excitation control system design of high response

Paper

Excitation Control System Design of High Response Excitation TypeSuperconducting Generator for Improving Power System Dynamics

Worawut Sae-Kok∗ Student Member

Akihiko Yokoyama∗ Member

Tanzo Nitta∗ Member

Superconducting generator (SCG) with superconducting field winding has many advantages such as small size, lightweight, high generation efficiency. A prominent advantage is to improve power system stability owing to lower syn-chronous reactance compared with the conventional generator. High response excitation type SCG has a rotor withthermal radiation shield without damping effect; it can enable very rapid change in field current and the magnetic fluxdue to the change of the field current can reach the armature winding quickly. The excitation power is large and hasa rapid change enough to affect the conditions of power system in self-excited operation of the generator. This effectis equivalent to the effect of a superconducting magnetic energy storage (SMES), then so-called “SMES effect”, andit is expected to contribuite to transient stabilty improvement. In this paper, excitation control system design for highresponse excitation type SCG in consideration of SMES effect is proposed for improving power system dynamics byemploying eigenvalue sensitivity. The SMES effect is considered to couple with the power system as a nonlinear load.The control performance of the proposed excitation control system is examined in IEEJ East 10-machine and West 10-machine systems. It is made clear that the SMES effect of SCG is utilized effectively and results in both transient anddynamic stability improvement. However, the performance depends on the locations of SCGs and fault contingencies.

Keywords: superconducting generator, high response excitation, excitation power, SMES, power system stabilizer, excitation sys-tem

1. Inroduction

Superconducting generator (SCG) with superconductingfield winding offers various advantages, such as high effi-ciency, small size, light weight, and so forth. A promi-nent advantage is to improve power system stability owingto the lower synchronous reactance compared with the con-ventional generator (1). High response excitation type SCGhas a rotor with thermal radiation shield without dampingeffect for screening time-varying magnetic flux to enter thefield winding part. It can enables very rapid change in fieldcurrent and the magnetic flux due to the change of the fieldcurrent can reach the armature winding quickly. The excita-tion power by the quick change of field current is large andhas a rapid change enough to affect the conditions of powersystem in self-excited operation of the generator (2). Since thefield winding is made of superconducting wire and is con-nected to the generator terminal through AC-DC converter,the field winding coupling with the converter can be consid-ered as superconducting magnetic energy storage (SMES);hence, the effect of the exciter coupling with the AC-DCconverter is called “SMES effect”. It is expected that highresponse excitation type SCG using appropriate excitationcontrol system can improve power system dynamics by ef-fectively utilizing the SMES effect.

In single machine to infinite bus system, influence of SCG

∗ Department of Electrical Engineering, The University of Tokyo7-3-1, Hongo, Bunkyo-ku, Tokyo 113-8656

with high response excitation on power system stability andits control system have been experimentally studied (2); highresponse excitation control of SCG with high response exci-tation for stability of superconducting field winding has beenconducted (3). In multi-machine power system, calculationmethods of the equilibrium state considering SMES effectof the power flow and dynamic simulation of multi-machinepower system including SCG with high response excitationand control system design based on energy function for SCGhave been proposed (4). It has been reported that the SMES ef-fect contributes to the improvement of system stability; how-ever, the control system is complicated and requires informa-tion from all generators in the system.

The control system design for SMES system includingthe selection of installing locations have been proposed fordamping electromechanical oscillation in multi- machinepower system (5). SMES has been considered as a nonlinearload coupling with the power system and small-signal modelof power system has been used to determine the systemeigenvalues. Eigenvalue sensitivities have been employed asindices to decide the installing location and to determine thecontrol parameters. It has been reported that the oscillationscan be damped well by the designed controller and eigen-value sensitivity has been proved as a powerful approach tocontroller design. Eigenvalue sensitivity based eigenvaluetechnique for control parameter optimization has also beenproposed in Ref. (6) and it has made clear that the designedcontroller is effective in stability improvement.

In this paper, excitation control system for SCG with

1112 IEEJ Trans. PE, Vol.125, No.12, 2005

Excitation Control System of Superconducting Generator

high response excitation considering SMES effect is de-signed for power system dynamics improvement in multi-machine power system by employing eigenvalue sensitivitybased eigenvalue control technique. The SMES effect of theSCG with high response excitation is considered to couplewith power system as a nonlinear load and is taken into ac-count in control system design through eigenvalue consid-eration. The control performance of the proposed excita-tion control system in several patterns of SCG installationsis examined by digital dynamic simulations in two powersystems, IEEJ East 10-machine system and IEEJ West 10-machine system. Comparisons with conventional generatorsare also conducted. It is made clear that the SMES effectof SCG is utilized effectively and results in improvement ofboth transient stability and dynamic stability. However, theperformance depends on the locations of SCGs and fault con-tingencies.

This paper is organized as follows: Section 2 describesthe structure and model of the SCG with high response ex-citation. Seciton 3 explains SMES effect of SCG with highresponse excitation. Section 4 explains the proposed con-trol systems for SCG considering SMES effect. Section 5shows the numerical example for examining the control per-formance of the proposed excitation control system. Conclu-sions are finally drawn in Section 6.

2. SCG with High Response Excitation

2.1 Structure A typical structure of SCG is shownin Fig. 1. SCG can produce very high magnetic field by itssuperconducting field winding cooled in the cryogen suchas liquid helium, which can not be realized by the conven-tional generator. Multi-cylindrical structure which consistsof field winding, thermal radiation shield and room temper-ature damper is adopted in the extreme cold region. Unlikethe low response excitation type SCG, the thermal radiationshield of high response excitation type has no damping ef-fect; it has been thus far made of stainless steel with embed-ded pipe where the cryogen flows inside. Since the generatedmagnetic field is so large enough that it is not necessary touse magnetic iron to enhance magnetic flux; instead, air-coregeometry can be used; SCG can have stator windings with the“air-gap” type, which leads to the advantage of avoiding fluxsaturation and low synchronous reactance. Magnetic shieldis installed outside for confining flux in the machine.

Fig. 1. Structure of superconducting generator

Fig. 2. Equivalent circuit of SCG with high responseexcitation

2.2 Equivalent Circuit The equivalent circuit ofSCG with high response excitation can be considered as thatof conventional generator. Since the thermal radiation shieldof SCG with high response excitation has no damping effect,it can be considered that SCG with high response excitationseems to have only one damping winding (from room tem-perature damper) in each axis, assume to be kd and kq, equiv-alent to amortisseurs winding in conventional generator. Theequivalent circuit of SCG with high response excitation isshown in Fig. 2.

3. SMES Effect

3.1 Concept SCG with high response excitation en-ables the rapid change of field current; the excitation powerbecomes large and changes rapidly enough to affect the con-ditions of power system in self-excited operation of the gen-erator. The simplified model of SCG with high response self-excitation is depicted in Fig. 3. The field winding is con-nected to SCG bus through the AC-DC converter. Since thefield winding is made of superconductors, the field windingcoupling with the converter can be considered as SMES andits effect is thus called “SMES effect”.

Like conventional generator, terminal voltage of rotor cir-cuit v f has the relation to e f as shown in Eq. (1).

v f =Rf

Xd − Xle f =

1T ′doω0

Xd − Xl

Xd − X′de f · · · · · · · · · · · · · (1)

And the excitation current is given in Eq. (2).

i f =eq1

Xa f=

eq1

Xd − Xl· · · · · · · · · · · · · · · · · · · · · · · · · · · · · (2)

The excitation power in SMES effect, which is the multi-plication of terminal voltage of the rotor circuit and the exci-tation current, is given in Eq. (3). It should be noted that theexcitation power of the SCG is smaller than that of the con-ventional generator in the steady state due to the very smallresistance of superconducting field winding.

Pf = v f i f =e f eq1

T ′doω0

(Xd − X′d

) · · · · · · · · · · · · · · · · · · · · (3)

where v f , i f , and e f are the applied voltage at rotor circuit,excitation current and excitation voltage, respectively. eq1 isthe generator internal voltage.

From the simplified model of SCG in self-excitation modein Fig. 3, the excitation system including the AC-DC con-verter can be considered as a nonlinear load. Active and re-active powers flowing into excitation system are assumed as

電学論 B，125 巻 12 号，2005 年 1113

Fig. 3. SCG with high response excitation in self-excitationmode

Fig. 4. Equivalent model for SCG with high responseexcitation considering SMES effect

Pex and Qex, respectively; hence, the equivalent model forSCG with high response excitation can be obtained as shownin Fig. 4. Without any loss in transfer power, active power Pex

is the same as the excitation power of SCG; reactive powerQex can be adjusted by self-commutated type converter forsupporting the generator terminal voltage.

3.2 Consideration of SMES Effect in Power SystemAccording to the assumption of SMES effect as a nonlinear

load in the power system, the relationship between currentflowing into excitation system and bus voltage at the SCGwith high response excitation in self-excitation mode can beobtained by Eq. (4) in d and q axes with consideration of Pex

and Qex.[IexD

IexQ

]=

1

V2t

·[

Pex Qex

−Qex Pex

] [VD

VQ

]· · · · · · · · · · (4)

where VD and VQ are d and q-axes terminal voltage compo-nents of SCG bus, respectively.

Assuming that IG is the vector of currents through gener-ators, Iex is the vector of currents of all nonlinear loads con-necting to generator buses, which includes the SMES effect,VG is the vector of generator bus voltages, VL is the vectorof load bus voltages, Y is the admittance bus matrix whichconsists of four sub-matrices YGG, YGL, YLG, and YLL, and allloads at non-generator buses are constant impedance types,the relationship of currents and bus voltages of power systemin a matrix-form is given in Eq. (5).[

IG − Iex

0

]= Y

[VG

VL

]=

[YGG YGL

YLG YLL

] [VG

VL

]

· · · · · · · · · · · · · · · · · · · · (5)

Rearranging Eq. (5);

IG = Y ′VG + Iex · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · (6)

Y ′ = YGG − YGLY−1LLYLG · · · · · · · · · · · · · · · · · · · · · · · · · (7)

Considering impedances and internal induced voltages ofall generators, generator current can be rearranged as shown

in Eq. (8).

IG = YGEG + KLIex · · · · · · · · · · · · · · · · · · · · · · · · · · · · · (8)

YG = (I + Y ′Z′)−1Y ′ · · · · · · · · · · · · · · · · · · · · · · · · · · · · (9)

KL = (I + Y ′Z′)−1 · · · · · · · · · · · · · · · · · · · · · · · · · · · · · (10)

where EG is the vector of internal induced voltages of genera-tors and Z′ is the matrix of internal impedances of generators.

By the above equations, SMES effect can be taken into ac-count in multi-machine power system. It should be notedhere that when there is no nonlinear load at any generatorbuses, Iex term appearing in Eqs. (6) and (8) will be neglectedand can be omitted from those equations.

4. Excitation Control System Design

4.1 Design Concept In this section, the excitationcontrol system for SCG with high response excitation is pro-posed. The excitation control system design of SCG withhigh response excitation is constrained by two issues: (1) thetransient open-circuit time constant T ′do being relatively large(even if in practical it may become smaller from the structureof high response excitation) and (2) the SMES effect.

Initially, AVR and speed governor are assumed to be usedin SCG as the 1st order time-lag transfer function controllerblocks. The first constraint leading to the slow response andlonger storing of energy in the field winding, requires thelarger AVR gain in SCG to step up the excitation voltage toceiling voltage level so fast that the voltage regulation can beachieved; however, large AVR gain results in degradation ofpower system stability. Power system stabilizer (PSS) is ageneral solution for this problem. The main function of PSSis to add damping to the generator rotor oscillation by con-trolling the excitation using auxiliary stabilizing signal(s). Itis expected to improve both transient and dynamic stability.By this reason, AVR and PSS will be used as excitation con-trol system for SCG with high response excitation. The con-trol scheme and design is described below.

4.2 Power System Stabilizer Output power is usedas an input of PSS which consists of stabilizer gain, signalwashout and phase compensation block as shown in Fig. 5.The PSS output is added to the AVR as a supplementary con-trol signal. Excitation limiter is required for the AVR outputto restrict the level of excitation voltage due to insulation andequipment limitation. Parameters of PSS are determined byeigenvalue sensitivity based parameter optimization methoddescribed in the next section.

4.3 Eigenvalue Sensitivity based Parameter Opti-mization The state equations shown in Eqs. (11) and (12)are obtained by linearizing nonlinear differential equations ofpower system around an operating point.

AG xg + BG∆VG = CG xg · · · · · · · · · · · · · · · · · · · · · · · · (11)

Y ′∆VG = DG xg = ∆IG · · · · · · · · · · · · · · · · · · · · · · · · · (12)

where xg is the vector of state variables of generators, ∆VG isthe vector of deviations of generator terminal voltages.

By eliminating ∆Vg from Eqs. (11) and (12),

xg = Axg · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · (13)

A = A−1G

(CG − BGY ′−1DG

)· · · · · · · · · · · · · · · · · · · · · (14)



Fig. 5. Power input type power system stabilizer (P-PSS)

Eigenvalue sensitivity can be evaluated by Eq. (15).

S i j =∂λi

∂α j= vT

i∂A∂α j

ui · · · · · · · · · · · · · · · · · · · · · · · · · ·(15)

where S i j is the sensitivity of ith eigenvalue (λi) with respectto parameter α j, A is the state matrix, vT

i and ui are the leftand right eigenvectors associated to λi, respectively.

This parameter optimization method employs eigenvaluesensitivity to form an objective function expressed in Eq. (16)and parameters can be evaluated under the conditions that alleigenvalues have their own real parts less than the thresh-old value sD (<= 0). It is constrained by limits of param-eter changes and parameters as shown in Eqs. (17) and (18)and all observed eigenvalues are guaranteed to be in LHP byEq. (19) (6).

min f =p∑

j=1

⎡⎢⎢⎢⎢⎢⎢⎣∑i∈UC

Re(S i j

)⎤⎥⎥⎥⎥⎥⎥⎦∆α j · · · · · · · · · · · · · · · · · (16)

∣∣∣∆α j

∣∣∣ ≤ dj · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · (17)

α j,min ≤ α j + ∆α j ≤ α j,max · · · · · · · · · · · · · · · · · · · · · (18)p∑

j=1

⎡⎢⎢⎢⎢⎢⎢⎣∑i∈UM

Re(S i j

)⎤⎥⎥⎥⎥⎥⎥⎦∆α j ≤ 0 · · · · · · · · · · · · · · · · · · · · · (19)

where p is the number of control parameters, d j (> 0) is thebound value of step change of α j, α j,min, α j,max are the mini-mum and maximum values of α j, UC is the set of controlledeigenvalues, UM is the set of observed eigenvalues.

4.4 SMES Effect as Load Model The second con-straint to be considered in excitation control system designis SMES effect. As shown in Section 3.2 that SMES effect isconsidered as a nonlinear load with current flowing out of thegenerator terminal as given in Eq. (6), the variation of thosecurrents can be obtained as shown in Eq. (20).[∆IexD

∆IexQ

]=

⎧⎪⎪⎨⎪⎪⎩−2

V4t0

·[

Pex0 Qex0

−Qex0 Pex0

] [V2

D0 VD0VQ0

VD0VQ0 V2Q0

]

+1

V2t0

[Pex0 Qex0

−Qex0 Pex0

]⎫⎪⎪⎬⎪⎪⎭[∆VD

∆VQ

]

+1

V2t0

[VD0

VQ0

]∆Pex +

1

V2t0

[VQ0

−VD0

]∆Qex

· · · · · · · · · · · · · · · · · · · (20)

Since Pex can be evaluated in terms of generator variablesas shown in Eq. (3), the variation of excitation power can bedetermined by Eq. (21).

Fig. 6. Reactive power controller

∆Pex = v f 0∆i f + i f 0∆v f · · · · · · · · · · · · · · · · · · · · · · · · ·(21)

The ∆i f and ∆v f are considered as state variables of gen-erators (SCG). SMES effect in the form of current can beconsidered by combining Eq. (21) into Eq. (20); if the reac-tive power of SMES effect Qex is omitted, the relationshipof the change of currents flowing into the nonlinear loads atgenerator buses, the changes of bus voltages, and the statevariables can be written in matrix form as shown in Eq. (22).

∆Iex = Yex∆VG + FG xg · · · · · · · · · · · · · · · · · · · · · · · · ·(22)

From Eqs. (6) and (22), (12) is changed by the consider-ation of SMES effect to the form as shown in Eqs. (23) and(24).

Y ′∆VG +(Yex∆VG + FG xg

)= DG xg = ∆IG · · · · · · (23)(

Y ′ + Yex)∆VG = (DG − FG) xg · · · · · · · · · · · · · · · · · (24)

Although Eq. (21) is used to represent the excitation power,it should be noted here that the change of excitation currentof SCG with high response excitation in the transient state isso very high and in turn of the excitation power.

The state-matrix of power system in Eqs. (11) and (24) areevaluated to determine the eigenvalues and eigenvalue sen-sitivity based parameter optimization method is employed todetermine optimal control parameters.

As for Qex, it can be adjusted by self-commutated type AC-DC converter according to the external control signal value.It can be set to be zero (without control) and that term can beneglected.

4.5 Excitation voltage of SCG The per unit systemused for the excitation voltage e f is the non-reciprocal perunit system; 1.0 per unit excitation voltage is assumed to bethe field voltage required to produce the rated synchronousmachine armature terminal voltage (1.0 p.u.) on the air-gapline. It depends on the ratio of field circuit resistance Rf andthe linkage inductance Lad. Since the resistance of field wind-ing circuit of SCG is very low, the excitation voltage in perunit for SCG becomes much lower than the conventional gen-erator even if their actual values are the same.

4.6 Reactive Power Control (Q-Control) As de-scribed in Section 4.4, the Qex can be adjusted by self-commutated type AC-DC converter according to externalcontrol signal. In order to support and improve voltage sta-bility, here, the simple control system in the form of the 1st

order time-lag transfer function control block with bus volt-age input as shown in Fig. 6 is applied to control reactivepower of SMES effect. Mathematical expression of the con-troller is shown in Eq. (25). It can be easily coupled into thepower system via Eq. (20) and system eigenvalues can be de-termined in the same way as described above.

∆Qex =1

Tq

(−∆Qex + Kq (Vt0 − Vt)

)· · · · · · · · · · · · ·(25)

電学論 B，125 巻 12 号，2005 年 1115

5. Numerical Examples

The digital simulations are conducted in two model powersystems, which are IEEJ East 10-machine and West 10-machine systems, to examine the performance of the pro-posed excitation control systems of SCG with high responseexcitation. In each model system, for one pattern of replacingone conventional generator (CG) by one SCG whose param-eters shown in Table 1, five cases of different generator types,excitation modes (operation), and excitation control schemesas shown in Table 2 are considered. In cases 2∼5, excitationcontrol systems are PSS-AVR (Q-Control is added only incase 5) whose parameters determined by eigenvalue sensitiv-ity based parameter optimization method.

5.1 IEEJ East 10-Machine System The modelpower system consisting of 10 generators and 47 buses isshown in Fig. 7. The structure of the system is a loop net-work with radial line connected at node 17. Three patterns ofreplacement of conventional generator by SCG with high re-sponse excitation at nodes 2, 5, and 10 are taken into account.Excitation control systems for each pattern are designed andeigenvalues of power system are evaluated; only dominant

Table 1. Parameters of SCG with high response excitation

Table 2. Generator types and control schemes

Fig. 7. IEEJ East 10-machine system

eigenvalues are considered and shown in Table 3.As shown in Table 3, comparing with case 1, when

PSS-AVRs are designed by employing eigenvalue sensitiv-ity based parameter optimization method for generators, itcan be seen that damping of system is improved well in allpatterns and all cases. When CG is replaced by SCG in cases3∼5, there are some cases that damping of one mode is im-proved but the other is not and there are also some casesthat both cases are improved and both are not. The reasonis that when SCG is introduced to the system at some loca-tions, it may itself degrade the dynamic stability of the sys-tem and coupling SMES effect by considering the exciter asa nonlinear load to the system may affect dynamic stability.

Table 3. Dominant modes of IEEJ East 10-machine system

Fig. 8. Rotor angle of generator 7




In patterns 1 and 3, the proposed excitation control systems(cases 4 and 5) can improve dynamic stability well.

As for transient simulation, in pattern 3, 3LG fault with100 ms duration is considered to occur at node 40. Gener-ator 3 is considered as a reference generator of the system.Figures 8 and 9 show the rotor angle differences of genera-tors 7 and 10, which are respectively in loop part and radialline. Figures 10, 11 and 12 show the output power, excitationpower and terminal voltage of node 10 where CG is replacedby SCG (node 10).

As shown in Figs. 8 and 9, when applying proposed

Fig. 10. Output power of generator 10

Fig. 11. Excitation power of generator 10

Fig. 12. Terminal voltage of generator 10

excitation control system (PSS-AVR) considering SMES ef-fect to SCG with high response excitation (case 4), amplitudeof oscillation in rotor angles of generators is greatly reducedcomparing with cases 1 to 3, and in case 5, where reactivepower control is applied, the amplitude of oscillation is morereduced than in case 4. From Fig. 10, after the fault is cleared,output powers and their slopes in cases 3 and 4 are very highand they get back to the steady state more quickly comparedwith cases 1 and 2; however, case 4 is better. This is thereason why the amplitudes of generator oscillation are less-ened. Although output power in case 5 does not change andget back to steady-state so rapidly as in cases 3 and 4, reac-tive power control is involved in improving system stability.It is clearly seen from Fig. 11, that the excitation power risesup to around 0.25 [p.u.] and then back to around −0.15 [p.u.]in case 4, which reflects much power to absorb (inject) from(to) the system at SCG node, mostly from (to) SCG; it resultsin increase (decrease) of output power. Excitation power incase 3 goes up to the same level as case 4, but less stabilityimprovement is obtained. This shows that SMES effect con-tribute to transient stability improvement. From Fig. 12, os-cillation of terminal voltage is improved well in case 5 fromthe effect of reactive power control.

If the fault duration time is increased to 1.0s at the samelocation (node 40) for pattern 3, cases 1∼3 becomes unstablebut the proposed control systems (cases 4 and 5) can stabilizethe power system as shown in Fig. 13. This shows the effec-tive utilizing of SMES effect for power system stabilization.

In order to assess the performance of the excitation con-trol system, an index as shown in Eq. (26) is introduced; it isthe total curve area of variables with respect to the referencevalues in all generators in a system.

AREA =∑g∈G

∫T

∣∣∣Xg − Xg0

∣∣∣ dt · · · · · · · · · · · · · · · · · · · ·(26)

where Xg is the considered variable, Xg0 is the steady statevalue of considered variable, G is the set of generators, andT is the period of time.

Here, the rotor angle difference of each generator is consid-ered as the variables for assessment and T is set to be 10 sec.In addition to the previous simulation, other fault contingen-cies are also considered for each pattern of SCG installation,which are 3LG faults with 100 ms duration at nodes 43, 45


電学論 B，125 巻 12 号，2005 年 1117

and 36. Assessment values are determined by Eq. (26) for allcases, and the results are shown in Table 4. It can be saidthat system stability can be improved well by the proposedexcitation control system (cases 4 and 5) of SCG with highresponse excitation; even if there are some contingencies thatthe control systems may not be effective.

5.2 IEEJ West 10-Machine System The modelpower system consisting of 10 generators and 27 buses isshown in Fig. 14. Three patterns of replacement of one con-ventional generator by SCG with high response excitationat nodes 1, 4, and 7 are considered and excitation controlsystems are designed as in the previous section. Dominanteigenvalues are shown in Table 5. They can be considered inthe same way as in the previous IEEJ East system.

As for transient simulation, in pattern 2, 3LG fault with100 ms duration is considered to occur at node 24. Genera-tor 10 is considered as a reference generator of the system.Figures 15 and 16 show the rotor angle difference of gener-ators 4 and 8 and Figs. 17, 18 and 19 respectively show the

Table 4. Total curve areas in IEEJ East 10-machine system

Fig. 14. IEEJ West 10-machine system

Table 5. Dominant modes of IEEJ West 10-machine system

output power, excitation power and terminal voltage at theSCG node (node 4).

It can be seen from the results that system stability is im-proved well by the proposed excitation control systems (cases4 and 5). The mechanism of how the excitation control sys-tem gets involved in system stabilization can be described bythe same way as the previous system. This shows that SMESeffect influences the improvement of system stability.

However, it can be observed in case 5 that, unlike in theEast system, the excitation power becomes very high in thetransient state; this reflects the difference of how reactive



Fig. 17. Output power of generator 4



Fig. 18. Excitation power of generator 4

Fig. 19. Terminal voltage of generator 4

Table 6. Total curve areas in IEEJ West 10-machine system

power control (Q-control) contributes to stability improve-ment in distinct system.

Other fault contingencies, which are 3LG faults with100 ms duration at nodes 21, 15 and 18, are also consideredand assessment values are shown in Table 6. It is made clearthat system stability can be improved well by the proposedexcitation control systems.

5.3 Discussion From the simulations in both test sys-tems, it can be seen that the proposed excitation controlsystem (cases 4 and 5) with consideration of SMES effectthat are designed by employing eigenvalue sensitivity basedparameter optimization method can improve both transient

stability and dynamic stability of the power system (compar-ing to case 3). Replacement of CG (cases 1 and 2) by SCGand coupling SMES effect into the power system may makethe system become less stable (dynamic stability viewpoint),but, with the eigenvalue control technique, eigenvalues arecontrolled to move to the more stable direction, the dynamicstability is then improved. However, there are also somecases that dynamic stability seems not to be finally improved,but with transient stabilization by SMES effect, the overallsystem stability is ultimately improved. It is also clearly seenthat the control performance of the excitation control systemsdepends on the locations of SCGs and fault contingencies.

6. Conclusion

Excitation control system design for SCG with high re-sponse excitation is proposed by applying AVR and PSSwhich is designed by employing eigenvalue sensitivity. Re-active power control at SCG node is also added into excita-tion control system. The SMES effect is modeled and putinto consideration when designing control system. The con-trol performance is examined in IEEJ East 10-machine andWest 10-machine power systems and the results show that,coupling with SMES effect, the proposed excitation controlsystem is effective to improve both transient stability and dy-namic stability. The control performance also depends on thelocation of SCGs and fault contingencies.

(Manuscript received Feb. 21, 2005)

References

( 1 ) J.L. Kirtley: “Large System Interaction Characteristics of SuperconductingGenerators”, Proc. IEEE, Vol.81, No.3 (1993-3)

( 2 ) T. Nitta, T. Okada, Y. Shirai, T. Kishida, Y. Ogawa, H. Hasegawa, K. Takagi,and H. Matsumoto: “Experimental Studies on Power System Stability of aSuperconducting Generator with High Response Excitation”, IEEE Trans.Power Syst., Vol.12, No.2, pp.906–912 (1997)

( 3 ) T. Nitta, Y. Shirai, T. Okada, Y. Imai, and Y. Kobayashi: “High ResponseExcitation Control of Superconducting Generator for Stability of Supercon-ducting Field Winding”, IEEE Trans. Appl. Superconductivity, Vol.10, No.1,pp.935–938 (2000)

( 4 ) F. Kawano, A. Yokoyama, and T. Nitta: “Control System Design of Excita-tion System of High Response Excitation Type Superconducting Generatorsconsidering SMES Effect for Stability Improvement of Multi-machine PowerSystem”, Proc. of 2000 Annual Conference of Power & Energy, IEE Japan,pp.17–23 (2000) (in Japanese)

( 5 ) L. Rouco, F.L. Pagola, A. Garcia-Cerrada, J.M. Rodriguez, and R.M. Sanz:“Damping of Electromechanical Oscillations in Power Systems with Super-conducting Magnetic Energy Storage Systems: Locaton and Controller De-sign”, 12th PSCC, Proc., pp.1097–1104, Dresden, Germany (1996-8)

( 6 ) Y. Sekine, A. Yokoyama, and K. Komai: “Eigenvalue Control of MidtermStability of Power System (Concept of Eigenvalue Control and its Applica-tion)”, 9th PSCC, Proc., pp.885–891, Lisbon, Portugal (1987-8/9)

Worawut Sae-Kok (Student Member) was born in Pattani, Thailandin 1980. He received B.E. from Chulalongkorn Uni-versity, Thailand in 2000 and M.E. from the Univer-sity of Tokyo, Japan in 2003. Now he is a doctorcourse student in Department of Electrical Engineer-ing, the University of Tokyo. His interest concernspower system stability and control.

電学論 B，125 巻 12 号，2005 年 1119

Akihiko Yokoyama (Member) was born in Osaka, Japan in 1956. Hereceived B.S., M.S. and Dr. Eng. all from the Univer-sity of Tokyo in 1979, 1981 and 1984, respectively.He has been with Department of Electrical Engineer-ing, the University of Tokyo since 1984 and currentlya professor in charge of Power System Engineering.He is a member of IEEJ, IEEE and CIGRE.

Tanzo Nitta (Member) was born in Hyogo Pref., Japan on August 2,1944. He received his B.E., M.E. and D.E. degreesin Electrical Engineering from Kyoto University, Ky-oto, Japan in 1967, 1969, and 1978, respectively. Heis now a professor of the University of Tokyo. His ar-eas of interest are applied superconductivity to powersystem apparatuses, electrical machines and networknet-work theory. He is a member of IEEJ and theCryogenic Engineering Society in Japan.


excitation control system design of high response

Documents