Transient Stability

836 Chap. 13
NUMERICAL
INTEGRATION METHODS [4-61
13.2

to be solved in power system stability an.

The differential equations
with known initial values:
nonlinear ordinary differential equations

dx = f(%,t)
dt (13.10A
variables and tis the independent Varato
where x is the state vector of n dependent able
to solve x as a function of i, with the initial values ofr
(time). Our objective is and
t equal to Xo and to, respectively.
of numerical inteprati
In this section we provide a general description
of the above form. In describing thece
methods applicable to the solution of equations
13.10A as if it were a first.
methods, without loss of generality, we will treat Equation
order differential equation. This simplifies presentation and makes it easier for a
features of each method.
novice reader to comprehend the special
We will first describe the Euler method, which by virtue of its simplicity

serves as a introduction to numerical integration, and then we discuss more

good
advanced methods.

13.2.1 Euler Method

Consider the first-order differential equation

dx = fx,) (13.10B)
dt

illustrates the principle of applying the Euler method.

with x=X at =l. Figure 13.6

True solution

X Tangent
Ax

At

Figure 13.6
 Sec. 13.2
Numerical Integration Methods
At x =*o» t=lo we can
approximate the
tangent having a slope curve 837
te
representing the true
true solution
solution by
d
dt x

Therefore,

Ax = |dx
At
dt x-o
The value of x at t= =lo +Ar is given by

dx
o+Ax =z At
(13.11)
The Euler method is equivalent to using the first two terms
of the Taylor series
expansion for x around the point (x,to):

X +ArG%)+G,)+
2! 31 )
(13.12)

After using the Euler technique for determining x=* corresponding to t, we can

1ake another short time step At and determine x^ corresponding to l= +Ar as tollows:

dx At (13.13)
dt |x-*

By ying the technique ofx can be determined corresponding

successively, values
to
different values of t. therefore, referred
of x and is,
derivative
The considers only the first Ar has be
towill
od accuracy for each step,
*Urder method. To give sufficient effort required
small.
. This
This will and the computational
be very high. incr
r o u n d - o f f errors,
a s e

important
to

methods, it
is very
the
In the integration made early in
Consider application
der the pro
of numerical
cause slight
errors

the propagation
propagation of error, which may
depends
on
later, the
Procese
proces stability errors

oferor. be magnified at later steps.

Numerical

significant
further
cause
other

ethod isIf said

early c a u s e no
errors

errors carry through but other hand,

early
on the
to be numerically stable. If, unstable.

errors later, the method is said to be

numerically
 Sec. 1 3 . 2
Numerical Integration Methods
At =X0 =lo we can
tangent having a slope oximate
OXImate the
the curve 837
its
representing the true solution
by
dx
dt
dt x*x 4 )

Therefore,

Ax dr At
dt x

The value of x at t=tj=lo +At is given by

dx |
X Ax =x^+ At (13.11)
dt x-o

The Euler method is equivalent to using the first two terms of the Taylor series
expansion for x around the point o)

Ar6,)+ ) ) (13.12)

to we can
I=l,
Alter using the Euler technique for determining x=x corresponding as tollows:
another short time step At and determine x, corresponding l=1 +
to
aKe

(13.13)
dx At
x dt x*x
be determined corresponding
values ofr can
By applying the technique successively,
to die"8 referred
different values of f. derivative ofx and is,
theretore,
to be
first Ar has
The method considers only the accuracy
for each step, will
0to as
as aa sufficient effort required
order method. To give computational
and the
small. TThis will
Snall n c r e a s e
round-off errors, important
to
be very high. it is very
the
integration
methods,
i made early in
In the pplication of numerical
errors

slight propagation
on the
cause

consider the opagation of error,

COnsider which may
stability
depends
f
fuur
rtth
heer
r
er
e rr
roor
rss
later,
later,
the
th
process to be magnified at later steps. Numerical

se no no
significant
signiticant errors
cause
other

error. If early errors carry through but o n the other hand, early
cause

method is said to be numerica

1cally
s t a b l e . If,
numerically
unstable.

b
be
to
to
later, the method is said
 Transient Stability
838 Chap. 13
13.2.2 Modified Euler Method

The standard Euler method results in inaccuracies because it uses the da

interval though it applied throughout the rivative
intvative
at the beginning of the as
rval. The
modified Euler method tries to overcome this problem by using the average
the
derivatives at the two ends.
The modified Euler method consists of the following steps:

(a) Predictor step. By using the derivative at the beginning of the step, the valiu
alue
at the end of the step is predicted

dx
o dt \-soAt (13.14)

(b) Corrector step. By using the predicted value of xi, the derivative at the end of
the step is computed and the average of this derivative and the derivative at
the beginning of the step is used to find the corrected value

dx
x Xo2 At (13.15)
dtK dt |x-xf
If desired, a more accurate value of the derivative at the end of the step can be
calculated, again by using x=xi. This derivative can be used to calculate a more
accurate value of the average derivative which is in turn used to apply the corrector
step again. This process can be used repeatedly until successive steps converge with
the desired accuracy.
The modified Euler method is the simplest of
predictor-corrector (P-C)
methods. Among the well known higher order P-C methods are the
method, Milne method, and Hamming method [4]. The applicability of these methods
Adams-Bashtor
to power system stability analysis has been investigated in reference 7 and has been
found to suffer from a number of limitations. They need
are not self-starting, more
computer storage and require smaller time steps than the Runge-Kutta methoas
described below.

13.2.3 Runge-Kutta (R-K) Methods [4,51

The R-K methods the
approximate
the Taylor series solution; however, u f
formal Taylor series solution, the R-K methods do not
derivatives higher than the first. The effects of require explicit evaua by
by
several evaluations of the first derivative.
higher derivativesnumber
are ine o f terms
Depending
the
effectively retained in the Taylor series, we have R-K methods ofnumo
on
different orders
Sec.
.
13.2 a 2 Numerical Integration Methods
Secona-order R-K metho
839
Referring to the
differential equation 13.10B,B, the
the
value of r at f=lo +Ar is second-order
r
R-K formula

A x =x+
2
where

k = f,)At

k fo+k.to*Ar)at
method is equivalent to
considering first and second derivative terms in the
Taylor series, error is on the order of Ars.
A general formula giving the value of x for the
(n+1st step is

n+1X*- 2
where

Fourth-order R-K method

ne general formula giving the value of x for the (n+1)° step is

(13.16)
, , + 24,*2k,+k,)
where
k = fo,,)At

k Sr,+k,t,+At)A
 Transient Stability
840 Chap. 13
is as followS:
The physical interpretation of the above solution

k= (slope at the beginning of time step)Ar

2 (irst approximation to slope at midstep)Ar
K3 (second approximation to slope at midstep)Ar
k (slope at the end of step)At
Ar
=1/6 (kj+2k +2k3*ka)
Thus Ax is the incremental value of x given by the weighted average of estimates
based on slopes at the beginning, midpoint, and end of the time step.
This method is equivalent to considering up to fourth derivative terms in
the
Taylor series expansion; it has an error on the order of ArP.

Gill's version of fourth order R-K method (R-K-G)

With x as the initial value of x at the
and 4 to denote four stages, each
beginning of a step and by using j=1,2.3
stage of the Gill method can be described as follows
4]:

k, a,Lf-1)-b,4.1
k,At (13.17)
9, 9,-13k-c,f-,t)
The values of the a, b, and c coefficients are as follows:

a 1/2, b 2,
C =a a =

1-y0.5, b =
1, c2 a2
C2
a l+/0.5, b, =1, c3 =aji a 1/6, b
=
=
2, C4 1/2

Solution the end of a time

at
step is given by xa. Initially qa=0, thereafter in
the solution, go for the next
step is equal to q4 of the advanc
The following are the previous step.
advantages of Gill's version of the R-K methoOd
(a) Roundoff errors are minimized (variable q is used for this purpose).
(b) Storage requirements are less than for the
original R-K method.
Care should be exercised in
or sudden changes in the rate of applying the R-K-G method when discon
uities

variable should not be change of variables occur. Within a time

limited, as a result
of sharp nonlinearities, h the
variable q takes on incorrect otnerw
e

values.
 848 Transient Stability
From Figure E13.5, for critical clearing, Chap. 13
A, A
or

A,4, A,4
Therefore, from the plots, we have

125.27
0.9(125.27-41.77) =1.1024sinô dô

1.3116 =

1.1024 (cosô +0.5781)

Thus the critical
clearing angle is

= 52.29
This agrees with the value
determined from time
responses.

13.3 SIMULATION OF POWER SYSTEM

DYNAMIC RESPONSE
13.3.1 Structure of the Power
System Model
Analysis of transient
stability of power systems involves the
their nonlinear
dynamic response to large computation ot
fault, followed by the isolation disturbances, usually a transmission network
of the faulted element by protective
Figure 13.8 depicts the general
to transient structure of the power relaying
stability analysis. This model structure systemmodel applicaoie
Chapter 12 (Figure 12.18) for is similar to that preseneu
stability analysis, nonlinear smail-signal stability analysis. However, for transie
discontinuities due to faults andsystem equations are solved solved. In addition,
addition, large
limits on network switching, and small to
system variables, appear in the discontinuities u
performance of protection systems are of system model. Bus voltages, line flow>.
related to the stability of interest, in addition to the basic intot
the system.
As seen in
models for the Figure 13.8, the overall includes
following individual power system
components: representa
Synchronous generators, and the
movers associated excitation system prime

Interconnecting transmission network

including static loads
 3.3 Simulation of Power
Sec. 13.3

System ynamic
Response 849
Stator equations and ER. E
axes transformation
R
Transmission
*****
- -

network Other generators

equations
Generator

Excitation including Motors

rotor circuit static loads
equations
System
Other dynamic
Acceleration or Prime mover devices, e.g.,
swing equation governor HVDC, SVC
Individual machine ****~----.
Common reference
reference frame: d-qq frame: R-I

Algebraic equations
** Differential equations

Figure 13.8 Structure of the complete power system model

for transient stability analysis

Induction and synchronous motor loads

Other devices such as HVDC converters and SVCs

Ine model used for each component should be appropriate for transient stability
analySis, and the system equations must be organized in a form suitable for applying

numerical methods.
model consists of a large
AS we willin what follows, the complete system
see

differential equations and large sparse algebraic equations. The

Ordinary initial-value problem.
aULSIEnt
stability analysis is thus a differential algebraic
13.3.2 Synchronous Machine Representation [8]

3 to unit 5. The per

modelling is presented in Chapters Modifications of
Cnronous machine Section 3.4.9.
summarized in
synchronou
these equations
achine equations are discussed in

in stability
studies are
representation
necessary for
Section
on 5.1. model for
transient stability
d-axis and
To illustrate the implementation of the generator a model with
model
one
one

analysis,
WO
we assume that the ger the g en
neerra
atto
orr is represented
by
However,
the equations
presented
number of

here q-axis
amortiidt 13.9. arbitrary
heres amnoortisseurs as shown in Figure model
with an

can be for a
eadily modified to account

amortisseurs.
 Transient Stability
850

L
L Chap.1's
La
ads Vag aqs
d ad R
ZRidea
Figure 13.9 Synchronous machine equivalent circuits

The following is a summary of the synchronous machine equations as a se

first order differential equations, with time t in seconds, rotor angle 8 in elerthi
radians, and all other quantities in per unit.
Equations of motion

pao, .-T,-K,ao,)
(13.22)
po = A,

where

@g 27% electrical rad/s

Ao, = pu rotor speed deviation

P derivative operator d'dt

Rotor circuit equations

With rotor currents

expressed in terms of rotor and mutual flux nnINag (See
Equation 12.159 of Chapter 12), the rotor
circuit dynamic equations ae

PVsa a (adaRa
Pp1d Rd
Lnd
(13.23)

p14 oiR,
p2
L24
 Simulation of Power System Dynamic
Sec. 13.3

flux
Response 851
q-axis
mutual
linkages are given by
a- and
The

ad-adiakas iya*Ladeind
= Lads
(13.24)
L
Vaaee-i ,2
(13.25)
where

Lads
Ladis La
(13.26)
L'ags 1
ags 14

Here Lads and Lags are saturated values of the d- and

q-axis mutual inductances given
by

Lads =Ksaadu
(13.27)
ags aqu
nd and
Ka are computed as a function of the air-gap flux linkage Vat as
escribed in
Chapter (Section 3.8.2).
3
Stator voltage equations

With stator transients (pypy.) and speed variations (o/0g) neglected as

Cnapter 5 (Section 5.1), the stator voltage may be written as follows:

e-Ria*(L)i,+Ej (13.28)

-Rai-(La)i,+E
 Transient Stability
Cha 13
852
with

E - ags
Lia L
(13.29
E wLads
La id

(13.30

the stator voltaos

Since we have neglected the effect speed
of variations on
tage
Consequently, oLg =Ag and oLq=Xg. Theabove
=0/0=1.0 in the above equations. machine d-q reference frame which rotates with the
are in the individual
equations
the interconnecting transmission network
machine's rotor. For the solution of
R-I reference is used. The relationships
equations, a synchronously rotating common
from one reference frame to
shown in Figure 13.10 are used to transform variables
the reference for
the other. The R-axis of the common reference frame also serves as
measuring the rotor angle ô of each machine.

e Epsino-E,cosô
E ***********
---

eE,sino+Ecoso
r (13.31)
E =e sinð+e,coso
E,esinð-e,cosð
ER
Figure 13.10 Reference frame
transformation
and definition of rotor angle d

For convenience in
the
the stator plete set of algebraic equa
organization of the complete c e frame 2s
voltage equations are set o
expressed in the common K-l erence fraac
indicated in Figure 13.8. stator voltage

Use of Equations 13.31 to

equations 13.28 yields transform u
 Simulation of Powe System Dynamic
Rest
S e c . 1 3 . 3

esponse 853
Ex -RRX'E
E,-X R,|E (13.32)
The elements of
ats of the impedance matrix are given by

RRR (Xa-Xq)sinð cosô+R,

R (X-X4) sinôcosô +R,
XRI Xgcos*ô +X*sin? (13.33)
XIR Xsin?ô+X cos 83
dearlier, o is assumed
As no
to be equal to 1.0 pu. The internal voltage
are given by components

ER E sinô+E" cosô
(13.34)
E E"sinð -Ej coso

If subtransient saliency is negligible, Lg =Lg. Then

RRR R= R
XRXIR = olj = X X =X"

nis case, ER tjE represents the voltage behind the subtransient impedance R, YA".
erwork solution, the generator may be represented by either of the Simple
quivalent circuits shown in
Figure 13.11.

Z"
Z" R,+X
Y" = 1/Z"

(a) Thevenin's equivalent (b) Norton's equivalent

circuits
13.1l Synchronous machine equivalent
subtransient saliency
wilh negligible
 Transient Stability
854 Chap. 13
saturation may introduce
due to
small amount of
a .

Variations in Lads and Laas condition. This is usually ino

a transient

computational ficant aand

subtransient saliency during of the
desired to take advantage
may be ignored, if it is circuits. simplicit
offered by the above equivalent
stator terminals are
Active power and reactive power at the generator

P,aia* (13.35
(136

The air-gap torque required for the solution of the swing equation (13.22) is

T. i , a
(13.37A)

Since we have assumed o =o/o, =1.0 pu in the stator voltage equations, in per unit the
air-gap torque is equal to the air-gap power (see Chapter 5, Section 5.1.2). Hence

(13.37B)
T, P P,+R,T?
The field current in the
reciprocal per unit system is given by

(13.38

The per unit exciter

output current Ia (see Chapter 8, Section 8.6) 13

(13.39)

Here we haveconsidered a generator

a X s

two
amortisseur circuits. For models with model with ne d-axis and changesto
theabove a different ircuits,
formulation of machine number of rotor hav we

assumed equal mutual

axis. Reference inductances
equations are straightrorwa However,e Ver,
r c u i t si n
each

8 between the armature and ro m o d e l with

4ith

unegual mutual provides a

description of the of a a0
r o a c h

on
inductances implemena
g e n e r a la p p r o a c h

for stability analvsis, using the abov

Sec. 13.3 Simulation of Power System
Dynamic Response
855
generator variables
aitial values of

t e d earlier, the transient

stability analysis involves the
rential
differe and algebraic equations with known initial solution of alarge
set of
set of
i e provides the initial values of network values. Prefault power
flow analy
variables,
ve power outputs and voltages at the generator terminals. including the active and
The nrocedurefor computing the initial values of
the generator
known terminal conditions is presented in Section 3.6.5 of Chapter 3. quantities for

model and infinite bus

Classical generator
For a machine represented by the classical model, Xp=XIg =X% and E" is
replaced by E. The rotor angle d is the angle by which E leads the R-axis. The
magnitude of E' is constant throughout the solution. The R and I components of E
change with 8, as determined by the solution of the swing equation.
For a node associated with an infinite bus, both the magnitude and angle of the
node voltage remain constant.

13.3.3 Excitation System Representation

in use are described in Chapter
Models for different types of excitation systems
these models into a transient stability
8. We will illustrate the method of incorporating in Figure 13.12. It
the excitation system model shown
program by considering as type ST1A in Chapter
excitation system (classified
represents a bus-fed thyristor (PSS).
and a power system stabilizer
an automatic voltage regulator (AVR) feedback is
8) with without transient gain reduction or derivative
A high exciter gain (K) to the excitation system model
used The model of Figure 13.12 is jdentical that we also account for the limits
except
considered in Chapter 12 (see Figure 12.14),
on exciter and PSS outputs.

Terminal voltage
Vref EF max
Exciter
transducer

E 1+SR
1
- KA
Emin
Phase Vsmax
compensation ma
Gain Washout
1+sT1
sT
KsTAB 1+STw 1-sT
5) Vs min
3 4
Power system stabilizer
AVR
and PSS
with
system
excitation

Thyristor
Figure 13.12
 856
Transient Stabilityy
Chap. 13
For a bus-fed (potential source) thyristor exciter, the voltages var
generator terminal voltage (E) and exciter output current ( ) :
vary with the

Efmar VRma-Ku
EFminVRmin
(13.40

From block 1 of Figure 13.12, we may write

(13.41)
TR
From blocks 3 and 4,

P sTAB PAa, h (13.42)

with pAo, given by Equation 13.22. From block 5,

py, 7Pv,*V-K,] (13.45)

12
with pv given by Equation 13.42. The stabilizer output v, is

(13.444)

with

(13.44B)
smar 2 V2 smin
Vemax

From block 2, the exciter output voltage is

(13454)

E K,VV,*V,]
with
(13.45B)

Efma
EFmaxEa2 EFmin
 Sec. 13.3
Simulation of Power System Dyn:
Both limits considered namic Response
here are
discussed in Chapter
8. windup limits. 857
discussenerator field voltage efa
is
The generator field
in the
Modelling of non-windup
ilimits
exciter output voltage Eja (see Chapter 8, reciprocal
Section 8.6) asper unit system
follow
lows:
is related to

efd
(13.46)
alternative
method
An alternative method
a of treating blocks 4
iobles are not required, is described in and 5, in which the derivatives of the input
variables

Chapter 12, Section 12.3.4.

Initial values of excitation system variables:

For any given steady-state generator output, the field

hy the generator equations (see Section 13.3.2). The excitationvoltage ea is determined
determined as follows: system quantities

L
Ega Radu
Rid (13.47)
VE
The AVR reference is

fd+V (13.48)
KA
to the
usref takes a value appropriate to the generator loading condition prior
disturbance.
Kepresentation of field-shorting circuit [8/

the exciter current usually cannot be

or static exciter,
c a s e of an ac circuit in the form of a
"crowbar" or a
a special field-shorting current to flow (see
nerefore, and allow negative field
provided to bypass the exciter
Ch
Chapter 8, Section 8.5.8). simulated
voltage
by setting the field
are
The field-shorting circuits the period when
Crects of appropriately during
the field resistance (R)
zero and For a
increasing is fixed for a crowbar.
the field curren resistance added Once the field
1s negative. The applied voltage.
Varistor rent
varistor the resi is a nonlinear
function of the
restored to its normal
value,
Sstance resistance is
Current becomes
uTent the field circuit output.
and the field voltage is established by uthe
positive, exciter
 26

2.5 DYNAMIC STABILITY MODEL OF MULTI MACHINE

POWER SYSTEM

In stability analysis of a multi-machine system, modelling of all the

machines in a more detailed manner is exceedingly complex in view of the
largenumber of synchronous machines to be simulated Theeforesimplifying
assumptions and approximations are usualymadein modellingthe system
In this thesis two axis model is used for allmachinesin thesample system
takenfor investugation.

2.5.1 Assumptions Made

In this work the synchronous machine is modeled using the two-

axis model (Anderson and Fouad 2003). Inthe Iwo-axis model the transient

effecis are accounted for. while the sub transient effectsare neglected. The
transienteffects are dominated by the rotor circuits. which are the field circuit
in d-axis and an equiyalent circuit in the g-axisfornied by the solid rotor. The

amortisseur winding effeots are neglected. An additional asaumption made in

this model is thal in stator voltige equätions the terns px, and pA, are

negligible compared to the speed voltage terms and that o= o, Ip.u. The
=

two-alis model
block diagram representation of the syichronous machine in
.

is shown in Figure 2.6

 quadrature axis open circuit time constant
T electnical torque of synchronous machine
T
Tm mechanical torque of synchronous machinee
D damping coeficent of synchronous machine
ED Equivalent stator emf corresponding to field voltage
qUadrature xE AImature cient
direct axis armature curent
ineriaconstantol synchronous machine n sec
trequency in Hz

A mult-machine power system 1s Shown n Figure2 and the

network has n machines and r oads. The active source odal voltages in
Figure 2.7 are taken as the terminal voltages V i = .2..n instead of the

interna EME S. The loads are repreented by constant impedäneeS and te

Hetwork has n active sources reprexemting the SynchroOnous mathines

-
- ::

Figure 2.7 Muil-nachine wih constant impedince loads

 in
is reduced to a n-node network shown
This network

voltage phases of each node

are expressed
Figure 2.8 in which the current and
frame.
in terms of the respective machine eference

Transinission
System

-02

o1

Flgure 2.8 Reduced n-port network

derive relations between V and

The objective here is to

i=1,2....n. and the state variables. This will be obtained in the fom of a
Vai.
the machine currents i and i and the
relation betweæn these voltages. .

will a complex notation as

angles o .
i=l.2.....n. For convenience we use

follows.

For a machine i we deline th phasors and as

21
V=VVdi :
 where

i / 3 Tdi 3

rwig nd where te axiS q 1S taken as the phasor reference in each Tben we

case.

detinethe complex vectors V and by

+jV

(2.13)
V,V +V

T + j

hk (2.14

The voltageV and thecurrentl ar referredto the q and d xes or

machine i. In the other words the different voltages and currents are expressed
i n ierms of different reference To obtain general network relationships, it is
esirable to cxpress the various branch quaAntities ta Lhe sanme reference which

IS given by equation (2.15

Xpraseu s

. 2.15)
 25.3 Convertlng to CommonReference Frane
i

the phasor V=V+jV, to

i
e t S 1ssue that wWe want to convert

t e common eference rame (moving at synchronous peed). Let the

same

voltage, expressed innew notation, be VTV as shown inFigure 2.9.

where, V=V, JV and V-V+V (2.16)

Figure 29Two frames of reference for phasor quantities

p m the FIgN 2

INcOsoV coso 1 (2.17)

(2.18)

:i.iftigi

+V coso+(V sin-V coso (2.19)

 J02
0
o2D2 0 Y2a2 (2.21)
0
Dn 0 0 0 ej n q d n

j62 0
(2.22)
0
0000eon

Theequation (2.20)can De Written as

mmamaa.am
TV (223)
hus isa transformation that transforms the d and q quantities ofall
machine s to the system frame, which a common frame is movingat
nchronous apecd. The transformation matrix T contains elements only at the
w.ww eading diagonaland hence we can show that T isorthogonal,i.e.T=T.

Now the equation (2.23) can be rewritten as

VT (2.24)

Sim
milariyor noIe urrenl

(2.25)
.
 1-T1 (226)

Substituting equation (2.25) and equation (223) in equation (2.15),

we get

=MV (221

where M=TYT (2.28)

Linearizing equation (2.27) and making necessary substitutions

(Anderson and Fouad 2003),the following equations are obtained.

AL GAV4-BAVatiYicos(0-6yo)AV

Ysin8)AV Y,(sin(,Vtcos p l
1,.n (2.29)

AlBiAV +G,AV+Ycos( ap)AVl

Ysin j-80)AVql+Ysin(9-opo Vao- cos ôpVplAo

i-l.n (2.30)

The state spaee model for linearized system is obtained by

linearizing the differential and algebraic equations at an operating point
While doing this linearization process, additional terms involving
terminal
equations. To express the voltage components in terms of state variables, the
machine curents are also linearized and expressed in terms of state variables
and voltage components. Finally the current components are eliminated using
the interconnecting network algebraic equations From the initial conditions,
Eai Eqo. Ig, lao, EeDio and & are determined.

Linearizing equation (2.11) we get

pAEai -AE--) Al, }/spi i=1,..n

pAE {AEpDi-Ei+(- Xai) Ala }/t i=1.n

pa@ {AT-o AE + Lgo AE +Edo Al +Eo AL

D / i=1,.n

pAo Ai-L. (2.31)

Substituting equations (2.29) and (2.30) in equation (2.31).

(replacingV by E):

pA

+-Xqi) Y (sin ( - Ouo) AEat-gi gi)Gi AE

qq)YIcos (6-1AE

(%-qi) k=EYa cos(e-ôko] Egko+Yu sin(a-ôao)EgkolAðl

pAE (l(a-xai) BuAE
do i

+a-a'dYsin ( a A k t -xa) G; ARii

+Fa-d rasinkal1AE

-d IYa cosano Ei Ya sin(Ceno)Edoldoht AErp)l

Il2 (2.33

pA@ [AT- D,Aa,-uo GEo-B Fl AE

1o t+B Edio + G E 1AE

ËY cos ( - t o E a ioH sin (9 o) Eiol AE

#1

Y sin (0j-Od)E4cos (Ha).E

Eio1AEi

- Y cos (6 ö0KEKGEREG)+Y sin ((e 8o)

k

+ oa6J