Sunteți pe pagina 1din 9

247

SUPPLEMENTARY DC CONTROLIN A TWO AREA AC/DC POWER SYSTEM

J.M. castellanos

SaSkPowex

Regina, sask.

RJ. Fleming

PowerSystemResearch

Group

university of Saskatchewan

saskatoon, sask.

Abstract:Thensearchreportedinthispaperconcernsthe

performance of supervisory control associated with a high voltage

Two Area Power System interconnected by HVAC and HVDC tie lines. The system includes supplementary DC control interacting with the conventional AC area controls. TaLing advantage of the fast controllability of DC links, supplementary control utilizing DC power modulation could simultaneously improve system dynamic stability,reduce area control error and inadvertent energy interchange. The supplementary DC power modulation is based on variation of the net power interchange due to step load disturbances in one of the control areas. A linear model was used for the power system to study small

changes around a base refmce state and only the power-

fresuency control aspects of the 0veraU power system control problems were considered. Eigen-structure analysis of the state space model of the power system was used to evaluate system stability. Control aptimization was achiewd by minimization of a performance index which is a function of the Area Control Error (ACE). Computa simulation was used to obtain powcr system dynamic rrsponses to step load disturbances and to estimate the optimal controller parameters.

kb”s

area Capacity (power base), MW

nominal system frequency, Hz

subscript referring to mtrol area (i=1,2)

generated powex deviation, pu

net power interchange deviation, pu

change in load demand (load disturbance), pu

frequency deviation, Hz

WAC tie line power deviation,

HVDC tie line power deviation, pu

AGC control signal

supplanentary control signal

area control mor (API2+ B, Afi), pu

speed governor rime constant, s

pu

geon regdarion (h”a*

‘c, Wpu

steam chest time constant, s

inertia constant, s

load frequency characteristic, plmz

&, = 1Di, Tpi= 2 FI,/fDi)

~frequencynsponse- ‘c (14 +DJ,pU&

synchronising coefscient, pu/Hz

DC system time constant, s

kpr

k,

kp

kl

s

PropOItional gain of AGC controller

integral gain of AGC controller

proportional gain ofsupplementary controller

integral gain of supplemen*

Laplaceoperator

controller

InMuctioa

~nthe last three ~esse~eralst~dies’~~*‘~*~~’~~avebeen conducted to emphasize the performance improvements of power

systemsbecause of HVDC controllability. Some of those studies

are related to the development of suitable models for studying hybrid AC/DC systems and a more detailed simulation of power

systems.In others, the influence of DC power modulation on the

stability of AC/DC power systems for various control schemes

hasbeen researched.

The

of this research was to extend the studies reported

above to investigate how HVDC control could be used to

improve the OVQau power system dynamic performance.

The requirements

for amtrol of frequency

and

net power

interchange, in a Two Area Interconnected Power System

(TAIPS), are implemeated bythe Automatic Generation Control

(AGC).

The

AGC control provides

automatic

variation

of

generation set points on the speed governors to keep system frequency and net power interchanges within scheduled values.

Usually HVDC systemsare opgated with constant power control

in accordance with a fixed scheduled DC power transmission; on

to load variation (small disturbance) from

time to time on the AC systems, AGC control action is required

to match the generation to power demand within predefined

fresuency limits.Utilidng the fastpower control charac“* ‘c of

controller to

modulate the DC power transfer, based on an AC system signal (e.g. net power interchange deviation from its scheduled value) it should be possible to simultaneously achieve the following objectives:

the other -,due

HVDC

systems with a suitable supplementary

i) improve system dynamic stabiity providing a positive

damping for the natural made of system oscillations,

5)

iii)

mini” the area control (ACE) in both AC systems to

avoid excessive governing system operation, and

reduce the inadvertent energy interchange 0

-

neighbouring powa Symms (control areas).

between

A simplified line diagram of an i”ected

power

system and its functional control scheme is illustrate in Figure 1.

0-7803-1319-4/93/$3.00 @ 1993 IEEE

AREA1

U

PL1

fi

! -

=

7 I

AGC system

control

I PAC L.

P12

248

Figure 1. The Two Area Interconnected Power System (TAPS).

The power system pmposed is composed of two power system areas interconne~ted by parallel HVAC and RVDC tie lines.

Each AC power system is represented by thermal generation (a turbindgenezator), area load (PJ and system inmdon bus. The DC link includes two converter stations, one performing asanztik at the sending end, and the otha performing as an inverkr at the receiving end.

The basic control system includes the speed governing system, the AGC control system and the DC system control. A supplementary control is added, to the basic control scheme, to introduce a new control signal (VJ to the DC system.The AGC control of the TAPS is centred around 'tie line bias control strategYw9.TheAGC control senses frequency deviation and net power interchange deviation to compute the ACE signal, which is processed by the AGC controller (PI controller) to provide the control signal Ui(reference) to the speed governing system.

Ui=h A=,

+4JACEidt

(1)

The supplementary DC umtroller (PI controk) provides a new control signal U, (reference) to the constant DC power system

control to modulate the DC power flow based on the net power

interchange deviation fromitsscheduled

value.

"he derivation of the model for the TAPS is based on the

following assumptions:

i) the electrical in"ch 'ons within each individual control area are very strong, compared with the ties between neighbowing areas, such that each area could be Charactenzed* by a single frequency and can be represented by one equivalent machine;

E) the voltage magnitudes at the high voltage AC buses were

I

AREA2

= 50 U2

PL2

h

control

maintaiued oonstant by voltage regulation action;

iii) incrernaral mechanical and electrical losses were

amsidered to be negligible, including those of the AC and DC links.

A smaIl perturbation transfer function block diagram of a two

area thermal system,interamnected

lines is shown in Fi 2. The fast supplementary

interacts with the conventional AGC controller to restore system hquency and net power interchange to scheduled values (zero steady state area control error).

via HVAC and HVDC tie

control

tatiQQ

The stak space equations of a two areas power system

interconnected via parallel AC and DC links can be derived from the dynamic model shown in Figure 2. The state variables are:

:frequency deviation, area 1

:deviation in generated power, area 1

:deviation in governor position, area 1 :fkquea~cydeviation, area 2

: deviation

in generated power, area 2

:deviation in govmor position, area 2

:AC power interchange deviation

: DC power interchange deviation

: control signal, area 1

: control signal,area 2

: control signal DC system

: change in load (disturbance), area 1

:change inload (man=),

area 2

r

Q

I

A

249

Figure 2.Transfer frmction block diagram ofthe AClDC TAIPS.

X=[x,

%

xloxll]= is the state vector,

V=[v,

v3'

is the

disturbance vector.

Translating the transfa functions (Figwe 2) which use Laplace notation into the state space form, tbe dynamics of the model are

rqrcsented by the aet of diffexential equations as follows:

j$ = --X1--%---.$ 1

11

%TGl

TGl

TGl

9

Theref~thesystemsh~inFigure2canberrprcsentedby

the vector space equatian:

x=M(+Fy

(14)

AisthesLatecoeaclea-

coefficient matrix, whose coefficlen' tsareobtainedfromthestate

space equations.

t matrix and Fistbe disturbana

A=

F=

1

-

TPI

0

1

-

R1 TGI

0

0

0

TI2

0

-42T12

kPT12

%I

--

TPI

0

0

0

0

0

0

0-

0

0

0

0

0

0

0

Kpz

--

TPZ

0

0

0

-kprB2Kpl

TPZ

0

0

0

1

-

TTI

--

1

TGl

0

0

0

0

0

0

0

0

--

0

0

0

1

-

TP2

0

1

R2TG2

-TI2

0

-kP 1T12

250

0

0

0

k!

-

TP2

1

-

TT2

0

0

0

0

The formal optimal control approach includes all systcm

state variables in the optimal feedback amtrol; however, in this

all the system state variables in

study, it is not feasible to include

the feedback loop because of the unavailability of all of the state

variables in this case. Moreover, the configuration of the system

0

0

0

0

1

-

TT2

-

1

-

TG2

0

0

0

[-LPlBIKfl

TPl

+141

- U1

0

4

[-kp1(-+-)

Bl%

TPI

+41

-1421

1

TDC

--+kILP '

TDC

0

0

1

-

TGl

0

0

0

00

0

0

00

--

0

0

1

TG2

0

0

0

000

0

0

0

0

0

0

1

TDC

kP1

TDC

--

--

--

Irp

TDC

controllers ispredeterrmn. ed (proportional plus integral control).

Thdm a pseudo optimal control was utilized in this research,

based only on the optimization of the controller parameters to minimize a selected performance index which includes a subset

of the state variables. The performance Criterion chosen must lead to the minimisation of the Area Control Error (ACE) in both control areas. The performance index selected is

represented by the time integral of the sum of the square of

control error (ISACE) in

in Area 2, averaged over a time period T.

Area 1 and the square of control error

ISACE =

1 I[(ACE,)' + (ACEJ'] dt

To

(17)

The performance index of Equation (17) is based directly on

generation requirements

a good measure of control effectiveness.

in both control areas, which represents

The performance index minimisation was carried out by a sub

routine called parameter Estimation of the TUTSIM program.

This subroutine minimizes the ISACE function by adjusting the controller parameters following a time response computer simulation. In the computation of the optimal parameters a

constraint of system frequency deviation was also included. The peak value of the first Swing of the system frequency deviation

was not allowed to exceed 0.025 Hz. This

is an arbitrary limit

based on accepted North American practice.

The estimated optimal AGC controller parameters and supplementary controller parameters, based on the system parameters shown in Appendix A, are:

251

*

lrpl =&

= 0.3 ; k,, = kn = 0.6

* kp = 0 ; k, = 12.0

sv-

w!z!&

The stability of an i"nd power system is

measured in tenas of its capability to return to normal or stable operation ah having been subjected to some lrind of

disturbance. Dynamic stabiity refers to the stabiity of a power system subject to a relatively small and sudden or periodic

disturbance, including the effect of automatic control devices. In

gend, natural oscillation modes are in the fresuency range of

0.01 to 0.2 Hz, assockd

with gwemors, and 0.2 to 3 Hz,

associated with exchanges of power between generating plants in a system.

sms

The system charactens. tic stability is a property of the system itself and does not depend on the input or driving

function of the system. The stability of a linear closed loop

system can be detemhed from the location of the closed loop poles in the s plane. Ifall closed loop poles lie to the left of the jo axis,any transient response eventually reaches equilibrium.

This represents a stable system. The fact that all closed loop

poles lie in the left half of the s plane

does not guarantee

satisfactory dynamic response characten'tics.s If dominant complex poles are located close to the jo axis, the dynamic

response may show excessive oscillations or a long settling time.

It is known that the eigenvalues of the state space dcient

matrix are the closed loop poles of the model system. For the

system npresented by the state equation :

x = AX+FV

(18)

the eigenvalue spectrum is the set of roots of the -c aption

:

-

'c

I U-AI = 0

(19)

The eigenvalues determine

the time domain characten'slic

of

every dynamical modes of the system response.

All the eigenvalues of the basic AUDC system without SC lie in

left side of the complex plane; this represents a stable system.

the

The dominant polcs of the system are the complex conjugate pair A,, A,. The osciuatiOn frequency of these modes is 3.76rad/s#:

Table 1. Eigenvalues of the AC/DC system.

mode. The dominant complex poles (A8, &) are moved to the left, from the previous location (-0.18 to -1.17), providing a positive damping to the system responses. Also, the oscillation frequency of this mode has deaeased slightly (3.76 to 3.09). Mode A, is slightly StllJitive to SC pmmeten variation. Fm 3and4illustrate thee&ctofSCcontrolla gainvaxiation onthe

dominant pole locatianS.

Q

U

//b:

b

x

kp = 0.0

10.0

c kr = 0.5

b: kI = 13.5

113.6

a.3.4

~~3.2

3.0

::2.*

2.6

13.2

**3.0

252

Figure 3 shows that A, moves to the left side in the s plane as

proportional gain is increased. The damping is increased and the oscillation frequency is decreased. Figure 4 shows that l9moves

left in the s plane as integral gain is increased. For low

values of gain, oscillation frequency is increased slightly and after

that it decreases faster around the optimal gain. The damping is continuously increased until the gain is a little less than optimal gain, after that starts to decrease slowly.

to the

The variation of supplementary controller pammetefi on system stability show improvements in system damping and oscillation frequency reduction as proportional gain is increased and the integral gain is set in a range close to its optimal value.

Svstem Dvnamic Re-

Simulation studies can be used to analyze and evaluate the

dynamic performance of a system. In

simulation process was canied out using a computer program called TUTSIM. It is an interactive program which simulates the dynamics of systems. The "UTSIM system accepts models in the form of block diagrams. It simulates these models and allows one to change system parameters interactively and compare resulting changes in system responses. In the research reported here the model input to the TUTSIhl system was based on the transfer

function block diagram of the Two Area Interconnected Power System (TAIPS) representation (Figure 2). Through this simulation program the dynamic response of the system was obtained for a specific input disturbance.

this research the computer

The simulated system responses were obtained for the two-area power system model assuming it had been subjected to a step

load disturbance in one of the control areas. It was assumed that each area was working independently with no mutual assistance of power transfer scheduled through the link except following a disturbance. Initially the power system was in steady state condition determined by normal system operation. That means there were no system variable deviations from their scheduled values, in other words all the state variables of the system model were zero. Subsequently the area load was increased by a step (APJ in one of the control areas, so the system model responses oscillated until they were settled down by the action of the AGC and supplementary DC controls. The supplementary DC control, the main object of thisresearch work, imposed power modulation of the DC link based on the deviation of the power interchange (AP,3 between the individual power system areas.

The inadvertent energy interchange which is represented by the time integral of the net power interchange, was also computed in the simulation process. Furthermore the simulation process was used for the optimization of controller parameters @ropomonal and integral gains) of the AGC and supplementary

DC controls,pxeviously described.

Since the smallest time constant of the system model utilized is ,T = 0.05 seconds @C system time constant) the simulation time step chosen is 0.01 seconds, which is small enough so as not

to compromise the accuracy of system responses. The time period

of the simulation was determined by the reset time of the slowest system response examined. In thisresearch work the time period of the simulation was 20 seconds.

The simulation tests fora two area powex system interconne!cted

by parallel AC and DC tie lines, with AGC and supplementary

DC control (SC), were based in the transfer

function block

diagram shown in Figure 2. The parameter values for this system are shown in Appendix A.

Simulation Results

A set of simulation results is presented for different

interconnection types and control system characteristics of the two area power system. These simulation tests were canid out

to compare system dynamic responses under similar Operation conditions of the power system. The following cases were considered in the simulation prm:

i) Two area power system interconnected by two AC tie lines, with AGC control.

ii)

iii)

iv)

Two area power system inter~~nne~tedby AC and DC tie lines, with AGC and constant DC power controls.

Two area power system interconnected by a AC and DC tie lines, with AGC and Supplementary DC controls.

Two area power system interconnected by AC and DC tie lines,with AGC control and supplementary DC control including DC power limiter.

and DC links were set to have the same capacity (1096 of

area capacity). The interco~ected power system was operating at a power angle of U)" between system areas and the power transfer between artas was 200 Mw (10% of area capacity). The step load disturbance applied to the model system in Area 1was 1% (20 MW) of area capacity for all the cases. The DC power limiter of the DC control system,included in the third case, was 10% (20 MW) of Dc link capacity.

AC

The parameters of the AGC and SC controllers were optimal. These values are given in a previous section with the exception of the system which included two AC tie lines. The optimal pamneters of the AGC controller for a two area power system ink~~~~e~tedbytwo AC tie lines are: &,=kp2=0; k,,=k,=0.5.

Figures 5 to 11 illustrate the system dynamic responses for

various interconnection types and control systems of the

under the operating conditions described above. In the systems where no supplementary DC control is included, the system dynamic responses have practically the same settling down times in both cases.The oscillation frequency of the system responses

is higher for the TAPS interconnected by two AC lines with

AGC control than the TAPS interconnected

lines with AGC and constant DC power control. Figure 5 shows the frequency deviation responses in Area 1. By the action of SC

control the frequency deviationresponse isdamped faster,sothe

system frequency comes back to scheduled values in about 10 scconds or less. The first swing of the frequency deviation in Area 1 islarger for the TAPS which includes SC control. Figure 6 shows the frequency deviation responses in Area 2. The SC control action increases the damping of the dynamic frequency deviation response sigdicaavely. Moreover the first swing of frequency deviation in Area 2 is largely reduced. Therefore the dynamic frequeacy changes are very small in the system area which is not subjected to load distmbance. Figure 7 shows that the power interchange deviation responses experience a

significant dampiug improvement because of the SC control action. The dynamic response settles down in about 10 s for the

SC control which include DCPL, and in about 7 s for the SC

control without DCPL. Figure 8 illustrates that the inadvertent

TAIPS,

by AC and DC

253

energy interchange is drastically reduced by the action of the SC control, reaching a minimum value for the system without DCPL. Figures 9 and 10 show the arta amtrol enor dynamics in Area 1 and Area 2 rrspectively. Inboth control artas the control error signal isreduced (mainly in Area 2) and quickly stabilized by the action of SC control. Therefore the govaning system operation

in both areas would be mini-

because of the control error

minimization. Since the system damping is mainly provided by the faster SC control, Area 2 has very low sensitivity, to load

1. Figure 11 illustrates the performance of the

changes in Area TAPS measured

by the performance indexISACE. By the action

of the SC control the ISACE is minimized, so the control

effectiveness isinneajed such that the dynamic performance of

the TAPS isimproved.

The fast supplementary DC amtrol (SC), based on net power interchange Variations, interacts with the conventional AC controls providing a positive damping to the dynamic system responses. The responses settle down faster, reaching a steady

state condition in about 10 seconds.

-1

I

I

I

I

I

I

0041 &&&

I

14

Figure 5. Frequency deviation

responses in Area 1.

I

I

I

I

I

oe40

I

I

14

I

I

I

io

I

10

I

I

i

m

I

m

-ao!

I

' &&A

004s

14

'

i

IO

I

18

1

Figure 7. Net power interchange deviation responsts.

- ,\ \
Y

d. -1aOj

I

;\

I

\

1.

9 :

.-.,

i.

* a.

\,i<.- a,-*,

I

oe4r

'

*-

"

c

&&A

e-

14

'

.C

-2.2.

,

ir

m

eo

Figure 8. Inadvertent energy interchange.

-MQ1

-0.ow-l

oe4r

I

I

' L&&'

I

1

l4YlOSo

ooo1,

9.a

'U

0.0

i:

254

Figure 11. Performance index ISACE.

conclusions

It

was

found

that

the

stability of the TAPS

was

significantly improved with the implementation

supplementary control. The system damping is increased oscillation frequency of the system responses is reduced.

of the optimal

and the

Simulation results illustrate a signifkxnt dynamic performance

improvement ovd for the TAPS with supplementary control. The DC power modulation, based on the time integral of the net power interchange, intesaCting with the AGC control makes it possible to restore net power interchange and system frequency back to their scheduled values in less than 10 seconds following a disturbance (change in load). Furthermore, the control optimization, minimizes the area control errors (reducing system governors operations) and reduces the inadvertent energy interchange IEI.Thus the control effectiveness is improved.

Since the system damping is provided mainly by the faster SC, the control area which is not subjected to load disturbance is practically insensitive to load changes in the neighbouring control

area.

The effects of the DC power limiter on system responses show a small influence on the performance index value, but it causes an increase of the inadvertent energy interchange. It increases the mutual assistance between the individual power systems through the AC tie-line and hence demands larger angle swings between

areas.

The integral action of the supplementary control causes a significant steady state deviation of the power transfer over the AC and DC links; one increases while the other decreases. Further research is required on this problem to devise a suitable means to compensate for or avoid it. It is suggested that the addition of a relatively slow "wash-out" action to the supplementary controller would accomplish this.

The values of optimal parameters of the supplementary controller depend on the size of the step load disturbance and of the value of the DC power limiter setting. This is due to the basic non-linear characteristic of the system and on-line tuning of such param- would be required to maintain optimal performance over a wide range of operation. This area requires further study.

References

1.

'Power

Springer-Verlag, klin/Hidelberg,

E. lhlman,

Transmission

by Direct

1975.

Cumt',

2. J. Arrillaga, 'High Voltage Direct Current Transmission', Peter Peregrinus Ltd.,London, UK 1983.

3.

IEEE Committee Report, 'Dynamic Performance

-C

of North

American

HVDC Systems for

Transient and Dynamic Stability Evaluations', IEEE

Trans.,Vol. PAS 100,No 7,July 1981,pages 33563364.

4.

R. L. Cresap and W. A. Mittelstadt, 'Small Signal Modulation of the Pacific HVDC Intertie", IEEE Trans.

Vol. PAS 95,1976,page 536.

 

5.

Y.

Yoshida and T. Machida, 'Study of the Effect of the

DC Link on Frequency Control in Interc~~ectedAC Systems', IEEE Trans. Vol. PAS 88, No 7,July 1969,

pages 10361042.

 

6.

Xia Zu-zhi, 'A

Study of Load Frequency

Control

for

Interconnected Power Systems with a HVDC Tie Line",

IEE 4" International Conference on AC and DC TransmissiOn, London, Sept. 1985,pages 416424.

7. J. J. Dougherty and T. Hillesland, 'Power System Stability

DC

conside.ration

with

Dynamically

Responsive

Transmission Lines',

January 1970.

IEEE Trans. Vol. PAS 89, No 1,

8. 0. I. Elgerd, 'Eldc Energy System Theory', Second Edition, McGraw-Hill, 1982.

9.

N.

Cohn, 'Some

Aspects of Tie Line Bias Control on

Interconnected

POWP Systems',

AIEE Trans. Vol. 75,

Feb. 1957,

1415-1436.

10.

J. M. castellanos, 'Supplementary DC Control in a Two Ana AUDC Power System', MSc. Thesis, University of Saskatchewan, 1992.

255

mdix A: Power System Parametus

area power system have

the same 'cs. So it is assume that both areas have

For the purpose of thisstudy both

equal system paramems

(equal

time

constants

and

gain

coefficients). The nominal system parameters are:

Pbpc= 2000Mw

f

Pn

Hi

Di

&

=mHZ

=2oooMw

=sMwS/Mw

=

=

8.33 x IO-)pu/Hz

2.4Wpu

TGi = 0.08s

TX

TpT

KpT

T,

Ti2

= 0.3s

= 20 s

= 120 WPU

= 0.05s

= 0.544 pu/Hz

TI, = 1.088 pu&

B, = 0.425pfi

('I" -ni

by two AC lines)

The

frequency bias constant (B3 is set qual to the area

frequency nsponse (Pi)to minimize the control action in the

control area where not disturbance

OCCUT.