integral controller (block PCR I)
is used to control the direct line current. T he ignition delay angle is lim ited beh veen Q m in
and 1800. T he loxver lilllit min ensures fhat there is suff cient forw ard voltage aCrOSS the
thyristor for a successful turn-on . T his lim it, as show n in F igure 17 .26, varies w ith the ac line
voltage at the rectiGer (blocks PCR 4 to PCR 7) so that there is alw ay s a m in im um am ount of
forw ard voltage aCrOSS the thyristor.
F igure 17.27 show s the inverter pole control representation . U nder norm al cond itions,
@

the inverter IS on cpnstant extinction angle (CEA ) control represented by

6X CId
-

CO S ? cosy
J Fd0

The control ensures thqt COSG corresponds to a condition Ahritll constant (see C hapter 10,
Section 10.9.3).
@S also equipped w ith a currnt controller that I
T he inverter I @S active only w hen the
firing angle derived from current control I *S sm aller than the f ring angle derived from the
C E A controller. T he inverter sw itches to the current control m ode w hen the rectif er hits Q ml
'n
and is unable to m aintain the desired direct current.
The low er lim itofthe extinction angle (ymin) is varied w ith the inverter ac bus voltage
so th at a m in im um com m utation m argin area is m aintained . F igure l7 .2 8 show s the log ic for
the calculation of Ymin.
If the inverter aC bus voltage drops below 0.5 Ptl, it I *S assum ed that com m utation
failure OCCUrS.T he rectifier and inverter are blocked if the inverter experiences m ore than tw o
com m utation failures.T he inverter recovers follow ing a com m utation fAilure w hen the ac bus
voltage subsequently gOeS above 0 .7 Pu .

System perform ance withoutsupplem entary control..

F igures 17.29 and 17.30 show the system response to a three-phase fau lt near bus 9
On One of the circuits betw een buses 8 and 9 . T he fau lt is cleared by isolating the faulted
circuit ill 83 m s. W e see from the results that the system is transiently stable. H ow ever,
oscillations betw een area 1 and area 2 generators have zero or slightly negative dam ping.
T able 17.9 sum m arizes the results of sm all-signal stability analysis.T he frequency and
dam ping ratio of the interarea m ode are listed for different system conditions. T he interarea
m ode is poorly or negatively dam ped in al1 the cases considered.
1 7 .2 S m all-s ig n al S tab ility E n h an ce m ent

20
X

O 0
% =
Q Q
@-
O

O o
Q -

20
d

l o
2o . k -

40
%
Q
=

60
0 2 4 6 8 10

T im e in seconds

400
/
300
*-

t
/ 2 00
=
O
Q 100
=
O
*-

0
1 z'
.

<
-

100 0 2 4 6 8 10

T im e l
@n seconds

l00
#

@
0
Q
*e

t
Y -

100
=

O
=
*J -

2 00 ,

300
0 2 4 6 8 10

T im e in seconds

F igu re 17.29 Sy stem response to an ac system fault near

the inved er, w ith no supplem entary contro l
115 6 M eth o d s o f Im P ro v in 9 S ta b ility C h a P . 17

75
R ectifler
#ehh

> jM - Z ---------s-.-+-..---%.-.-.----x.-e------x.
lnv ed er
*-

O
25
=
>
c
Q 0

25
0 2 4 6 8 10

T im e 1
@11 seconds

12 5

E
100
Q
=
75

O 50
=
'J
D
O
25

0 0 2 4 6 8 l0

T im e in seconds

75

O 50
=
Q
*-

t
t 25
O
>
Q

0 0 2 4 6 8 10

T im e in seconds

F igu re 17.30 H V D C lilllk reSPOnSe to an aC system fau lt near

the inved er, w ith no supplem entary control
S ec . 1 7 .2 S m all-s i9 n al S tab ility En h ance m ent 1 15 7

T ab le 17.9 F requency and dam p ing ratio of interarea

m ode w ithout supplem entary control

M W F low from A rea 1

C ircuits out of lnterarea M ode
C ase N o . to A rea 2
S ervice
D C Tie A C Tie Freq.(H z) (
1(a) 200 200 N one 0.575 0.0076
1(b) 200 200 8 to 9 (1 cct) 0.495 -0.0054
l(c) 200 200 7
8t
to 8 (1 cct
) 0 440. -0.0 167
o 9 (1 cct)
2(a) 50 352 N one 0.560 0.0052
2(b) 50 352 8 to 9 (1 cct) 0.466 -0.0 ll0
2(c) 50 352 7 t
8 t
o 8 (1 cct) 0 397
. -0 .02 54
o 9 (1 cct)

S upp lem entary control /t? im p rove dam p ing :

T he general procedure for designing the supplem entary control is the sam e as that for
a n SV C using the pole placem ent technique as described in Section l7.2 .2 .

B ased on observability considerations, active pow er through the line betw een buses
7 and 8 IS selected aS the feedback signal. T he initial target for the P lacem ent of the
eigenvalue associated w ith the interarea m ode I@S chosen to be -0.64+/3.t. T h is corresponds
.

to a frequency of 0.5 H z and a dam ping ratio of 0.2.The m agnitude and phase of the OPCn-
loop transfer function beh veen the PoW er order signal of H V D C m aster contrpl and the active
pow er flow in-the line betw een buses 7 and 8 at a com plex frequen cy s = -0.64+/3.l for .

different operating conditions are listed in T able 17.10. T he results show that the phase
@

com pensation required is betw een 85O and l500. A s a com prom ise, the phase colTlpensatloll
IS chosen to be 1000.F igure 17.3 1 show s the block diagrarn of the supplem entary control.

T ab le 17 .10 M agn itude and phase of the open-loop

tran sfer function at s = -0 .64+/3.l.

e- .

C ase N o. O pen-L oop T ransfer F unction

1(a) 0.815Z -146.920

1(c) 0.773Z -86.840
2(a) 0.716Z -142.200
2(c) 0.237Z -112.900
 1 15 8 M eth o d s o f Im 9 rov in g S tab ility C h ap . 17

25 # order
+

10 0N
. 0 1 + 0 .55N 3 +
Xlctline 7-8) 1 + 10 .0N . 25 1 + 0 .20N E T o m aster control

25

F igu re 17.3 1 Supp lem entary control block diagrann

T he w ashout tim e constant is 10 s.T he param eters of the phase-lead block are chosen
tll provide 1000 at the com plex frequency -0.64+73.1. The gain is lim ited to 0.25 so as to
ensure adequate dam ping of al1 system m odes.R elatively large output lim its (+25 M W ) are
used to allow a high level of contribution from the supplem entary controller during large
sw ings.
T he frequency and dam ping ratio of the l 'nterarea m ode w ith the supplem entary control
are sum m arized in T able 17 .11 for the different operating conditions considered in T able l7.9.
T he interarea m ode is w ell dam ped ill 1tl1 CaSCS w ith the supplem entary control.

T able 17.11 F requency and dam ping ratio of interarea

m ode w ith supplem entary control

lnterarea M ode
C ase N o .
Freq.(H z) (
1(a) 0.607 0.1730
1(b) 0.506 0.1572
1(c) 0.467 0.2512
2(a) 0.588 0.1474
2(b) 0.475 0.1007
2(c) 0.422 0.103 1

F igures 17.32 and 17.33 show the transient response of the system w ith supplem entary
control for a three-phase fault near bus 9 on a circuit betw een buses 8 and 9. T he oscillations
are nOW w ell dam ped.
1 7 .2 S m a ll-s ig n al S ta b ility E n h a n ce m e nt

20
X
Q

O 0
< =
Q
2 -
19 o -

20
d o

Q o
o >
y 2
.A
-

40

60
0 2 4 6 8 10

T im e 1
*11 seconds

400
Y
Q
@- A 300
t
/
= 200
Q
m
l
O
*eA
l00

<
0 0 2 4 6 8 l0

T im e in seconds

100

Y
0
Q
@<

t
o
-
100
=
=
O
=
*m -

2 00
O

-
300
0 2 4 6 8 10

T im e in seconds

F igu re 17.32 Sy stem response to an ac system fault ne1

the invel er, w ith supplem entary control
1 16 0 M eth o d s o f Im p rov in g S ta b ility

80

R ectifler
60 -.
> 4 h .. - - . . .. . . . . . . . .. . . . . . . . . .. . . . . . . . . . . . . . . .

Inverter
*-
40
X

>
20
c
Q 0

20
0 2 4 6 8 10

T im e in seconds

12 5

*
O
100
O
=

*-
75

O 50
=
'>
O
O
25

0 0 2 4 6 8 l0

T im e l
*l1 seconds

80

60
O
=

@>A 40

t
t
O
>
20

0 0 2 4 6 8 l0

T im e l
*n seconds

F ig u re 17 .33 H V D C link reSPOnSe to an ac sy stem fau lt

near the invel er, w ith supplem entary control
 R efe re n ces 1 16 1

R E FE R E N C E S

(11 R .O .Berglund, W .A .M ittelstadt, M .L . Shelton,P.B arkan, C .G .D ew ey, and

K .M . Skreiner, (to ne-c ycle F ault lnterruption at 500 kV : System B enel ts an d
B reak er D esign,'' IE E E Trans. V ol. P A S -93, PP . 1240-12 5 1, S eptem ber/
O ctober, 1974 .

(21 J.Esztergalyos, M .T.Y ee, M . C ham ia and S . L iberm an, kky j,e D evelopm ent
and O peration of an U ltra H igh Speed Itelaying System for E H V Ll
*nes,55
C IG R E 34-04, 1978.

(31 C .Lindsay and V .Shenoy, ttReliabilit)? and Selection of Characteristics for

E H V T ransform ers,'' C IG M 12-08, 1978.

(41 G .D .Breuer,H .M .Rustebakke,R .A .G ibley,and H .O .Sim m ons,tt-l-he U se of

Series C apacitors to O btain M axim um E H V T ransm ission C apability,'' IE E E
Trans., V ol. P /tS-83, PP . 1090-1 10 1, N ovem ber 1964 .

(51 E .W . K im bark, itlm provem ent of System Stability by Sw itched Series

C apacitors,'' IE E E Trans., V ol. P A S-85, PP . 180-188, F ebruary 1966.

r6q IEEE Task Force Report,ttk?j D escription of D iscrete Supplem entary C ontrols
for Stability,''IE E E Trans., V ol.PA S-97,pp.149-165, January/F ebruary 1978.

(71 H .M .Ellis,J.E .H ardy, A .L . B lythe, and J.W . Skooglund, tD ynam ic Stability

of the P eace R iver T ransm ission Sy stem ,''IE E E Trans., V o1.P A S-85,PP .586-
600, June 1966.

(8q M .L . Shelton, R .F. W inklem an, W .A . M ittelstadt, and W .L . Bellerby,

B onneville P ow er A dm inistration 1400 A4SV B raking R esistors'' IE E E Trans.,
V ol. P A S-94, PP . 602-6 11, M arch/A pril 197 5.

(9) CIG RE SC38-W G 02 Report, (gtatd of the A rt 1

@11 N on-classical sleans to
lm prove P ow er System Stability,55 E lectra, N o . 1 18, PP . 88-113, M ay 1988.

(101 EPRIReport E L -5859, ET echnical L im its to T ran sm ission Xy stem O p eration ,''
@

F inal R eport of Project 5005-2, prepared by P ow er T echnologles lnc., June

1988.

(111 IEEE W orking G roup , tsingle P ole Sw itching for Stability and R eliability ,''
R eport of a panel discussion held at the 1984 P E S Sum m er M eeting, IE E E
Trans *# V ol. P W R S-I, N o. 2, PP . 25-36, s4ay 1986.
1 16 2 M eth o d s o f Im P ro v in g S ta bn
ility C h a P . 17

(121 R .D .D unlop,R .M .M aliszew ski,and B .M .Pasternack,GtA pplication of Single

P hase Sw itching On the A EP 765-kV System ,51Proceedings of the Am erican
P ow er Conference, V ol. 42, PP . 46 1-462, 1980.

(13) ?Y.J.Fakherietal., U se of R eactor Sw itches in Single P hase Sw itching,''

C IG R E 13-06, 1980.

(141 N .K nudsen, ttsingle-p hase Sw itching of T ransm ission L ines U sing lteactors
for E xtinction of the Secondary ytrc,55 C IG R E 3 10, 1962.

(151 E .W .K im bark,tCSuppression of G round-Fault ytrcs On Single-pole Sw itched

L ines by Shunt R eactors,'' IE E E Trans., V ol. P A S -83, PP . 2 85-290, M arch
1964.

(161 R .M .H asl
kbar A .C .L egate,J.11.B runke, and W .G .P eterson, tcT he A pplication
,

of H igh-speed G rounding Sw itches for Single-p ole R eclosing on 500-kV

P ow er S y stem s,'' IE E E Trans., V ol. P A S-IOO, PP . 15 12-15 15, A pril 198 1.

(171 JV.J. G onzales, G .C . K ung, C . Raczkow ski, C .W . Taylor, and D . Thonn,

GtE ffects of Single-and-T hree P ole Sw itching and H igh-speed R eclosing On
T urbine-G enerator Shafts and B ladesr''IE E E Trans*5 V ol. PA S-103, PP .32 18-
3228, N ovem ber 1984.

(181 A N SI Standard C50.13-1989.

(19) R .C . Buell, R .J. Caughey, E .M . H unter, and V .M . M arquis, ttG overnor

P erform ance during System D isturbances,'' A IE E Trans., V ol. 50,pp .354-369,
M arch 193 1.

(201 F.P.deM ello, D .N .Ew art, M .Tem oshok, and M .A .Eggenberger, (iTurbine
E n ergy Controls A id 1
@11 Pow er System Perform ance,'' Proceedings of the
A m erican Power Conference, V ol.28,PP.438-445, 1966.

(211 R .H . Park, tlm proved Reliability of Bulk Pow er supply by Fast Load
C ontrol,'' P roceedings of the Am erican Power Conference,V ol.30,PP.l128-
114 1, 1968.

(221 R .H .Park,GF ast T urbine V alving,''IE E E Trans., V ol.PA S-92,PP . 1065-1073,

h4ay/Jun e 1973.

(231 E.W .Cushing, G .E .D rechsler, W .P .K illgoar,H .G .M arshall and H .R .Stew art.

(F ast V alving as an A id to P ow er System T ransient Stability and P rom pt
R esynchronization and R apid R eload after F ull L oad R e'
Jection ,'' IE E E Trans.,
V ol. PA S-9 1, PP . 1624-1636, July/A ugust 1972.
R efe re n ce s 1 16 3

(241 H .F.M artin, D .N . T apper and T .

M . A lston , tsu stained F ast V alv ing A pp lied
to T V yk's W atts B ar N uclear U nits,'' P aper 76-JP G C -P W R -5, presented at the
1976 Joint P ow er G eneration C onference, B uffalo, N .Y ., Septem ber 1976 ,

A SM E Trans., Series A , V ol. 99, N o . 1, 1977.

(25) P.L.M cG aha and T.L.D resner, N uclear Steam Turbine lntercept V alve
and C ontrol System for F ast V alving,'' P aper presented at the l977 Joint
P ow er G eneration C onference, L ong B each, C alif., Septem ber 1977.

(26) L.Edw ards,R .J.Thom as,D .C .H o gue, P .H ughes, W . N ovak , G . W eiss and
J.E , W elsh, (sustained F ast V alving at T V A 'S C um berland Steam P lant'
.
Background and TestResults,55Proceeding of theA m erican Power Conference,
V ol. 43, pp . 142-152, A pril 198 1.

(271 IEEE W orking G roup R eport of panel discussion, tkT urbine F ast V alving to
A id System Stability : B enes ts and O ther C onsiderations,'' IE E E Trans., V ol.
P W R S-I, PP . 143-153, F ebruary 1986 .

(28) W .A . M organ, H .B . Peck, D .R . H olland, F .A . C ullen, and J.B . R uzek ,

utu odern Stability A ids for C alvert C liffs Y XCCS SS IE E E Trans., V ol. P A S-90,
PP . 1-10, January T ebruary 197 1.

(291 P.K undur and J.P.B ayne,((:4k Study of Early V alve A ctuation U sing D etailed
P rim e M over and P ow er System E9ilzatllfttitlzl,'' IE E E Trans., V ol. P A S-94, PP .
1275-1287, Ju1y/A ugust 1975.

(301 D .L.O sborn,yast V alving and N eutral Resistor A pplication on a 600 M W

*t,', P aper A 76608-0, presented at the 1976 Joint P ow er G eneration
Fossil U n1
C ohference, B uffalo, N .Y ., Septem ber 1976.

(31) I4.J. B alu, tiyast Turbine V alving and lndependent P ole T ripping B reaker
* 55
A pplications for P lant Stabillty, IE E E Trans., V ol. P A S-99, PP . 1330-1342,
July/A ugust 1980.

(32) R .M .M aliszew ski,B .M .Pasternack,and R .D .Rana,(Tem porary FastTurbine

V alving on the A EP SYStcm 5'DProceedings of theAm erican PowerConference,
*<' V o1. 44, PP . 118-125, 1982 .

(331 T.D .Y ounkins, J.14.Chow , A .S.Brow er, J.K ure-lensen, and J.B .W agner,
y ast V alving Asritlz R eheat and Straight C ondensing Steam T urbines,'' IE E E
Trans., V ol. P W R S-Z, PP . 397-405, M ay 1987.

(341 P.K undur,R .E.Beaulieu,C .M unro,and P.A .Starbuck,tgteam Turbine Fast

V ajVl
*ng . B enel ts and T echnical C onsiderations,'' 51 267 position Paper,
 1 164 M etho4s of lm prov ing Stability C haP . 17

presented at the C anadian E lectrical A ssociation , spring m eeting, M arch 1986.

(351 R .H . H illery and E .D . H oldup , tioad R ejection Testing of Lafge Therm al

E lectric G enerating U nits,'' IE E E Trans., V ol. P A S-87, PP . 1440-1453, June
1968.

(361 V .F.Carvalho, tjkjSe of G enerator and Custom er Load Rejection to lncrease

P ow er T ransfer L im its,'' C IG R E P aper 32-79-00-56, presented at the Study
@

C om m ittee 32 M eeting, M inneapolls, M ilm *7 1979.

(371 P.K undur and W .G .T.H ogg,ttgyo of G eneration Rejection 1

*l1Ontario H ydro
to lncrease P ow er T ransfer C apability,'' Paper presented at p anel SC S SIO
* II O n
gen erator tripping, IE E E P E S w inter m eeting, N ew Y ork, January/F ebruary
1982 .

(38J R .G . Farm er, (Elndependent D etection Schem e to lnitiate W estern System

Islanding for P acific A C Intertie O utages,'' Paper presented at the panel session
on controlled separation and load shedding, IE E E P E S sum m er m eeting, July
1985.

(39j C .W .Taylor,F.R .N assief,and R .L .C resap, CN orthw est P ow er P ool T ransient

Stability and L oad Shedding C ontrols for G eneration-lwoad Im balances,'' IE E E
Trans., V ol. PA S-IOO, PP . 3486-3495, July 198 1.

(40) D .C . Lee and P.K undur, (tA dvanced Excitation Control for Pow er Kystem
Stability E nhancem ent,'' C IG R E 38-0 1, 1986.

(411 C .W . Taylor, J.Ik. M echenbier, and C .E . M atthew s, (tTransient Excitation

B oosting at G rand C oulee T hird P ow er P lant: P ow er S ystem A pp lication and
F ield T est,'' P ap er 92 SA4 533-0 P W R S, presented at the IE E E P E S SunAnAcr
m eeting, Seattle, July 12-16, 1992.

(421 IEEE Com m ittee R eport, (CD ynam ic P erform ance C haracteristics of N orth
A m erican H V D C System s for T ran sient and D ynam ic Stability E valuation s,''
IE E E Trans., V ol. PA S-IOO, PP . 3356-3364, July 198 1.

(43) IEEE Com m ittee R eport,CIS V D ; C ontrols for System D ynam ic P erform ance,''
IE E E Trans., V o1. P W R S-6, N o. 2,PP . 743-752, M ay 199 1.

(441 C .E . G rund, G .D . Breuer, and R .P. Peterson, C:A G/D G Xystem '
D ynam ic
P erform ance - T ran sient Stability A ugm entation w ith D ynam ic R eactlve P ow er
C om pen sation ,'' IE E E Trans., V ol.P A S-99, pp . 1493-1502, July/A ugust 1980.

(451 P.L.D andeno,A .N .K aras,K .R .M cclym ont,and W .W atson,CkEffectofH igh-

R efere n ces 1 16 5

Sp eed R ectif er E xcitation S ystem s On G enerator Stability L l

*m l
*ts,:5 IE E E
Trans., V ol. P A S -87, PP . 190-20 1, January 1968.

(461 W . W atson and G . M anchur, ttE xperience w ith Supplem entary D am ping
Signals for G enerator Static E xcitation S ystem s,55 IE E E Trans*> V ol. P A S-92,
PP . 199-203, January/F ebruary 1973.

(471 W . W atson and M .E . Coultes, Kgstatjc Exciter Stabilizing Signals On Large

G enerators M echanical Problem s,5, IE E E Trans., V ol. PA S-92, PP .204-2 1 1,
January/F ebruary 1973.

(481 P.K undur,D .C .Lee,and H .M .Zein El-D in, ktg gw gy System Stabilizers for
T herm al U nits.
* A nalytical T echniques and O n-site V alidation,'' IE E E Trans.,
V ol. PA S-IOO, PP . 81-95, January 1981.

(491 F.P. deM ello, L.N . H annett, and J.M . lllltlrill, itpractical A pproaches to
Supp lem entary Stabilizin g from A ccelerating P ow er,'' IE E E Trans., V ol.P A S -
97, pp .15 15-1522, Septem ber/o ctober 1978.

(501 D .C . Lee, R .E. Beaulieu, and J.R .R . Service, ((z!k Pow er System Stabilizer
U sin g Speed and E lectrical P ow er lnputs D esign an d F ield E xperience,''
IE E E Trans., V ol. P A S -IOO, PP . 4 15 1-4 167, Septem ber 198 1.

(511 E.V .Larsen and D .A .Sw an, (tA pplying P ow er System Stabilizers, P arts 1, l1,
and 111,'5 IE E E Trans., V ol. P A S-IOO, PP . 30 17-3046, Jun e 198 1.

(521 F.P. deM ello, J.S. Czuba, P.A . Ruche, and J. W illis, tD evelopm ents 1
@l1
A pplication of Stabilizing M easures through E xcitation C ontrol,''C IG R E P ap er
38-05, 1986.

(531 P.K undur,M .K lein, (J.J.R ogers, and M .S . Z yw no, tk

A pplication of P ow er
System Stabilizers for E nhancem ent of O verall System Stability,''IE E E Trans.,
V ol. P W R S-4, PP . 6 14-626, M ay 1989.

(541 M .K lein,(J.J.R ogers, S .M oorty, and P .K undur, ttA nalytical lnvestigation of

@

F actors lni uenclng P ow er System Stabilizer P erform ance,''IE E E Trans., V ol.

E C - , PP . 382-388, Septem ber 1992 .

(551 Canadian Electrical A ssociation Report, Gtlnvestigation of L ow F requency

#

lnter-A rea O scillation Problem s 1l1 L arge Interconnected P ow er S y stem s,''

Report of Project 294 T 622,prepared by Ontario H ydro, 1993.

(56) E.V . Larsen and J.11. Chow , jj444(y Control D esign for System D ynam ic
P erform ance,'' IE E E Special Sym posium On A pplication of SV S for System
1 16 6 M eth o d s o f Im p ro v in g S ta b ility C h aP . 1 7

D ynam ic P erform ance, P ublication 87T H 0 187-5-P W R , 1987.

(57) N .M artins and L .T.G .Lim a,Kt

D eterm ination of Suitable L ocations of P S S and
SV C for D am ping E lectrom echanical O scillation s 1
@1I L arge P ow er S y stem ,''
Proceedings of the 1989 Power Industry Computer Application Conjrence,
PP . 74-82, M ay 1989.