You are on page 1of 8

144

IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 19, NO. 1, FEBRUARY 2004

Power System Stabilizers as Undergraduate
Control Design Projects
Joe H. Chow, George E. Boukarim, and Alexander Murdoch

Abstract—The use of power system stabilizers (PSS) to damp
power system swing mode oscillations is of practical importance.
The design of PSS is taught in graduate level courses on power
system dynamics and control, and has been the topic of numerous
M.S. and Ph.D. theses. This paper discusses the experience in assigning PSS projects in an undergraduate control design course to
provide students with a challenging design problem using three different techniques and to expose them to power system engineering.
The details of the PSS design projects using root-locus, frequencydomain, and state-space methods are provided.

In both the VR and PSS designs, realistic control specifications
such as the rise time and the overshoot of step responses were
given. The PSS structure and the lead-lag compensator formulas
were provided to allow students to concentrate on selecting the
PSS parameters. The final designs were verified by time simulation using Simulink. These design projects are described in
Sections III–V. Some observations on the projects are provided
in Section VI.

Index Terms—Frequency-domain compensation, power system
stabilizers, root-locus techniques, state-space methods, undergraduate design projects.

II. POWER SYSTEM MODEL

I. INTRODUCTION

T

HE undergraduate level control systems engineering
(CSE) course at Rensselaer Polytechnic Institute and
many other universities covers the root-locus technique, frequency-response compensation, and state-space methods for
control design [1]. The prerequisite of the course is the signals
and systems course [2]. A course in modeling of dynamic
systems [3] is helpful but not required. In the Rensselear CSE
course, the students are required to do a sequence of design
projects, each corresponding to a different design technique.
Past projects included a ball-and-beam system and an inverted
pendulum system [4], which are challenging but tend to be
more of textbook- and laboratory-type problems. To introduce
the students to real-world design problems, in the Fall 2002
semester, three power system stabilizer (PSS) design problems
were assigned to about 40 students. As part of the projects,
the students were also required to design the voltage regulator (VR). The MATLAB package, with the Control System
Toolbox and Simulink, was used for the design [5].
In assigning these problems, we needed to distill the PSS design methodologies into steps that the students could accomplish with basic design knowledge. Most of the CSE students
had not taken an introductory power system analysis course,
but did have notions of dynamic systems. In each project, the
same single-machine, infinite-bus system model in state-space
form, presented in Section II, was used. The VR was obtained
first using basic design guidelines offered in textbooks [1], [6],
[7]. Then with the VR loop closed, more specialized techniques
were used to design the PSS to add damping to the swing mode.
Manuscript received June 24, 2003.
J. H. Chow is with the Electrical, Computer, and Systems Engineering Department, Rensselaer Polytechnic Institute, Troy, NY 12180-3590 USA.
G. E. Boukarim and A. Murdoch are with Power System Energy Consulting
Department, General Electric Company, Schenectady, NY 12345 USA.
Digital Object Identifier 10.1109/TPWRS.2003.821003

A single-machine infinite-bus system (Fig. 1) was used as the
design model. The machine, modeled with subtransient effects,
to the infinite bus. The voltage
delivers the electrical power
regulator controls the input to a solid-state rectifier excitation
system [8], which provides the field voltage to maintain the genat a desired value
. The states
erator terminal voltage
for the machine are its rotor angle , its speed , and its direct, ,
, and . The exciter is
and quadrature-axis fluxes
modeled with the voltage state . All of the variables are normalized on a per-unit (p.u.) basis, except for which is in radians.
The power system model is linearized at a particular equilibrium point to obtain the linearized system model given in the
state-space form
(1)
where denotes the perturbation of the states, input, and outputs from their equilibrium values, with
(2)
(3)
The matrices for (1) derived from typical machine parameters
are given in Appendix A. In the sequel, the symbol will be
dropped to simplify notations. The dominant poles of (1) are the
associated with the field voltage response,
real pole
and the electromechanical (swing) mode
with a small damping ratio
, representing the oscillation of machine against the infinite bus.
Starting from (1), the students were required to first use
to design a high-gain
the terminal bus voltage signal
. Because the VR destabilized the swing mode, a
VR
using the machine speed signal was used to add
PSS
damping to the swing mode. The feedback control system block
diagram implemented in Simulink is shown in Fig. 2. Note that
the gain is set to 1 for the root-locus and frequency-response
designs. The input signal to a speed-input PSS is derived from

0885-8950/04$20.00 © 2004 IEEE

step in of the open-loop system (1) up to 10 s. 4.u. yielding increases.1-p. Control structure of the single-machine infinite-bus system. Perform another root-locus analysis for (6) and select to achieve a for the swing mode. Single-machine infinite-bus system. Then with the PSS-loop open. how. but the oscillation due to the swing mode becomes less damped. Thus.1-p. Find the angle of departure of the root-locus branch leaving the swing mode with the positive imaginary part. These specifications reflect the requirements of modern high-gain VRs. the proportional gain has to be kept below 47. 1. late the closed-loop step input.CHOW et al.u. and A good way to approach task R3 is to set up a grid of values and simulate the step response by sweeping through (5) and plot the closed-loop response to a 0.u. the closed-loop step a 25% steady-state error. Discussion of Design The first project in the sequence was the design of the VR and the PSS using root-locus techniques. III. balanced by a sign inversion in the feedback path. Note the slow open-loop response which settles to 0. as increases. 3 shows the open-loop step response and the responses for the different values. the voltage mode moves left shown in Fig. ever. the PSS path is not active. such that in steady state. The swing mode. because the MATLAB root-locus function assumes negative feedback. these ideas were translated into procedures that could be followed by students with basic control system design skills.and closed-loop transfer functions required in various design stages are generated from the Simulink diagram by opening appropriate connections. such that the angle of departure of the compensated system is about 180 . 2. simulate the response (R1) For a 0. such a high gain is nonrobust because any reduction in the system gain due to changing operating condition would pull the swing mode back into the right half-plane.u. assuming gain control. At . (R3) Apply a PI controller for the VR Tasks R1 and R2 reveal important properties of the system and the proportional control. and crosses the imaginary axis at is increased Although the system becomes stable again as beyond 1260. repeat the simulation for (1) conwith trolled by a proportional VR . In the PSS projects. the machine speed passed through a washout filter and several banks of torsional filters [9]. A. ROOT-LOCUS DESIGN and such that the rise time is less than 0. thus improving .1-p. Fig. Detailed discussions of PSS design techniques based on the synchronizing and damping torque concept can be found in many excellent references such as [13]–[15]. the output of the transfer function from to be a proportional torsional filter. design a second-order phase-lead com(R5) Based on pensator (6) (7) using the phase-lead properties described in Appendix B. output was not used in the projects. the feedback system is unstable with a growing oscillation. Select the parameters from . which are usually taught first in a control systems course. It can be used if The an instructor wants to extend the designs to a dual-input PSS with both the machine speed and the output power as the input signals [10]–[12].. Check if s and have been satisfied. damping ratio in the Simulink diagram and simu(R6) Implement response to a 0. Fig. step input. The aggregate phase lag effect of the torsional filters is represented by (4) Note that the conventional PSS path comes into the summing junction with a positive sign.: POWER SYSTEM STABILIZERS AS UNDERGRADUATE CONTROL DESIGN PROJECTS u Exciter Synchronous V term machine Pe VR Vterm N Transmission line ω _ Voltage regulator + 145 Vref Normalization Gain Step Infinite bus t Clock Vref time numVR(s) denVR(s) Voltage Regulator KV(s) numPSS(s) denPSS(s) PSS Kd(s) x’ = Ax+Bu y = Cx+Du w Power System State–Space Model wf numTor(s) denTor(s) torsional filter 10s 10s+1 Washout Filter Pe -1 Sign Inversion Fig. is destabilized. (R2) Make a root-locus plot of the voltage regulation loop using the proportional controller and find the gain when the lightly damped swing mode becomes unstable. The washout (derivative) filter is a high-pass filter having a dc gain of 0. and per(R4) Close the voltage regulation loop with form a root-locus analysis of the PSS loop using the to . As from . Here. As response becomes faster and the steady-state error smaller. The instability can be studied with a root-locus plot. The open. Design Tasks B.5 s and the overshoot is about 10%.0747 p. . we use a negative sign. The process was specified in several tasks: .

3 Pole: -0. 6. and closed loop for K p = 10. One set of values that will and .5 -1 -0. . The root-locus analysis is repeated with the phase compensation in the feedback loop. although older analog VRs usually have a high proportional gain for tighter steady-state regulation and a lag-lead block for transient gain reduction.04 0. . we close the VR loop and add in the washout filter to generate the transfer function from to the and PSS input. 7 shows the swing mode having a departure angle close to 180 .13 Pole: -1.13 to achieve a 15% damping ratio.0747 pu Kp = 20 0. We leave some margin in because a PSS will slow down the time response. the response to a 0. responses to step in V V 14 with K (s) (5). . VOL. with a PSS providing additional damping.5 0 -25 0.5 Fig.1 pu step response amplitude 0. s and . From Fig. 5. Taking the PSS to be a proportional controller. responses to step in V V 6 8 10 Time (sec) .7 Frequency (rad/sec): 8. we . To add purely damping to the swing mode. has step in s and . 5. 5.00206 + 9. -20 System: untitled1 Gain: 1. open loop. 30 25 20 Imag Axis the grid values. In task R4.06 0. Thus. -10 -5 0 5 10 Root-locus plot of swing mode using proportional gain PSS. On the other hand. 3.1 with PSS without PSS 0.1-p. this 120 compensation requires two first-order phase-lead compensators in series.06 to 0. the overshoot will be reduced.02 0 0 10 2 4 Time (sec) Fig. the departure angle should be 180 . 1. allowing -15 Real Axis Real Axis Fig.9 Frequency (rad/sec): 9.12 Kp = 50 0. Note that new commercial digital VRs have integral actions.12 0.146 IEEE TRANSACTIONS ON POWER SYSTEMS.36 + 8. satisfying the design specifications.08 Kp = 30 0.1 0. Setting the two compensators to be identical.1 pu step response amplitude 0. 50.u. Root-locus plot of voltage regulation loop. satisfy the design specifications is whose step response is shown as the dashed curve in Fig. shown as the solid curve in Fig. 20 12 15 Imag Axis Imag Axis 10 System: GV Gain: 47.83i Damping: 0.93 15 10 5 0 -5 -18 -16 -14 -12 -10 -8 -6 -4 -2 0 Real Axis Fig. 17. . 20. 7. NO. Fig. Root-locus plot of swing mode with PSS (8). for task R5. 6). 4.152 Overshoot (%): 61. 19. the departure angle of the root-locus branch leaving the swing mode is about 60 (Fig. we need to add a 120 compensation at the swing-mode frequency in the feedback loop using a phase-lead controller.33 rad/s obtain as the center frequency . . the selection of the PSS gain at 1.02 0 Open Loop 0 2 4 6 8 0.08 0.4i Damping: 0.4 8 6 4 60 deg 10 5 2 0 0 -5 -2 -3 -2. .000219 Overshoot (%): 99. The PSS thus has the form (8) When the PSS loop is closed. with the swing mode frequency of 9.04 Kp = 10 0. FEBRUARY 2004 0.5 -2 -1. Fig. The integral action has with removed the steady-state error.

p.CHOW et al. For the designed with plot the frequency response of . indicating a low-frequency gain of 200 and a phase margin of 85.747 and a local magnitude peak rad/s due to the lightly damped swing mode. Plot the frequency response of filters to form . which will be improved by the PSS. Regarding the power system model as a second-order mass-spring path system. The resulting system ture 147 (6) such that (F5) Design a phase-lead controller is between 0 to from the phase of 2 to 12 rad/s (approximately 0. . with the dc gain at 0. and are column vectors. improved damping is possible whenever the swing frequency. and the pole to be (12) (13) (10) where . and generate the linear model from with the loop closed. Again.1-p. at This swing mode also determines the gain margin to be 33 dB of at 40.3 to 2 Hz). step input.1-p. Due to the local magnitude peak of the swing mode. Decomposition of speed control loop. Furthermore.67).6 rad/s. late the .2H _ tion machine torque inertia __ 1 s machine speed ω ___ 377 s machine angle δ Q(s) Effect of speed on electical torque due to machine dynamics Fig. We set the high-frequency gain of the compensated that the new gain-crossover frequency system is about 3. such that (14) The frequency-response plot of the compensated system is also shown in Fig. Find . the root-locus departure angle from the swing mode will be slightly less than 180 . Find . whereas the PSS design required decomposing the accelerating torque on the synchronous machine into its synchronizing and damping components. Design Tasks The second project required the use of frequency-response design techniques. as shown in matrix has the strucFig. Note that the phase of is at 2 rad/s . the zero. Find the dc gain and the gain and phase margins of the uncompensated system. in the Simulink diagram. (F1) Plot the frequency response of the system from to . Use this system to perform a filter. is a row vector. To guide the design process. the washout (F6) Connect . so (44. the following tasks were specified. such that the swing mode frequency will increase. 10. The closed-loop system response to is shown as the dashed curve in Fig. 9. step in s and . using a frequency range of 1 to 100 rad/s. Plot the compensated system frequency response and find the gain and phase margins. 8. disconnect into the input summing the feedback path from to the junction.u. decompose the model to isolate the from the speed to the electrical torque. 11. The rais close to zero tional is that if the phase of and slightly negative. FREQUENCY-RESPONSE DESIGN A. and the steady-state error. in the Simulink diagram and simu(F7) Implement response due to a 0. Appendix B will be helpful. Find the output of gain and phase margins of the compensated system. and to achieve a damping root-locus analysis to select ratio for the swing mode. . in series with .u. falls in this range. we construct in Fig. the gain margin is only 0. we set the controller low-frequency gain. 8. (F2) Design a phase-lag controller (9) such that the dc gain of the compensated system is at least 200 and the phase margin is at least 80 . and is a square with the matrix. B. 405]. Discussion of Design The uncompensated system frequency response plot in Fig. the speed feedback loop will add mostly damping.85 dB.5 . (F8) As a final check. and the steady-state error. (F3) Close the voltage regulation loop and simulate the response of the closed-loop system to a 0. Construct the state-space model machine speed as the input and the electrical torque as the output (11) in series with the washout and torsional Connect . the phase-lead compensator characteristics in .u. By specifying a frequency range. IV. The VR design was quite standard. a 0. which can change according to the operating condition. step input. improving the synchonizing torque. with a phase of about . with and plot its frequency response In task F4. Using the phase-lag design procedure in [1.1-p.: POWER SYSTEM STABILIZERS AS UNDERGRADUATE CONTROL DESIGN PROJECTS synchronizing torque damping torque mechanical torque + K D _ _ ___ 1 accelera. from to (F4) Generate the state-space model with the lag controller loop closed. 9 shows that the magnitude of the open-loop system is less than unity.

step in is shown as the solid curve in Fig. To set the gain.u. VOL. which are similar this response. step input achieves s. with much of its dynamics not important. we need to add at least 113 at 12 rad/s.04 0.u. and drops to at 12 rad/s. use as the input and as the output. V responses to step in V with lag controller (14).1 25 without PSS with PSS 0. and zero steady-state voltage regulation error. the system response to a 0. V. obtained from task S1 is sev(S2) The controller enth order. In this task. and rad/s are selected for both the values lead stages.15 Overshoot (%): 62 Frequency (rad/sec): 9. 12. the gain margin in the voltage loop improves to 10 dB. 12. According to the phase specifications. as shown in Fig. Fig. we perform a root-locus analysis. function. State-space design is not commonly used for PSS design because the controller order may be high. The design tasks were as follows: (S1) With the PSS-loop open. leaving all of the other poles unchanged.02 20 System: untitled1 Gain: 3. to those obtained from the root-locus design. plot the frequency response of in the zero-pole-gain form and Then. Thus. Note that in s and . the PSS has the form (15) When the PSS loop is closed.1 pu step response amplitude 0.IEEE TRANSACTIONS ON POWER SYSTEMS.33 rad/s used in (8). . STATE-SPACE DESIGN A.25i Damping: 0. The center frequency of the desired lead controller will be higher than the value of 9. the VR and PSS [16]. Fig.12 0.35 15 10 5 0 0 0 2 4 6 8 10 -20 Time (sec) -15 -10 -5 0 Real Axis Fig. obtaining a gain of 3.1-p.1-p. design an observer so that the system states can be estimated by ensuring that the voltage control tranfrom sient due to the observer pole decays faster than the full-state feedback pole.08 Imag Axis 0. 30 0. First.41 + 9. Thus. full-state feedback laws and observers derived from pole placement were used to design and implement Root-locus plot of swing mode for (15). with the PSS adding damping and reducing the local magnitude peak due to the swing mode. NO. Phase Compensated F(s) 102 Frequency (rad/sec) Frequency responses of voltage regulation loop. it was essential that some model reduction steps be taken to obtain low-order controllers. Implement the observer-based and the scalar gain such that for a controller . we will preserve only its voltage regulation .75 to achieve a 15% damping ratio of the swing mode.75 Pole: -1. FEBRUARY 2004 Magnitude (dB) 50 Compensated 0 Uncompensated -50 -100 -150 0 -45 -90 -135 -180 -225 -270 10-2 Uncompensated Phase (deg) Phase (deg) Magnitude (dB) 148 Compensated 10-1 100 101 102 0 -20 -40 -60 -80 0 Phase Compensated -90 -180 F(s) -270 -360 100 103 101 Frequency (rad/sec) Fig. 10. In task F8. Then. After a few design iterations.06 0. Frequency responses of F (s) and F (s)K (s). in this project. express . 10. Design a full-state feedback control to place the closed-loop voltage regulation pole at the desired location to reduce the voltage control time constant. the closed-loop response 0. single-output model. 9. 1. to obtain a single-input. 19. 11. Design Tasks In the third project.

fast rise time VI. One lecture was spent on a detailed discussion of each design project. 13. PEDAGOGY In a typical undergraduate control systems course. the time response of the closed-loop system. from to (S3) Generate the state-space model with (16) implemented. the shown as the solid curve in Fig. –35. AND GAIN OF (17) whose frequency response no longer has a notch filter characteristic. response using . 14 processes only a small amount of oscillations. and for the observer design.1-p. whose frequency response is shown in Fig. 14. 0. For zero .: POWER SYSTEM STABILIZERS AS UNDERGRADUATE CONTROL DESIGN PROJECTS 149 perform approximate pole-zero cancellations to retain only a first-order reduced controller TABLE I ZEROS.243.6.1 pu step response amplitude where is selected so that and have the same dc gain.7. server-based controller is shown in Fig. with s and . The frequency response of exhibits a notch filter characteristic at the swing mode.33 –114. In task S1. 14. 16) root-locus plot using moves from the swing mode almost parallel to the negative real axis. An interesting aspect of the state-space design is that in the . –0.19. POLES.u. This pole will be retained in except for the pole at (16) such that 35 10 0 0. we set response to a 0. The closed-loop steady-state voltage error. Note that this response has a s.4.–26. Find and . 15 shows that the frequency responses of and match closely. Discussion of Design .04 0. The controller is designed to improve the damping ratio of the swing mode to 15% by first obtaining a full-state feedback control to add damping to the swing mode and then an observer so that the states of the system . The resulting seventh-order Table I.12 KV with PSS B. it is Because the open-loop voltage pole is at essential that this pole be made faster in order to improve the voltage response. 14.479 ± j9. –114. 13.u. V responses to step in V for K (s) and K (s). Frequency responses of K (s) and K (s). the swing mode to achieve a slightly higher than 15% is shifted to damping ratio. these design projects were assigned concurrently with their lecture discussions. Plot the frequency response versus that of .06 0.001. but a higher overshoot . this pole will be shifted to for the full-state feedback design and to in the is given in observer design. is oscillatory. each of the three design techniques are covered in about three weeks. step in is loop is closed.59 392. At Rensselaer.u. the root-locus branch (Fig. Thus.6. All of the other shifted to poles are fixed.1-p.1 KV without PSS 0.1-p.–34. –3. providing purely damping enhancement. Students were allowed . When the PSS response to a 0. Construct the 11th-order can be estimated using observer-based controller . This design strategy is not robust. Fig.8. 15. because the swing frequency can vary with the operating condition.08. 13. Compare it to the step a 0. a fifth-order controller obtained. the swing mode is to ensure fast tracking. Table I shows that the zeros and poles form approximate pairs .6 zeros poles gain (16) K (s) 0. In task S3 for the full-state feedback design.u. Close the voltage of and perform a step response with loop with step input.1-p.–13. Using the funcis tion with a tolerance of 0. The frequency response of the 11th-order ob- Magnitude (dB) 1st order 30 7th order 25 20 15 Phase (deg) 1st order 7th order -45 -90 10-1 100 101 102 Frequency (rad/sec) Fig. reducing its effect on the swing mode. input. Note the phase-lead characteristic of .CHOW et al. step in shown as the dot-dashed plot in Fig. –26. which also has some dynamics that are not important. Control System Toolbox to obtain a controller Do a root-locus analysis of the damping controller .–0. shown as the dashed plot in Fig. as shown in Fig. Close the damping control loop loop using and simulate the response of the with step overall closed-loop system to a 0.02 0 0 2 4 6 8 10 Time (sec) Fig.08 KVo 0.–2.8.609 ± j9. Apply the minin the MATLAB imal realization function .

19. Many students were amazed at how well their VRs and PSSs performed. and state-space methods. PHASE-LEAD COMPENSATION The single-stage phase-lead controller (18) has its zero closer to the origin of the -plane than its . 60 Root-locus plot of swing mode using K (s). to form two-person teams and had about two to three weeks to complete each project. NO. used in industry and PSS tuning during commissioning. considering that the design required closing two single-input. one of the co-authors (GEB) of this paper gave a lecture to the students on PSS design tools In this paper. PSS designs for systems with multiple operating conditions can be considered. 16. for an undergraduate control systems course. For a graduate-level design course. The system data can also be downloaded from the web site www. The course survey indicated that the students were generally pleased with the design activities.7 Frequency (rad/sec): 9. At low frequencies. Many students also included these design projects on their resumés. it has the At the center frequency defined as maximum phase lead of (19) . VII.172 Overshoot (%): 57.com.IEEE TRANSACTIONS ON POWER SYSTEMS. FEBRUARY 2004 Phase (deg) 25 20 15 10 5 0 -5 180 80 11th order 5th order φm (deg) Magnitude (dB) 150 0 2 4 6 8 11th order Fig.13 8 6 4 2 0 -6 -4 -2 0 2 Real Axis Fig. APPENDIX VIII. based on the root-locus. STATE-SPACE MODEL The matrices for the state-space model (1) are shown in the equation at the bottom of the top page. IX. we have presented three PSS design projects. Although the projects could be accomplished solely using the MATLAB command language. 1. A saturation block can be added to the output of the PSS to limit its contribution in the voltage regulator input. In the design projects. single-output loops. Feedback and closed-loop systems are among the hardest concepts for many engineering students to have a good grasp. depending on the needs of the course. frequency-response. applying different techniques to the same design problem provides insights into the linkage of the design concepts. More advanced techniques such as fuzzy logic [17] and genetic algorithms [18] can also be applied. Maximum phase angle  10 αld 12 14 16 18 20 versus . In addition.brookscole. reducing the possibility of errors. VOL. At high frequencies. its gain is unity and its phase pole and its phase is zero. 15. and the design experience became a favorite discussion topic in their job interviews. This was an important consideration when most of the students were not proficient MATLAB users before taking this control systems course. CONCLUSION 18 16 14 12 Imag Axis 20 0 Frequency (rad/sec) 10 System: untitled1 Gain: 1. with advance planning. The design specifications can be either relaxed or tightened.01 Pole: -1. Although such industry interactions can be readily arranged at Rensselaer because of her proximity to a turbine-generator manufacturer. is zero. Toward the end of the semester. one for regulation and the other for stabilization. 17. Any one of the PSS design projects can be adopted for use with modifications by instructors of similar courses. 40 5th order 90 10-1 Fig. Good design projects with powerful simulation tools can be very helpful in reinforcing theoretical analysis. In addition. The projects provided students with some realistic and challenging design experience and exposed them to a well-known power system design problem. we provided a Simulink diagram for the students to implement the controllers.57 + 9i Damping: 0. which provides resources for the text [4]. a PSS design for multiple-machine systems or the dual-input PSS design can be addressed. 135 100 101 102 Frequency responses of K (s) and K (s). its gain is . it is possible for other schools to do likewise. operations with Simulink diagram removed most of the drudgery of keeping track of the feedback structure.

D. His areas of interest include rotating machine modeling. which is useful for seto achieve a desired phase lead [6]. He joined General Electric in 1975 and has worked primarily at PSEC but also for Drive System. and advanced control theory. Falcao. parts I–III. J. Lawson. working in the area of power system dynamics and control. [14] E. 3017–3046.. Wright. 271–277.” IEEE Trans. 8th ed. pp.S. pp. Minneapolis. Inc. 189–202. Automatic Control Systems.” IEEE Trans. Lawson. 1994. respectively. vol. 1996. A. and the M. May 1989. M. R. Murdoch. pp. M. Newell. NY. NY. degree from Worcester Polytechnic Institute. Computer. Pearson. Close. [18] A. [2] A.. Kundur. 1978. Worcester. Chow. 1994. A. 4th ed.: POWER SYSTEM STABILIZERS AS UNDERGRADUATE CONTROL DESIGN PROJECTS Fig.. [10] Digital Excitation System Task Force of the Equipment Working Group. Syst. He also worked for the Transmission Technology Institute. [12] A. A.” IEEE Trans. 1997. [13] F. 14. G. “Minimization of power system stabilizer torsional interaction on large steam turbine-generators. J. 2001. 2003. design. Frederick and J.. Kundur. in 1972 and 1975. ABB Power T&D Company. 17 shows as a function of . Power Syst. Signals and Systems. Willsky. Feb. 3rd ed. 11. Swann. respectively. Power System Stability and Control./Feb. vol. Schenectady. R. and Ph. and Systems Engineering at Rensselaer Polytechnic Institute (RPI) . Chow received the B. B. and 1998. degrees from the University of Illinois at Urbana–Champaign. [16] J.S. 2002. Jan. and Ph. NC. 1999. pp. Dr. Sept. Modern Control Systems. [15] P.D.” IEEE Trans. 1664–1672. Venkataraman. 614–626. Concordia. H. Swann. Reading. “Application of power system stabilizers for enhancement of overall system stability. MA. H. Using MATLAB and Using Simulink. IN. NJ: Prentice-Hall. MA: Addison-Wesley. Feb. PAS-100. Troy. Lawson. A. Klein. for a short period from 1997 to 1998. degrees in electric power engineering from Rensselaer Polytechnic Institute. 2000. D. PAS-97. PWRS-4.E. Oppenheim and A. Franklin. Murdoch. and R. S. vol. pp. Energy Conversion. Energy Conversion. His research interests include modeling and control of power systems.. pp. Larsen and D. and tuning methodology. and M.E. he is a Consulting Engineer with General Electric. Do Bomfim. Zywno. Dec. Power App. Power App. Power Syst. deMello and C. p. Joe H. N. and A. and Ph. working in Power Systems Energy Consulting (PSEC). Dec.” IEEE Trans. and G. C.. Reading. Feedback Control Problems using MATLAB and the Control System Toolbox: Brooks/Cole. New York: McGrawHill. pp. vol. NY. Taranto. S. Currently. vol. in 1987. Bishop. [9] R. West Lafayette. F. Venkataraman. Dorf and R. K. 559. “Integral of accelerating power type PSS—Part 1—theory. “Integral of accelerating power type PSS—Part 2—field testing and performance verification. K. and W. Syst. G.” IEEE Trans. Apr. and J. New York: Wiley. and B. Sept. His interests include robust multivariable control.E. MA: Addison-Wesley. vol. Energy Conversion. 1988. pp. Power App. lecting the appropriate REFERENCES [1] G.Math degrees from the University of Minnesota. Before joining RPI. Sanchez-Gasca. Hiyama. 607–615. V. pp. C. 451–459. Schenectady..D. [3] C. Alexander Murdoch received the B. Englewood Cliffs. 9th ed. 1658–1663.” IEEE Trans. D. degrees from Purdue University. H. S. [6] R. “Simultaneous tuning of power system damping controllers using genetic algorithms.D. “Computer models for representation of digital-based excitation systems. 2002. Kuo and M. Golnaraghi. and the M. pp. 1989. NY.CHOW et al.” IEEE Trans.” IEEE Trans. Chow and J. “Robustness of fuzzy logic power system stabilizers applied to multimachine power system.E. Rogers. 15. From 1988 to 1994 and from 1998 to the present he was with the Power Systems Energy Consulting (PSEC) Department of the General Electric Company. [4] D. in 1970. A. 1999. he worked in the power system division of General Electric Company. M. L. Raleigh. [7] B. Power Syst.S. Currently. J. excitation system design and testing. Salem.S. “Applying power system stabilizers.” IEEE Trans. Feedback Control of Dynamic Systems. NY. 9. 2nd ed. 2000. Energy Conversion. VA. 14. 4. George E. M.S. Frederick.. F. New York: Wiley. Murdoch is a member of the Excitation System Subcommittee in the IEEE. Syst. “Concept of synchronous machine stability as affected by excitation control.E. Boukarim received the B. Schenectady. 163–169. and D. 1969..E. [8] P. 2002. 183–190. Emami-Naeini. Powell. . Schenectady. and power system dynamics and control. S. 151 [17] T. Troy. [11] A. NY. vol. PAS-88. June 1981. V. vol.E. vol. Modeling and Analysis of Dynamic Systems. he is Professor of Electrical. [5] The MathWorks. F. P. “Pole-placement designs of power system stabilizers. vol.