See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/3463038

A Mathematical Model for Dynamic Simulation of HVDC Systems

Article  in  IEEE Transactions on Power Apparatus and Systems · July 1983

DOI: 10.1109/TPAS.1983.317918 · Source: IEEE Xplore

CITATIONS READS

13 890

3 authors:

Nicholas A Vovos G.D. Galanos

University of Patras Southern Illinois University Carbondale
92 PUBLICATIONS   547 CITATIONS    31 PUBLICATIONS   394 CITATIONS   

SEE PROFILE SEE PROFILE

Γαβριήλ Β. Γιαννακόπουλος
University of Patras
59 PUBLICATIONS   326 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Βιβλία Ν. Α. Βοβός View project

Voltage source inverter for RES connected in microgrids View project

All content following this page was uploaded by Nicholas A Vovos on 20 October 2013.

The user has requested enhancement of the downloaded file.

 IEEE Transactions on Power Apparatus and Systems, Vol. PAS-102, No. 6, June 1983 1755

A MATHEMATICAL MODEL FOR DYNAMIC SIMULATION OF HVDC SYSTEMS

N. Vovos G. Galanos G. Giannakopoulos

Power Systems Laboratory
School of Engineering
University of Patras, Patras Greece

Abstract - This paper develops a mathematical sor techniques and this increases considerably the fle-
model for the dynamic simulation of ac-dc systems. The xibility of the computer program.
basic features of the proposed model are: The basic disadvantage of the model is the exten-
All the subsystems and components of a realistic sive computation required and is, therefore, suitable
HVDC scheme are included in the model. only for applications where accuracy is essential, as
The power system components are modelled using the for the final testing of new converter control systems.
exact differential equations derived from their three
phase equivalent networks and therefore the model can I NTRODUICT I ON
give the instantaneous values of voltage, current and
power at any point of the system. The range of applications of HVDC transmission can

Fig. 1. The system model led

The considerable amount of computation involved is be extented and the transient performance of the dc
substantially reduced using tensor techniques as the links can be improved using suitable control stra-
differential equations of the system are automatical- tegies. One of the potential applications is the impro-
ly assembled to deal with the time varying topology vement of the transient stability of the interconnected
of the network caused by the switching action of the ac systems using dc power flow control, provided that
valves. the operation of the dc link is not interrupted due to
Line and bus faults are also simulated using ten- commutation failures during the ac faults. Converter
control and protection method can be initially tested
on simplified system model but for their final testing
and for their optimization of their parameters a more
accurate system model is essential.
A number of dc power flow control strategies have
been proposed (1),(3),(5),(18) and tested using si-
83 WM 037-9 A paper recommended and approved mulation techniques. The ac-dc power system models used
by the IEEE Power System Engineering Committee of fall in one of four main groups:
the IEEE Power Engineering Society for presenta- a) Analog computer simulation or scaled down power
tion at the IEEE/PES 1983 Winter Meeting, Nejw system models (1),(5),(17).
York, New York, January 30-February 4, 1983. b) Detailed digital simulation of the HVDC conve-
Manuscript submitted September 3, 1982; made avail- rters and representation of the ac systems with simple
able for printing December 22, 1982. ac sources (7),(8),(9),(14),(15),(16).
0018-9510/83/0600-1755$01.00 O 1983 IEEE
1756
c) Digital simulation of the HVDC converters and
representation of the other power system components
with their transfer functions (2),(6),(12),(18). I-I
d) Digital simulation techniques for ac/dc systems
with particular emphasis on the detailed modeling of
the transmission lines,mainly suitable for the study of ec
ovOrvoltages and currents caused by switching and elec-
tromagnetic transients C4),(l1).
All the above methods, althouth they can provide a
good indication of the response of an integrated ac-dc
system subjected to a particular control strategy, they
do not provide exact and detailed waveshapes for the
systems voltages and currents.
The aim of this paper is the development of a di-
gital simulation method for the detailed and exact re-
presentation of rea-listic configurations of ac-dc sy- Vd
stems,suitable not only for transient stability studies
but also for the study of injected harmonics and any o-
ther application requiring the calculation of the in-
stantaneous values of the system's state variables.
DESCRIPTION OF THE SYSTEM MODELLED
A power system consisting of two finite control
areas interconnected via two ac tie lines and a dc link
is shown in figure 1. Each control area is represented
by a turbine-generator set equipped with power frequen-
cy and voltage-reactive power control systems. To each
convertor bus are also connected all the essential com-
ponents used in real ac-dc schemes such as:
a) AC filters for the characteristic harmonics and Fig. 2 The equivalent circuit of the converter.
a high pass filter.
b) Switchable capacitor banks for reactive power
compensation.
c) Static loads of variable magnitude and power
factor.
d) Generator,line and converter transformers.
e) Protective relays and autoreclosing circuit
breakers.
The analysis and mathematical modelling of all the
subsystems and components shown in figure 1 are descri-
bed in the following sections,together with the deve-
lopment of a flexible method for simulating bus faults,
load switching,unbalanced loading conditions and all ty- Ie
pes of line faults at any desired point of the system.
An important advantage of the model is that the
firing angle control systems of the converters are com- I
pletely independent and therefore any HVDC transmission
control strategy can be easily simulated and tested
using the model.

MODELLING OF THE DC LINK

The main difficulty in modelling dc systems is Fig. 3 The graph of the equivalent network and the
that the dimensions of the converter loop impedance tree selected.
matrix are changing with time as a result of the
switching action of the valves.
The basic principle of the method used for the mo-
del-ling of the dc link is as follows:
The loop impedance matrix of the converter network
is assembled assuming that all the valves are con-
ducting. The equivalent network is shown in Figure 2.
This matrix is termed the general matrix and is used
for the construction of the loop impedance matrix of
the converter for any conduction pattern of the valves.
This is done using an incidence matrix (Cn) (7, 8, 9). L R
The elements of the incidence matrix are directly rela-
ted to and computed from the conduction state vector of ie(1) I"_ ie(2)
the converter valves. Vdl Vcl1 C C VC2 Vd2
The general matrix of the converter can be con-
structed using the graph of the system shown in Fig. 3.
The general matrix of the dc link can be constru-
cted from the general matrices of the two converters
,_
_1 _ +

and the matrices introduced by the interconnecting

equipment on the dc side. Fig. 4 The equivalent circuit of the dc link.
 1757

With reference to Figure 4 we obtain: The incidence matrix is constructed as follows:

Vdl Vcl and Vd2 = Vc2 (1)

The state equations for the dc link are: CO1) 0 0

Vc1 Vc2 +ZLi'l

- 0 (2) [Cnl O0 C(2) (1 )
ie(l) -

tl
-

YcVcl = 0 (3)
i e (2 ) O 0 U
iI -
- Ycvc2 = 0 (4)
Thus this representation of the line introduces three
new state variables, the voltages Vcl, Vc2 and the line
current i1. where: [CO()] and [C(2)] are the incidence matrices
The general loop impedance matrix is constructed for the converters 1 and 2
as follows: respectively (7),(8).
and [u] is a 3x3 unit matrix.
0 0
z (1) 0 0 0 Analysing [Zn Rn S[ we finally obtain:
BI B2
0 0
dt[In] [Ln} [[u] - [Rn] [Inn] (12)
[Zt] = 0 Z (2) 0 + 0 (5)
Ci C2
REPRESENTATION OF THE FILTERS AND THE STATIC LOADS
A high pass and two harmonic filters per phase
Al A2 A3 0 0 0 are connected to each converter bus. The harmonic fil-
ters are of the series R LC type tuned to the 11th and
13th harmonics. The high pass filter consists of a ca-
pacitance in series with a paral lel RL network.
The differential equations for a tuned filter are:
where: [Z(l)] and [Z(2] are the generalized loop im-
pedance matrices of the con- dI1F 1
verters 1 and 2 respectively d = L (VS-VC-RIF) (13)
From the equations (1) we obtain: dV c 1
d = C IF (14)
[Bl] =[ ] [B2] =[ ] (6)
For the set of the two harmonic filters:
[C1] =[o oo O , [c2] [o o] (7)
From the state equations we obtain:
dt IF =[L ] [Vs] - Vcg - [R j I F (15)
OO 1 OO
where: The index i takes the values 1 and 2 for the bu-
[Al = o O O O O , [A2 = O 0 1 (8) ses 1 and 2 respectively:
The differential equations for the high pass are:
O OO OO O OO O O
di j=-1 (Vs-Vc)
d
(16)
[yc dVc 1 R (v-v)1

LA3] = O (9) dt C R
(17)
1 The static loads are represented by series RL net-
works. The differential equations for the loads are:
The loop impedance matrix of the link for a parti-
cular conduction state of the converter valves is ob- (18)
tained using an incidence matrix (7),(8),(9). [L] [LL] [[Vs] - [RL [I L]]

where: The index i takes the values 1 and 2 for the bu-
[Zn] r [z
[Cn]t ]
[Cn] (10) ses 1 and 2 respectively.
t
REPRESENTATION OF THE CIRCUIT BREAKERS
where: LCnJ is tensor relating the independent
a
currents of the general network to The switching action of the circuit breakers is
the independent currents correspond- represented by a resistance which increases with time.
ing to the topology of the link de- From experimental resistance-time curves we obtain
fined by the state of the valves. the analytical function using polynomial approximation.
1758

REPRESENTATION OF THE SYNCHRONOUS MACHINES The mechanical power input PTi is calculated from
the equations of the turbine and the power-frequency
The synchronous generators in the system are re- control system.
presented using the three phase terminal equations. The
step-up power transformers are represented by their THE POWER-FREQUENCY CONTROL SYSTEM
leakage inductance in series with an ideal transformer.
The equivalent circuit of the machine and the step-up The block diagram of the power-frequency control
transformers is shown in Figure 5. The damper windinqs system is shown in Figure 6. A non-reheat turbine is
are omitted in this stage.Their effect,however,is taken assumed and is represented with a single time constant
into account in the swing equation of the generator. transfer function.
From the equivalent circuit the mathematical model From the block diagram the differential equations
of the machine is derived as follows: of the power-frequency control loop are derived as fo-
l lows:
-1

[Lrs] N [Vsl]
dtd Pci = - KA ifi
i (22)
¶[i r] l rr Vr
dt i TGT (Pci Afi Y ) (23)
dt trr
r
pEs ] 1 E]
.t- PTi= TT i (yi-PTO
d 1
(19) (24)
I -

[r] rr Ir _ where: The index i takes the values 1 and 2 for the tu-
rbines 1 and 2 respectively.
It
ai

lac

r cc

jcII

rr Irr

tr
Vr

I:7

Fig. 5 The equivalent circuit of the machine.

The elements of the inductance matrices in equati-
on (19)
are functions of the angular displacement of
the rotor 5, which is calculated from equation 20.
Hj d26 d6i (20)
PTi-PGi= fo dt Di d PuMW

where: Di(d6i/dt) is the per unit power in the damper

windings and Hi is the per unit inertial con-
stant of the turbine-generator set.
The electrical output of the generator is:

PGi= [[Vs + [ d1d [it ]].

[it] pu. MW (21) Fig. 6 The block diagram of the P-f and Q-V
control systems
 1759
TGi: Is the time constant of the governor. passive components.
Ma thema t i ca lIy, th i s can be ach i eved by t rans fo rm i ng
TTi: Is the time constant of the non-reheat turbine. the loop impedance matrix of the general network using
incidence matrices. The elements of these matrices are
Kit: Is the gain of the integrator. logical functions of the state of the system's switches.
The system shown in Ficure 7 has 11 nodes,22 bran-
The initial values of the variables of the control sy- ches and 12 independent loop equations.The graph of the
stem are: system is shown in Figure 8 where a selected tree is
0 0 0 0 indicated with solid lines.
TPC r y= PTvTii = PG
ci = Yg (25) 20
The range of the variable Yi is limited by:
0 < Yj < 'yimax (26)
where Yimax corresponds to the rated output of the tu-
rbine Pimax.
THE VOLTAGE-REACTIVE POWER CONTROL SYSTEM
The block diagram of the Q-V control loop is shown
in Figure 6.The differential equations representing the
control system are derived from the block diagram:
d
Uri = T-j (KEiU2i Uni ) (27)

U2i = 1:-
[KAi (I V Iref IV[ -
Usi) -
U2] (28)

dtU TE T i (KsiKEiU2i - KsiUri TEiUsi) (29) Fig. 8 The graph of the system.
TEi ,KEi Are respectively the time constant and The loop equations of the generalized transmission
the gain of the exciter.
system are:
TAi KA Are respectively the time constant and
the gain of the amplifier.
Tsi KKsi Are respectively the time constant and
the gain of the stabilizing transformer. v
i i
i (30)
Variable ranges: U2imin < U2i < U2imax, O < Uri < Urmax
where (Vi)
and (I;) are the vectors of the voltages and
currents of the Independent loops respectively.
REPRESENTATION OF THE TRANSMISSION SYSTEM
The equation (30) is valid only when all switches
The ac transmission lines are represented by their are closed. For any other state of these switches the
equivalent i-networks. All possible line faults can be independent currents of the transmission system are gi-
simulated using the general network shown in Figure 7. ven by:
The various types of line and bus-bar faults can be si-
mulated by switching in and out branches of the network
and ad usting accordingly the values of the remaining [Ii] = [C] [Ip] (31)
ial Lal Rai La2 Ra2 ia4 where Cp is an incidence matrix and its elements de-
pend on the topology of the system corresponding to the
particular conduction state of the switches.This matrix
has dimensions 12x(9+m) where m takes values from 0 to
3 and it can be partitioned as follows:

C1
[Cp] =

whe re: C I] is a 6x6 unit matrix

[c2] is a 6x(3+m) zero matrix

[c3] is a 6x6 zero matrix

Fig. 7 The generalized transmission system. c can be partitioned as follows:

1760

C41 C42
[c4]= tEc
LE e 3l t E 3E E [
C43 C44

whe re: and therefore:

[C41] is a 3x3 unit matrix

VI ~~Zixlj
(35)
[C42] is a 3xm zero matrix t = t
C4xV2 C4xZ4xC4XI2
[C43] is a 3x3 zero matrix
since [zI] = - [Zc] where:
[C440 is a 3xm matrix,the elements of which take
values 1,-i and 0 depending on the topoloqy of the net- diagonal 1 ,1 1 1 1 1
work. This matrix relates the currents i10, i11, i12 to [Zc] =
the independent loop currents corresponding to the par-
[Cal Cbl Ccl Ca2 Cb2 Cc2 j
ticular topology of the fault simulating network. The we obtain:
value of m is therefore equal to the number of links
in the network. Since the element 13 is selected as
branch m is equal to the number of the network elements d [V1]I=
that are switched in minus one. [Zc] [ii] (36)
The values of the elements of (C44) can be easily
calculated from the state of the switches of the fault
simulating network. Analyzing [Z4] = [R4] + s [L4] and substituting in e-

Assuming that the incidence matrix is known, then:

quation (35) we finally obtain:
i Z c (32)

and therefore:
[c4] [V2] = [C4] [ER4 + s [L4]] [C4] [12] (37)

Definina:
[Cp] [v.] = [Cp]t [Zi] [Cp] [Ip] (33)

Equation (33) can be written as:

[Vm] [c4]t [V]

[VP] [Z ]
[zp] (34) [Lm] =[cijt [L4] [C4]

where:

The equation (37) can be written as:

[vp] = [cp] [vi] and [zp] [Cp] [Z;] [Cp]
[Vm] [[Rm] ±s [Lm]] [2] (38)
Equation (34), however, is not very useful because
of the sparsity of the resulting modified impedance ma-
trix. Better results can be obtained by partitioning In the time domain equation (38) becomes:
the matrices as follows:

dt I2] [Lmj [[1 - [Rm] [12]] (39)

[12] [j Zj
[Vjl]= F [ Ij= 1[Zj]= The equation (36) can be further simplified by parti -
tioning the matrices involved as follows:
V2

Vs IF7]
where: 1] [2] [I2]
[ 1], have dimensions 6xl
LZc]=
L i= j
I1
and [ z z have dimensions 6

The equation (33) can now be written as follows: Then equation (36) becomes:
 1761

it is therefore necessary to express them in terms of

IL
[s1] [Zcl] [II] and d[V]- [Zc2] [I Il other state variables of the system. The configuration
of the system's bus-bar is shown in Figure 9. Applying
REPRESENTATION OF TWO PARALLEL LINES Kirchhoff's law we obtain:
The simulation method developped in the previous
section can be easily extended for the case of the sy-
stem shown in Figure 1, where two parallel ac lines are [il] [Eu] [i}-1 L ]]-1 [L -
present.
We define the current vectors:
- LB] [CL4] [' - [B] [C] [ -
Ijl
Ij2 i L] I 2][] 2
II
1j3
ID
-i (46)
1j4 i
115

[Ii] I6 III
[ 1J = [[u] + LZc2 Zc2] I[sg + 111I1

[ Ij fL -D - EJ
1j7
Ij8 J
Ijg
jlO
'III

J
2I
j (47)
Lj 1
jl 12
iV where: 1-O-OimOi
1 00100

Where the index j refers to the lines 1 and 2 respecti-

vely.We also define the bus voltage vectors as follows: _ , _

l [Vall va2
[u] is a 3x3 unit matrix
! VblI LVsj
and = Vb2j
v1
[VclJ [Vc2J ii the current vectors of the static loads

[I the current vectors of the ac-dc converters

By partitioning the matrices
[4i] =
we obtain:

dt { s [Zi] [I] (40)

dd [v =- Zi [11 (41)

(42)

d 2= [2mj
dt 12 [Lm]
][]
Vm [Rm 12 (43)

(44)
[1 I [Zcl2 [zc2 [I
(45) Fig. 9 The bus-bar of the system under considerati-on
The model developed could be extented by adding
REPRESENTATION OF THE BUS-BARS lines and busses but this will result in substantial
increase of the computation required and it could make
The equations developped in the previous section the simulation program impractical. For larger systems,
are expressed in terms of the line capacitor currents. therefore, the use of simpler models could be a better
These currents are used to define the bus voltages and solution.
1762

THE DYNAMIC SIMULATION techniques are used for accurate determination of the
firing and extinction instants of the valves.
The basic structure of the dynamic simulation
program is shown in the flow chart of Fig. 10. The THE CONTROL SYSTEMS SIMULATED -
converter control system although is independently si-
mulated is locked to the model since it uses as inputs To demonstrate the effectiveness of the model de-
the instantaneous values of the commutation voltages ve l oped two cases of converter control were simulated:
and the converter loop currents. The control system si- Case 1: Predictive control (13)
mulation program processes these inputs and produces Inverter : Constant extinction angle control
the firing pulses to the valves. The system model Rectifier: Constant current control
checks the necessary condition, at the firing instant Minimum ignition angle 50
and if they are favorable the valve is fired. At this Constant extinction angle 100
instant the main program recalculates the incidence Constant current slope 400 Ohm
matrix (Cn) and the differential equations are reas- Rectifier current setting 1174 A
sembled to account for the new conduction pattern. The Current margin 509 A
matrix (Cn) is also recalculated when a thyristor cur-
rent approaches to zero due to extinction of a valve. Case 2: Equidistant control (10)
The loop impedance matrix of the system is stored Inverter : Minimum extinction angle control
and inverted using sparsity techniques in order to re- Rectifier: Constant current control
duce the storage and inversion time required. The correction of the firing angle at the
For the integration of the equations a 5th order inverter is given by:
Runge-Kutta method with error detection is used. Theba-
sic integration step is 0.1 msec but special simulation La= K(&5min-6o)
Read circuit data and where: 5min is the minimum extinction angle per cycle
calculate the initial values 60 is the extinction angle setting
for all the state variables K is taken equal to 0.02
This control is combined with a protection scheme
Form (C44) for initial configuration against commutation failures (2) which fires valves out
of sequence in order to improve the development of the
T = Tstep fault and advances temporarily the firing angle for
further protection.
DYNAMIC SIMULATION RESULTS
The following sequence of events is simulated:
The system is operating in the steady state when a
single phase short circuit is inserted on phase a 50 Km
from the inverter bus. Two cycles later the circuit
breakeriis activated and at the first current zero the
fault is cleared. The circuit breaker is reclosing
successfully after 5 cycles.
The transient performance of the system for the
two cases is shown in the Figures 11 and 12.
150r

Calculation of ac,f3
and all the state variables

-1401I r
c
(c)

Fig. 11 Case 1: (a) Commutating voltages of the recti-

fier bus (b) The firing angles of the converters
Fig.lO The flow chart of the dynamic simulation program (c) The dc voltage across inverter bridge 1
 1763
150
(4) "Analyzing Transients in AC-DC Systems with the
BPA Electromagnetic Transients Program", EEE
International Conference of Overvoltages and Com-
pensation on Integrated AC-DC Systems.
(5) C. N. Weygandt, and R. F. Chu," Resent results on
the stabilizing effect of a DC transmission link,'
-150 (a) presented at the 1977 Winter Meeting,New York, NY,
180 Jan. 30-Feb. 4, Paper No. F 77 061-5.
(6) D. B. Goudie," Steady-State Stability of paral lel
h.v. AC-DC power transmission systems," Proc. IEE,
m90 Vol. 119, No. 2, pp. 216-224, February 1972.
-~ o

64%. (7) S. Williams and 1. R. Smith,"Fast digital computa-

tion of 3-phase thyristor bridge circuits i" Proc.
O
I
IEE, Vol. 120, No. 7, pp. 791-795, July 1973.
C5
70 (8) J. Milias-Argitis and G. Galanos," Dynamic simula-
tion of HVDC transmission systems," IEEE Transac-
tions on Power Apparatus and Systems, Vol. PAS-92,
-the May/June 1976.
-70
(9) J. Milias-Argitis, G. Giannakopoulos , G. Galanos,
140 "Dynamic simulation for multiterminal HVDC sys-
Fig. 12 Case 2: (a) Commutating voltages of the recti- tems, " IEEE Transactions on Power Apparatus and
fier bus (b) The firing angles of the converters systems, Vol. PAS-97,pp. 587-593,March/April 1978.
(c) The dc voltage across inverter bridge 1
(10) J. Arrillaga, G. Galanos and E. T. Powner, "Direct
Digital Control of HVDC Converters", IEEE Transa-
CONCLUSIONS ctions on Power Apparatus and Systems, Vol. PAS-89
No. 8, pp. 2056-2065, Nov. Dec. 1970.
The paper presents an exact and detai model for
dynamic simulation of HVDC systems. Some of the results (11) "Digital Simulation of DC Lines and AC Machines"
obtained using the model are presented in the paper.
The analysis of the results shows that both the model D. A. Woodford, A. M. Gole, R. W. Menzies, IEEE
and the dynamic simulation program are correct and ac- PES 1982 Summer Meeting.
curate.
All types of converter faults, line and bus faults, (12) t;. K. Carter, C. E. Grund , H. H. Happ , and
unbalances, load changes, line switching and autore- R. V. Pohl, "The dynamics of AC-DC systems with
closing can be easily simulated as a result of the controlled multiterminal HVDC transmission," IEEE
formulation of the model using tensor techniques. Transactions on Power Apparatus and Systems Vol
Although a three phase representation of the sy- PAS-96, pp. 402-413, March/April 1977.
.

stem is used, by applying sparsity techniques both the

computer storage and the computation time required are (13) N. G. Hingorani and P` Chadwick, "A new constant
subs tant i a y reduced. extinction angle control for AC/DC/AC static con-
The basic disadvantage of the model is the sub- vertors", IEEE Transactions on Power Apparatus and
stantial amount of computation required and is, there- Systems, Vol. 87, pp. 866-872, March 1968. Disc.,
fore, suitable only for studies that require detailed p. 872.
representation of the system. Such studies could be the
testing of converter control systems and the optimiza- (14) J. Arrillaga and A. E. Efthymiadis, "Simulation of
tion of their parameters. convertor performance under unbalanced conditions"
The model developed could be extented by adding Proc. IEE, Vol. 115, No. 12,pp. 1809-1818,December
lines and busses but this will result in substantial 1 968.
increase of the computation required and it could make
the simulation program impractical. For larger systems, (15) D. B. Giesner and J. Arrillaga, discussion on,"Be-
therefore, the use of other models could be a better haviour of HVDC links under balanced-ac-fault con-
solution.
ditions," Proc. IEE, Vol. 119, No. 6, pp. 700-704,
December 1971.
REFERENCES
(16) J. K5uferle, R. Mey, and Y. Rogowsky , "HVDC sta-
(1) H. A. Peterson , P. C. Krause , J. F. Luini , and tions connected to weak AC systems," IEEE Transac-
C. H. Thomas, "An analog computer study of a pa- tions on Power Apparatus and Systems, Vol. PAS-89,
rallel AC and DC power system," IEEE Transactions pp. 1910-1917, September/October 1970.
on Power Apparatus and Systems Vol. PAS-85 , pp.
,

191-209 , March 1966. (17) H. P. Lips, and H. Ring, "The performance of AC

systems with predominant power supply by H.V.D.C.
(2) J. Arrillaga and G. Galanos, "Fault-development inverters," IEEE Transactions on Power Apparatus
control in a.c.-d.c. convertors", Proc. IEE, Vol. and Systems, Vol. PAS-94, No. 2, pp.408-417,March/
116, No. 7, pp. 1201-1208, July 1969. April 1975.
(3) H. A. Peterson , and P. C. Krause , Damping of (18) N. A. Vovos and G. D. Galanos, " Damping of Power
power swings in a parallel AC and DC system," IEEE sw^ings in ac tie lines using a parallel dc link o-
Transactions on Power Apparatus and Systems , Vol. perating at constant reactive power control", IEEE
PAS-85 pp. 1231-1239, December 1966.
,
Trans. on PAS, Vol. PAS-98, pp. 416-425,TA 1979.

View publication stats