Anda di halaman 1dari 7

474

IEEE TRANSACTIONS ON ENERGY CONVERSION, VOL. 20, NO. 2, JUNE 2005

Inuence of Excitation System Control Modes on the Allowable Penetration Level of Distributed Synchronous Generators
Walmir Freitas, Member, IEEE, Jose C. M. Vieira, Student Member, IEEE, Andre Morelato, Member, IEEE, and Wilsun Xu, Senior Member, IEEE

AbstractThis paper investigates the impacts of different excitation system control modes on potential factors that can limit the maximum allowable number of synchronous generators connected to power distribution systems. Excitation systems acting as either a voltage regulator or a power factor regulator are examined. The impacts on steady-state voltage, angle stability, short-circuit currents and voltage stability are determined. Simulation results show that the usage of the excitation system as a voltage regulator can increase the maximum penetration level of synchronous generators in distribution systems. Index TermsDistributed generation, excitation system, shortcircuit, steady-state voltage, synchronous generator, transient stability, voltage stability.

I. INTRODUCTION

HE USAGE of distributed generation has increased worldwide driven by market deregulation and continuous demand rise [1][3]. Despite the fact that considerable attention has been paid to new generation technologies, e.g., fuel cells and photovoltaic arrays, nowadays, most distributed generation sites employ synchronous machines [1][3]. Although such technology is well known, there is no consensus among different utility practices on what is the best mode of controlling the excitation system of synchronous generators connected to distribution systems [2]. Typically, there are two different modes of controlling the excitation system of distributed synchronous generators. One aims to maintain constant the terminal voltage (voltage control mode) and the other one aims to maintain constant the power factor (power factor control mode) [1], [4]. This decision depends on the operational rules adopted by the utility. In view of these facts, it is important to understand the impacts of different excitation system control modes on factors that can potentially limit the number of synchronous generators connected to a typical distribution system, i.e., the maximum allowable penetration level of distributed synchronous
Manuscript received September 29, 2003. This work was supported by the National Sciences and Engineering Research Council of Canada (NSERC) and by the Fundaco de Amparo Pesquisa do Estado de So Paulo (FAPESP), Brazil. Paper no. TEC-00275-2003. W. Freitas and W. Xu are with the Department of Electrical and Computer Engineering, University of Alberta, Edmonton, AB T6G 2V4, Canada (e-mail: walmir@ieee.org; wxu@ee.ualberta.ca). J. C. M. Vieira and A. Morelato are with the Department of Electrical Energy Systems, State University of Campinas (UNICAMP), C.P. 6101, 13081-970 Brazil (e-mail: jcarlos@dsee.fee.unicamp.br; morelato@dsee.fee.unicamp.br). Digital Object Identier 10.1109/TEC.2004.841526

generators. This paper presents research results considering distinct scenarios and technical factors. The factors analyzed are steady-state voltage prole, angle stability, short-circuit currents and voltage stability. Simulation results have shown that the usage of the excitation system as a voltage regulator is the best strategy to increase the allowable penetration level of distributed generation. The paper is organized as follows. In Section II, the network component models adopted in this study are described. The impacts of the excitation system control on the steady-state voltage are analyzed in Section III. Section IV investigates the inuences on the transient stability. In Section V, the short-circuit currents supplied by synchronous generators under both control modes are determined. The impact on the voltage stability margin is addressed in Section VI. II. DISTRIBUTION SYSTEM COMPONENT MODELS In this work, all network components were represented by three-phase models. In the steady-state voltage and stability studies, the network variables were represented by phasor models, as it is usual in transient stability simulations. Whereas electromagnetic transient simulation was employed in the short-circuit current studies, i.e., the network variables were represented by instantaneous values. Distribution feeders were modeled as series RL impedances. Three-phase transformers were simulated taking into account the core losses (T circuit). Loads were represented by exponential voltage-dependent models [5], [6]. In the calculation of steady-state voltage proles, the active and reactive components of loads were assumed as constant power. Otherwise, in the calculation of short-circuit currents and stability studies, active power loads were represented by constant current models and reactive power loads by constant impedance models, as recommended in [5] for dynamic simulations. A. Synchronous Generator In the electromagnetic simulation studies, the dynamic behavior of synchronous generators was represented by an eighthorder three-phase model in the dq rotor [6]. In the transient stability simulations, such representation was reduced to a sixthorder model, i.e., the stator transients were neglected. Usually, distributed generators do not participate in the frequency regulation of the system and, therefore, they operate in constant active power mode. Thus, the mechanical power was considered constant.

0885-8969/$20.00 2005 IEEE

FREITAS et al.: INFLUENCE OF EXCITATION SYSTEM CONTROL MODES ON DISTRIBUTED SYNCHRONOUS GENERATORS

475

Fig. 1.

Generic excitation system.

Fig. 2. Single-line diagram of the system 1.

B. Excitation System A functional description of excitation systems acting as a voltage or power factor regulator is provided in [4]. Therefore, only the basic concepts of these regulators are presented in this section. Furthermore, in [4], the steady-state effects of different excitation control modes on nodal voltages of a transmission system is analyzed. Here, dynamical simulation analysis is applied to distribution systems in order to investigate the impacts of different types of excitation control modes on the allowable penetration level of distributed generation. The general structure of an excitation system is presented in Fig. 1, which consists of measurement and signal processing circuits, a regulator and an exciter. A determined error signal, obwith the measured tained comparing the reference value signal is provided for the regulator. Then, the exciter eld voltage is adjusted based on the regulator output. The regulator is normally equipped with over and underexcitation limiters, which limit the maximum reactive power injected or consumed by the generator [7], [8]. is Voltage Regulator: in this case, the measured signal given by (1), where is the terminal voltage phasor, is the terminal current phasor, is the operator and is the compensation reactance. Reactive droop compensation (positive ) is adopted to share the reactive current among generators connected to the same bus. Line drop compensation (negative ) is used to regulate the voltage at a remote bus, usually the transformer high-voltage bus. In this work, the reactive . In this case, current compensation was neglected the terminal voltage is directly compared with the voltage reference. The usage of reactive current compensation should not be confused with power factor regulators [4] (1) Power Factor Regulator: in this case, the measured signal is the power factor. The eld voltage is adjusted to maintain constant this power factor. Such type of regulator is often used for excitation control of synchronous motors [4]. In the case of

distributed generators, this control strategy is usually adopted by independent producers to avoid penalties due to excessive reactive power consumption [1]. In consequence, unitary or capacitive power factor operation is adopted. Otherwise, such strategy of control may also be required by regulatory agencies to minimize the steady-state voltage rise in the presence of distributed generators [2]. In this situation, inductive power factor operation is required. Thus, the generator can operate in unitary, inductive or capacitive power factor, depending on the regulatory operating rules. Therefore, in this study, the following different values of power factor reference were simulated: unitary power factor, 0.95 inductive power factor and 0.95 capacitive power factor. III. STEADY-STATE VOLTAGE One of the factors that can limit the number of distributed generators connected to distribution systems is the steady-state voltage rise [2], [9], mainly during low demand intervals. Before installing (or allowing the installation of) a distributed generator, the Distribution Network Operator requires that the worst operating scenarios be analyzed to ensure that the network voltages will not be adversely affected due to the presence of the generators. Typically, these scenarios are [9]: no generation and maximum demand; maximum generation and maximum demand; maximum generation and minimum demand. In this work, it was considered that the minimum demand correspond to 20% of the maximum demand. Moreover, the allowable steady-state voltage variation was adopted equal to (0.95/1.05 pu). The single-line diagram of the system used in this section is shown in Fig. 2. Such network consists of a 132-kV, 60-Hz, subtransmission system with short-circuit level of 1000 MVA, represented by a Thvenin equivalent (Sub), which feeds a 33-kV transformer. distribution system through one 132/33 kV, The feeder X/R ratio is 4.3. The substation transformer tap was adjusted to maintain the nodal voltage in all buses within the

476

IEEE TRANSACTIONS ON ENERGY CONVERSION, VOL. 20, NO. 2, JUNE 2005

TABLE I ALLOWABLE PENETRATION LEVEL OF SYNCHRONOUS GENERATORS FOR DIFFERENT EXCITATION SYSTEM CONTROL MODES CONSIDERING STEADY-STATE VOLTAGE ISSUES

TABLE II VOLTAGE INDEX (V )

TABLE III VOLTAGE INDEX (V )

Fig. 3. Voltage prole considering different control modes of the excitation system.

allowable range for minimum and maximum demand consid. Assume that ering the case without generator an independent producer wishes to install six 5-MW, 0.90 power factor, synchronous generators at bus 8 through dedicated transformers. For this situation, simulation studies were carried out to verify what control mode of the excitation system allows such installation. The nodal voltage prole for the maximum and minimum demand cases is shown Fig. 3. In this gure, the allowable values of nodal voltage (0.95/1.05 pu) are represented by horizontal dotted lines. In this case, it was simulated that the six generators were injecting nominal active power (5 MW) into the network. It can be seen that some nodal voltages will violate the superior limit during minimum loading if capacitive or unitary power factor control mode is adopted. Whereas, if inductive power factor control mode is implemented, then some nodal voltages will be below the inferior limit during maximum loading. On the other hand, if constant voltage control mode is employed, then the nodal voltages will remain within the allowable range for both demand cases. If voltage control mode is chosen, the machine power factor will vary from 0.9731 inductive to 0.9981 capacitive for the minimum and maximum demand values, respectively.

To determine the maximum number of synchronous generators that can be installed without voltage violations, the nodal voltages were calculated for each generator added in an one-by-one basis (from 1 to 6 generators). The results are presented in Table I. It can be seen that only the voltage control approach allows all the six generators to be installed without steady-state voltage violations. In the other cases, there will be voltage violations. The third column in Table I shows what will be the problem if a new generator is installed. The most restrictive cases are related to the constant inductive and capacitive power factor control modes. A. Steady-State Voltage Variation Due to Generator Tripping Another issue on steady-state voltage variation is to determine the impact on the voltage prole when synchronous generators are suddenly tripped off. It is desirable that this variation be as small as possible. To examine this issue in quantitative terms, the following global voltage index can be used:

(2)

is the total number of buses, is the nodal voltage where is the of bus in the presence of distributed generators, and nodal voltage of bus without distributed generators. The results obtained, considering that the six generators are tripped off during maximum and minimum demand, are presented in Table II. It can be observed that the unitary power factor case

FREITAS et al.: INFLUENCE OF EXCITATION SYSTEM CONTROL MODES ON DISTRIBUTED SYNCHRONOUS GENERATORS

477

Fig. 4. Single-line diagram of the system 2.

leads to the smallest voltage variation for maximum and minimum demand. On the other hand, the capacitive power factor case implies in the largest voltage variation for both the loading situations. B. Steady-State Voltage Regulation It is desirable that the voltage regulation, i.e., the difference in the voltage prole between the maximum and minimum demand cases, be as small as possible. The following global voltage index can be employed to quantify this issue: (3) is the nodal voltage of bus for maximum demand where is the nodal voltage of bus for minimum demand case and case. The results are shown in Table III. It can be veried that, in the presence of the distributed generators, the inductive power factor mode implies in the worst voltage regulation (maximal variation). On the other hand, the voltage control mode leads to the best voltage regulation (minimal variation). IV. ANGLE TRANSIENT STABILITY Fault clearance time in distribution systems may be very long. Additionally, the inertia constant of distributed generators is typically low. In consequence, angle stability issues can limit the amount of active power injected by synchronous generators. In this section, the angle transient stability of distributed generators is analyzed using the system presented in Fig. 4. Such network comprises a 133-kV, 60-Hz, subtransmission system with shortcircuit level of 1500 MVA, represented by a Thvenin equivalent (Sub), which feeds a 33-kV distribution system through two transformers. A synchronous generator with 132/33 kV, capacity of 30 MVA is connected at bus 6, which is connected transformer. This to the network through a 33/0.69 kV, machine can represent an equivalent of various generators in a thermal or small-hydro generation plant. In order to verify the transient stability of this system, a three, phase-to-ground short-circuit was applied to bus 4 at which was eliminated at 250 ms by tripping branch 24, when the synchronous generator was injecting 25 MW into the network. The rotor angle responses for different excitation control modes are presented in Fig. 5, where the results were split into two graphs for better visualization. If the unitary or inductive power factor control modes are chosen, the system is unstable. This happens because the machine rotor eld voltage decreases during the short-circuit interval, as shown in Fig. 6. This behavior of the eld voltage affects the system dynamic adversely, contributing to accelerate the rotor angle, since the objective of

Fig. 5. Dynamic behavior of the rotor angle.

Fig. 6. Dynamic behavior of the eld voltage.

the excitation system is to control the power factor not the terminal voltage. In the case of capacitive power factor control mode, although the eld voltage also decreases during the fault application, the system is stable. This occurs because, in this case, the pre-fault rotor angle is much smaller than it is in the other cases. Thus, even though the rotor acceleration increases due to the behavior of the excitation system, the rotor angle does not reach the critical value during the transient time interval. The values of the pre-fault rotor angle for all cases are shown in Table IV for comparison. A brief explanation on why these pre-fault rotor angles are different from each case can be seen in Appendix. The system is stable under the voltage control mode, as shown in Fig. 5. In this case, the rotor eld voltage rises fast, reaching

478

IEEE TRANSACTIONS ON ENERGY CONVERSION, VOL. 20, NO. 2, JUNE 2005

TABLE IV PRE-FAULT ROTOR ANGLES

Fig. 7. Dynamic behavior of the terminal voltage. TABLE V MAXIMUM ALLOWABLE PENETRATION LEVEL OF EXPORTED ACTIVE POWER DUE TO TRANSIENT STABILITY ISSUES

the maximum value during the contingency, since the objective of the excitation system is to control the terminal voltage. The dynamic behavior of the generator terminal voltage for the stable cases is shown in Fig. 7. A. Critical Power Another issue that deserves investigation is the inuence of the excitation control mode on the maximum active power (critical power) that a distributed generator can inject into the system without provoking transient stability problems. Several simulation studies were carried out considering different fault clearance times and excitation control modes. The results are summarized in Table V. It can be seen that the voltage and capacitive power factor control modes lead to the highest values of critical power. Otherwise, inductive power factor control mode can limit signicantly the generator output. Observe that if the fault clearance time is smaller than 15 cycles, both the voltage mode and the capacitive power factor mode have similar impacts on the critical power. However, if the clearance time is larger (18 cycles, for example), then the voltage control mode presents a superior performance. This can be explained by the facts observed in the simulations showing that if the eld voltage remains low for a long time, when capacitive power factor control is adopted, then the system becomes unstable even if the pre-fault angle is small. V. SHORT-CIRCUIT CURRENTS It is well recognized that extensive usage of distributed generation may increase the values of short-circuit currents. As a conFig. 8. Phase-A stator current of the synchronous generator.

sequence, this issue can limit the maximum allowable number of synchronous generators connected to a determined network. In this section, the impact of different excitation control modes on the short-circuit currents supplied by synchronous generators is investigated. The system analyzed is the same shown in Fig. 4. All simulations were obtained using electromagnetic transient analysis. Different short-circuits were applied during 300 ms, to , and cleared without line tripfrom ping. The dynamic behavior of the phase-A stator current supplied by the synchronous generator for a three-phase-to-ground short-circuit applied at the middle of branch 45 is shown in Fig. 8. It can be observed that the current responses are different depending on the adopted control mode. In Table VI, the numerical value of the phase-A stator current of the synchronous generator for four different faults is presented. The types of faults simulated were a three-phase-to-ground fault or a phase-A-to-ground fault applied at bus 5 and in the middle of branch 45. The peak value of this current is shown in the rst column. In the remaining columns, the root-mean-square (rms) values of the current are presented at different instants after the fault application. It can be noted that the highest values of peak current occur when capacitive power factor control is adopted.

FREITAS et al.: INFLUENCE OF EXCITATION SYSTEM CONTROL MODES ON DISTRIBUTED SYNCHRONOUS GENERATORS

479

TABLE VI PHASE-A SHORT-CIRCUIT CURRENT PROVIDED SYNCHRONOUS GENERATOR (P.U.)

BY

Fig. 9. PV curve of bus 4.

TABLE VII PRE-FAULT TERMINAL VOLTAGE (P.U.)

On the other hand, the lowest values are obtained when inductive power factor control mode is employed. This fact can be explained considering the pre-fault values of the terminal voltage shown in Table VII. The higher the value of the pre-fault terminal voltage, the higher the maximum value of the stator current. The magnitude of the pre-fault terminal voltage is fundamental to determine the maximum value of the short-circuit current, however, the dynamic behavior of the eld voltage determines the sustained response. Such fact can be veried by analyzing the rms value of the short-circuit current along the time. It can be seen that after 12 cycles of the fault application, the voltage control mode leads to highest values of short-circuit current. In the inductive power factor case, the generator is practically unable to supply sustained short-circuit currents. This is an important fact because, in this case, the generator protection system may not be able to detect the fault in the network. VI. VOLTAGE STABILITY Normally, the installation of synchronous generators can enhance the voltage stability of distribution networks. How-

ever, the amount of voltage stability margin improvement will depend on the excitation control mode chosen. In order to investigate this issue, the PV curves of the system shown in Fig. 4 were determined. Usually, distributed generators are not rescheduled by the system operator. Thus, the PV curves were obtained varying the active and reactive loads and keeping constant the active power injected by the synchronous generator at the nominal level (30 MW). The PV curve of bus 4 is presented in Fig. 9. It can be seen that the presence of the generator increases the system voltage stability margin, independently of the control mode chosen. However, the largest gain of the voltage stability margin is obtained when the excitation system acts as a voltage regulator. Whereas the smallest gain of the margin occurs when the excitation control is operating in inductive power factor mode. In the case of voltage control mode, the generator can operate using its reactive power limit dynamically, improving the performance of the system voltage stability. On the other hand, if the power factor control mode is adopted, then the generator cannot dynamically supply proper reactive power support, limiting the benecial effect of the generator installation on the system voltage stability.

VII. CONCLUSIONS Among the various factors that can determine the penetration level of synchronous generators in distribution networks, the excitation system control mode is one of signicant importance. Therefore, in this work, the impacts of excitation systems acting either as a voltage regulator or as a power factor regulator on the allowable penetration level of distributed synchronous generators were investigated with respect to steady-state voltage, angle stability, short-circuit current and voltage stability issues. Based on simulation results, obtained using dynamical models, it can be concluded that the voltage control mode for the excitation system is the most advantageous strategy to increase the allowable penetration level of distributed synchronous generation in typical distribution networks.

480

IEEE TRANSACTIONS ON ENERGY CONVERSION, VOL. 20, NO. 2, JUNE 2005

: Generator operating in inductive power factor increases, consequently, inthe numerical value of creases. Therefore, it is necessary to increase the generator internal angle if it is desirable that synchronous generators consume reactive power from the network. REFERENCES
[1] N. Jenkins, R. Allan, P. Crossley, D. Kirschen, and G. Strbac, Embedded Generation, 1st ed. London, U.K.: IEE, 2000. [2] (1999) Dispersed Generation. CIRED Working Group 4. [Online]. Available: http//www.cired.be [3] Impact of Increasing Contribution of Dispersed Generation on the Power System, CIGR Working Group 37.23, 1999. [4] J. D. Hurley, L. N. Bize, and C. R. Mummert, The adverse effects of excitation system var and power factor controller, IEEE Trans. Energy Convers., vol. 14, no. 4, pp. 16361641, Dec. 1999. [5] IEEE Task Force on Load Representation for Dynamic Performance, Load representation for dynamic performance analysis, IEEE Trans. Power Syst., vol. 8, no. 2, pp. 472482, May 1993. [6] P. Kundur, Power System Stability and Control. New York: McGrawHill, 1994. [7] Recommended models for overexcitation limiting devices, IEEE Trans. Energy Convers., vol. 10, no. 4, pp. 706712, Dec. 1995. [8] Underexcitation limiter models for power stability studies, IEEE Trans. Energy Convers., vol. 10, no. 3, pp. 525531, Sep. 1995. [9] C. L. Masters, Voltage rise: The big issue when connecting embedded generation to long 11 kv overhead lines, Power Eng. J., vol. 16, no. 1, pp. 512, 2002.

Fig. 10.

Single-machine system.

APPENDIX PRE-FAULT ROTOR ANGLE CALCULATION The values of pre-fault angle are different when using distinct excitation control modes due to the reactive power injected by the generator for each case. It is important to remember it is not possible to neglect the - coupling in distribution systems due to high X/R ratio. This fact can be veried by analyzing the single-machine system presented in Fig. 10. Such system consist of a synchronous generator, which is represented by a voltage source behind an impedance, connected to a system bus through a line. It is possible to show that, neglecting the rotor saliency, the steady-state internal angle of the generator is given by [6, pp. 101] (4)

Walmir Freitas (M02) received the B.Sc. and M.Sc. degrees from Sao Paulo State University (UNESP), Sao Paulo, Brazil, in 1994 and 1996, respectively, and the Ph.D. degree from the State University of Campinas (UNICAMP), Campinas, Brazil, in 2001, all in electrical engineering. He is currently a Post-Doctoral Fellow at the University of Alberta, Edmonton, AB, Canada. His research interests are in power system stability and control, distributed generation, and power electronic applications.

where , , is the generator stator resistance, is the feeder series resistance, is the is the feeder series reacgenerator synchronous reactance, is the magnitude of the terminal current injected by tance, is the magnitude of the nodal voltage of the the generator, system bus, and is the angle between the phasors and . Multiplying the numerator and denominator of in (4) by , the generator internal angle can be computed by (5)

Jose C. M. Vieira (S97) graduated in electric engineering from the University of Uberlandia (UFU), Uberlandia, Brazil, in 1996, and received the M.Sc. degree in 1999 from the State University of Campinas (UNICAMP), Campinas, Brazil, where he is currently pursuing the Ph.D. degree. From 1999 to 2003, he was with FIGENER as a Consulting Engineer. His research interests are distributed generation, optimal power ow, energy markets, and power system control and dynamics.

where is the active power at the receiving-end bus and is the reactive power at the receiving-end bus. Theoretically, the generator internal angle varies from 0 to 90 degrees. Consequently, the internal angle increases and vice increases when the numerical value of versa. Thus, the following analysis can be carried out. Generator operating in capacitive power factor : the numerical value of decreases, consequently, decreases.

Andre Morelato (M89) graduated in electronics engineering from the Instituto Tecnologico de Aeronautica (ITA) in 1970 and received the Ph.D. degree in 1982 from UNICAMP, Campinas, Brazil. He is currently a Full Professor of Electrical Engineering at UNICAMP. During late 19911992, he was with Hitachi Research Laboratory, Hitachi, Ltd., Japan. His general areas of research interest are power system control and stability, distributed generation, and parallel processing applications.

Wilsun Xu (M90SM93) received the Ph.D. degree from the University of British Columbia, Vancouver, BC, Canada, in 1989. From 1989 to 1996, he was with B.C. Hydro as an Electrical Engineer. He joined the University of Alberta, Edmonton, AB, Canada, as an Associated Professor in September 1996, and is presently a Full Professor of Electrical Engineering. His research interests are power quality, power system stability, distributed generation, and distribution automation.

Anda mungkin juga menyukai