reactance scan crossover-based approach for investigating...

742 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 28, NO. 2, APRIL 2013

Reactance Scan Crossover-Based Approach forInvestigating SSCI Concerns for DFIG-Based

Wind TurbinesYunzhi Cheng, Member, IEEE, Mandhir Sahni, Member, IEEE, Dharshana Muthumuni, and

Babak Badrzadeh, Senior Member, IEEE

Abstract—This paper presents a reactance crossover-basedtechnique to investigate subsynchronous control interaction(SSCI) concerns associated with doubly fed induction gener-ator-based wind generation resources. A theoretical discussionserving as the mathematical premise for the proposed approachis presented. The driving point reactance, as seen from the WGR,across the subsynchronous frequency range is determined usingfour different approaches. Specifically, two of the four methods(Methods 3 and 4) are unique in terms of the approach utilizedfor performing frequency scans. System-side frequency scansare augmented with turbine-side frequency scans. A dynamicfrequency-scanning method for the turbine side is developedwhich takes the turbine nonlinearities and its active behaviorinto account. The presence of crossover points in the reactancescans, as obtained from four approaches, in conjunction withthe turbine-side frequency scans is utilized to draw conclusionson potential SSCI concerns. A portion of the Electric ReliabilityCouncil of Texas grid model has been utilized for the case study.The observations/inferences drawn via the reactance scans arecorroborated via electromagnetic transients simulations.

Index Terms—Doubly fed induction generator (DFIG), nonra-dial conditions, reactance crossover index, subsynchronous controlinteraction (SSCI), wind generation resources (WGRs).

NOMENCLATURE

, Total impedance, reactance for thewind turbine looking into the systemfrom the wind turbine terminal bus.

, Impedance, reactance associatedwith the connection/path fromthe wind turbine terminals to thePOI, including pad-mount turbinetransformers, medium-voltage cablesegments, substation transformers,and high-voltage tie lines.

, Impedance of the series-compensatedlines.

Manuscript received March 01, 2012; revised July 20, 2012; accepted Oc-tober 01, 2012. Date of publication January 18, 2013; date of current versionMarch 21, 2013. Paper no. TPWRD-00218-2012.Y. Cheng and M. Sahni are with PWR Solutions, Dallas, TX 75019 USA

(e-mail: [email protected]).D.Muthumuni is withManitoba HVDCResearch Center,Winnipeg,MBR3P

1A3, Canada.B. Badrzadeh is with Vestas Technology R&D, Århus 8200, Denmark.Color versions of one or more of the figures in this paper are available online

at http://ieeexplore.ieee.org.Digital Object Identifier 10.1109/TPWRD.2012.2223239

Reactance of the series capacitors.

, Equivalent impedances in the circuitassociated with Method 3.

, Imaginary part (reactance) of ,.

, , Equivalent impedancescorresponding to the externalsystem in the circuit associated withMethod 4.

, , imaginary part (reactance) of , ,.

Denotation of nominal frequency (60Hz).

Pure imaginary number.

I. INTRODUCTION

S ubsynchronous interaction (SSI) issues associated withwind generation resources (WGRs) have come into promi-

nence since the ERCOT event of October 2009 [1]. Sincethen, a considerable amount of research and development hasbeen dedicated toward investigating SSI issues associated withWGRs. Given the relatively varied and complex nature ofequipment comprising the modern grid, the first step in thatdirection has been an attempt to classify the subsynchronousphenomenon into specific categories. SSI phenomena, in gen-eral, can be classified into the following broad categories [2]:• Subsynchronous resonance (SSR): This phenomenon cor-responds to the traditional concern associated with the res-onance between the mechanical characteristics of the gen-erator turbine-shaft system and the electrical characteris-tics of the series-compensated grid. The phenomenon ofSSR has been observed to manifest itself in the form of tor-sional interaction (TI), induction generator effect (IGE), ortorque amplification (TA).

• Device-dependent subsynchronous oscillations (SSO):This phenomenon relates to the subsynchronous oscilla-tions originating from the interaction of turbine-generatortorsional systems and power system network components.More recently, this term has also been loosely associ-ated with subsynchronous oscillations originating from

0885-8977/$31.00 © 2013 IEEE

CHENG et al.: REACTANCE SCAN CROSSOVER-BASED APPROACH FOR INVESTIGATING SSCI CONCERNS 743

unstable behavior of the control system associated withnetwork components under unique grid conditions.

Subsynchronous control interaction (SSCI), as defined for thepurposes of this paper, relates to the interaction between the se-ries-compensated transmission grid and the control systems as-sociated with a DFIG-based wind turbine. As evident from theaforementioned definition, unlike TI or TA, SSCI is a purelyelectrical phenomenon. While the basic phenomenon character-izing SSCI differs from the existing classifications of SSI, it canbe considered to be another category of SSI albeit specific toDFIG-based wind turbine technology.Reference [3] discusses the development of a Z-bus-based

frequency-scanning program for performing SSR screening.This paper discusses the utilization of the resistance and reac-tance scans to assess three manifestations of subsynchronousissues typically encountered in synchronous machines (i.e.,IGE, TI, and TA). Reference [4] presents a case study asso-ciated with utilizing frequency scans to investigate SSR andpotential torsional interaction problems associated with 35 tur-bine-generator units located at 27 generating plants. Reference[5] presents the findings associated with the investigation ofthe first ever known instance of SSR (i.e., Navajo project).The analytical techniques used, potential problems discovered,solutions considered, and solutions being applied are reviewedin [5]. This paper indicates potential SSR concerns for con-ventional generation units when the frequency-scan resultsindicate the presence of a subsynchronous frequency at whichthe reactance is zero (or close to zero) and the resistance isnegative. Reference [6] presents a comprehensive approach forsubsynchronous resonance (SSR) screening analysis using adeveloped frequency-scanning tool. However, the focus of theinvestigation is on conditions that result in the conventionalgenerator being radial or near radial to the series compensation.A majority of the recent research and development presented

in [7]–[15] focuses on the investigation of potential WGR SSIissues under conditions radial to the series compensation. Thediscussions presented in the aforementioned references are in-tended to investigate SSI and/or TI issues associated with windgeneration via varying techniques ranging from detailed EMTsimulations to small-signal analysis. Discussions on impact ofvarious turbine technologies on the potential for such phenom-enon and associated mitigation methods are also presented.This paper aims to provide the mathematical premise as-

sociated with the potential for SSCI concerns associated withseries-compensated grid conditions and DFIG wind turbinesspanning radial and nonradial conditions. In the context ofthe discussion presented in this paper, radial refers to therelative location of the WGR vis-à-vis the series-compensatedtransmission segment. To that effect, this paper presents areactance scan crossover-based approach to identify potentialSSCI concerns under nonradial system conditions. Reactancecrossover refers to the range of subsynchronous frequency be-tween the two instances wherein the reactance scan crosses overfrom being positive to negative or vice-versa. Four differentapproaches have been utilized to determine the driving pointreactance across the subsynchronous frequency range as seenfrom the WGR. Methods 3 and 4 present unique approaches

for determining frequency scans relative to SSCI screening.While only one such approach may be sufficient in terms ofperforming the frequency scan, each method has its relativemerits/demerits over the others in terms of model requirements,ease of implementation, and subtle nuances in interpretationof results. A comparative analysis of the driving point reac-tance scan across the four approaches has been presented. Thesystem-side frequency scans are augmented with turbine-sidefrequency scans. Traditional harmonic impedance-scanningtechniques are rendered ineffective when nonlinear, active,and/or power-electronic device-based scans are encountered.A dynamic frequency-scanning method for the turbine side isdeveloped which takes he turbine nonlinearities and its activebehavior into account. The apparent resistance and reactance ofthe DFIG-based turbine across the subsynchronous frequencyrange are determined. A vendor-based turbine model has beenutilized for the turbine-side scans. The presence of reactancecrossover points in the driving point reactance scan, as obtainedfrom the four approaches, in conjunction with the turbine-sidescans, is utilized as the major indicator to identify the presenceof potential SSCI concerns under all conditions includingthose where the WGR is nonradial to the series-compensatedsegment. A portion of the ERCOT 345-kV grid has beenutilized to present a case study. System side reactance plotshave been derived using the four approaches outlined before.The inferences drawn using the reactance crossover techniqueand turbine-side scans have been corroborated using EMTsimulations based on the manufacturer-specific turbine model.

II. DRIVING-POINT REACTANCE SCANS

A. Method 1—PSCAD Frequency Scan Technique

The first of the four methods utilized to assess the drivingpoint reactance associated with the WGR across subsyn-chronous frequency ranges is the Harmonic Impedance Scantool in PSCAD. The system positive-sequence model is con-verted to PSCAD using third party software. The area ofinterest associated with the WGR under study is retained incomplete detail while the rest of the system is equivalenced.The following aspects related to the WGR under study are alsoincluded in the study model:• WGR station transformers;• pad-mount turbine transformer equivalent;• generation interconnection tieline, if applicable;• aggregated collection system model [16].The harmonic impedance scan tool is then connected behind

the pad-mount turbine transformer equivalent looking into thesystem. It is important to note that the turbine model and/orpower-flow equivalents of the same are disconnected when per-forming the system-side reactance scans. The scan is performedacross the entire range of subsynchronous frequencies with a1-Hz increment.A key advantage of this method is that the frequency depen-

dency of the line parameters (line inductances) can be accountedfor in the calculations.


Fig. 1. System frequency scans, frequency-dependent versus frequency-inde-pendent line model.

B. Method 2—Short-Circuit Calculation on theFrequency-Dependent Network

In this method, the driving-point reactance (Thevenin’s reac-tance) at various subsynchronous frequencies can be obtainedby performing a solid three-phase short circuit at the point ofinterest [17]. The following key steps comprise the computa-tion of the driving-point reactance using this method.• The reactance of the network components is linearly scaledacross the entire range of subsynchronous frequencies ofinterest.

• The network data at each subsynchronous frequency of in-terest is subjected to a solid three-phase short circuit at thepoint of interest (i.e., behind the pad-mount turbine trans-former equivalent).

• The three-phase short-circuit current (I3p, per unit) atthe point of interest is utilized to obtain the XTOTAL

at different subsynchronous frequencies (0 to60 Hz).

The key advantages of this method are ease of implemen-tation for large networks and the absence of any need to con-vert to PSCAD and/or use an EMT environment. While the fre-quency-dependent effects of the line inductances are not consid-ered, this is a valid approximation as illustrated in Fig. 1. Thefrequency dependence of the line inductances is minimal in thesubsynchronous frequency range of interest.

C. Method 3—Frequency-Scaling of the Equivalent NetworkDerived From Short-Circuit Calculation

This method is similar to Method 2 except that there is noneed to scale the entire system data for each subsynchronousfrequency of interest. Conversely, three-phase short-circuit cal-culations are performed for two scenarios at 60 Hzwith the windfarm disconnected in both scenarios:• three-phase solid fault at POI bus, noted as Isc_a;• three-phase solid fault at the POI bus with the series com-pensation out-of-service, noted as Isc_b.

Using this method, it is possible to represent any WGR con-nection to the system in the form of the IEEE 2nd SSR Bench-mark model [18] as shown in Fig. 2.The short-circuit currents obtained from the two scenarios

outlined before can be used to represent the WGR connectionto the system in a manner shown in Fig. 2. The short-circuit

Fig. 2. Simplified representation of the WGR connection.

currents for the two scenarios can be related to the network pa-rameters depicted in Fig. 2 using (1) and (2)

(1)

(2)

Assume that and are linearly dependent(proportional) on the frequency

(3)

(4)

In accordance with Fig. 2, and can bewritten as

(5)

Imaginary

(6)

In addition, the reactance of the wind farm (not including theturbines) and line is frequency dependent as shown

(7)

(8)

However, for the series capacitors, the reactance is inverselyproportional to the frequency as shown

(9)

Since the transmission system is typically characterized by ahigh X/R ratio (typically 10 or more), can be rewrittenas

(10)


Fig. 3. Simplified representation using the network reduction method.

Assuming will be infinite (corresponding to the ra-dial connection between the wind farm and the series compen-sation), will be

(11)When the frequency is below , becomes negative

and may cause the subsynchronous interactions

(12)

However, in the presence of other lines in parallel to the se-ries compensation at the POI bus, will not be infiniteanymore. In such nonradial conditions, when the issmall, which corresponds to a strong connection, willbe positive through the entire subsynchronous frequency range.On the contrary, when is large, which corresponds toa weak connection, could be negative at a particularsubsynchronous frequency. The mathematical premise for theaforementioned concept is as shown below.Assume 0, 0, the reactance of the parallel circuit

of and will be which implies

when (13)

when (14)

Like Method 2, the key advantages of this method are easeof implementation for large networks and the absence of anyneed to convert to PSCAD and/or use of an EMT environment.However, unlike Method 2, one key limitation of this approachis the inability to accurately reflect the frequency dependencyof the network represented in the equivalence systems. This isof particular significance if the equivalence system is comprisedof series-compensated lines within close electrical proximity ofthe boundary bus. Application of this method requires carefulreview and in-depth knowledge of the system to identify theboundary buses.

D. Method 4—Network Reduction Technique

The system network under study can be reduced to thetwo-port equivalent network with the two buses (ports) cor-responding to the two ends of the series-compensated line asshown in Fig. 3. The external system is represented by two

voltage-source equivalents and three impedances ( , , and).The network reduction can be achieved by reducing the large

network full admittance matrix to a small admittancematrix corresponding to the retained buses. The net-work full admittance matrix includes all branches, gen-erators, loads, switch shunts, etc. The reduction process can beperformed by a complex, sparse LDU decomposition routine[19]. The internal admittance matrix is formed to rep-resent the retained/internal network component (series compen-sation and the line). The admittance matrix of the equivalent ofthe external system can be formed by the following formu-lation:

According to [19], , , and will be calculated byfor 60 Hz. After , , and

have been determined, can be calculated as

(15)

(16)

Assume 60, 60, and60, then can be evaluated using

(15).The key limitation of this method is the same as that described

for Method 3.

III. TURBINE SIDE SCANS

The impedance characteristics of passive electric network el-ements can be determined by any of the four approaches out-lined before. In either of the aforementioned approaches, theimpedance seen from a particular point is determined by calcu-lating the equivalent (driving point) impedance seen from thatpoint, based on the series or parallel connections of network ele-ments ( , , and ). However, the presence of dynamic devicesand/or power-electronic devices in the network under study canimpact the impedance characteristics of the network. The effec-tive impedance as seen from the point of interest at differentfrequencies can be influenced by the characteristics of such de-vices and the control systems contained therein.In such situations, voltage or current injection-based

impedance scanning methods have to be used to obtain ac-curate impedance characteristics over a frequency range ofinterest. When considering a highly nonlinear and active device,such as a wind turbine, the frequency-scanning method utilizedmust be able to take these nonlinearities into account. The in-jected signal should be a wideband signal, such as an impulse orwhite noise, to avoid the need for multiple simulation runs foreach subsynchronous frequency. An initial challenge dealingwith the wind turbine or other similar nonlinear elements is thatdue to nonlinearities, the principle of superposition does not


Fig. 4. Turbine-side scan connection.

Fig. 5. Turbine-side frequency scan.

hold true. The system, however, can still be considered linearfor small changes around an operating point. The harmonicimpedance of a nonlinear system, linearized about an operatingpoint, can be determined with a fast Fourier transformation(FFT). This forms the basis for the turbine-side frequency-scan-ning technique utilized to perform turbine-side scans.To obtain frequency scans of the wind turbine, the turbine

model was disconnected from the interconnected network andconnected to an ideal voltage source. This is because the focusof the investigation is the turbine-side impedance scans and thecharacteristics are solely dependent on the response of theWTGand its controls. Fig. 4 depicts the connection of the wind turbinewhen performing turbine-side scans.It is important to select the injected current or voltage signals

judiciously in order to obtain accurate results. The magnitude ofthe injected signal as well as the injected location has to be care-fully selected based on the phenomena that are under study. Avery important consideration is that the system is running at thedesired steady state (power, reactive power, voltage, machinespeed, etc. are steady and at a desired operating point) and theinjection of the voltage or current signal should not alter the op-erating point. The injection should only cause “small” oscilla-tions around the steady-state operating point. The magnitude ofthe selected signal should be small enough to satisfy the afore-mentioned condition while being large enough to provide accu-rate measurements to determine the impedance characteristics.Turbine-side scans were performed on the DFIG wind tur-

bine model. More important, instead of plotting the positive-se-quence impedance as obtained from the current injection tech-nique, the resistive and reactive components of the turbine ap-parent impedance were plotted. Fig. 5 depicts the plots of theturbine apparent resistance and reactance as obtained from thefrequency scans.Interestingly, the apparent resistance of the turbine model is

negative across the entire range of subsynchronous frequenciestested. In other words, the controls associated with the DFIGwind turbine result in the machine exhibiting negative dampingat subsynchronous frequencies. It is important to note that the

Fig. 6. Relevant portion of the ERCOT grid model utilized for the case study.

turbine-side scans depicted in Fig. 5 are specific to the windturbine model studied here. Turbine-side scans for variousmanufacturers DFIG models are expected to vary from modelto model. In addition, it can be concluded that the potentialfor SSCI (per the definition) is expected to exist if the systemexhibits resonant frequencies within the subsynchronousfrequency range over which the DFIG turbine resistance isnegative.

IV. CASE STUDY

Each of the four reactance scan techniques has been appliedto a portion of the ERCOT grid model to assess the ability touse reactance crossover in conjunction with turbine-side scansto assess the potential for SSCI concerns. Furthermore, the re-sults of the reactance scans have been further corroborated bymeans of EMT simulation using the ERCOT grid model and amanufacturer-specific turbine model.

A. System Description

A section of the ERCOT grid model has been utilized for thecase study. Fig. 6 depicts the relevant portion of the ERCOTgrid model. All buses depicted in Fig. 6 are rated at 345 kV. Thedetailed system and turbine data are provided in the Appendix.As evident from Fig. 6, the proposed WGR, sized at 500 MWand comprised of Type 3 turbines, is connected at Bus 5. Theseries compensation is located between the following buses:• Bus 6–Bus 7;• Bus 8–Bus 9.The series compensation level is around 50% of the total line

impedance between Bus 5 and Bus 10. It is important to note thatwhile only a limited section of the grid model has been shownin Fig. 6, a larger albeit equivalenced section of the transmis-sion system was utilized for the system-side frequency scans.Although the frequency scans were performed at numerous lo-cations within the system to verify the robustness of the ap-proaches, a specific location is used for demonstrative purposes.

B. Reactance Scans—Method 1–Method 4

Table I depicts the scenarios pertaining to the system condi-tions that were utilized for the system-side scans. The drivingpoint reactance scans of the system were performed for each of


TABLE ISCENARIO DEFINITIONS

Fig. 7. System impedance scans, Scenarios 1–5.

Fig. 8. System reactance scans, Methods 1–4, Scenario 1.

the scenarios depicted in Table I using each of the four methodsoutlined in Section III.It is important to note that the only condition of those out-

lined in Scenarios 1– 5 resulting in the WGR being radial tothe series capacitor is Scenario 5. Fig. 7 depicts the system-side impedance scans associated with Scenarios 1–4. As evi-dent from Fig. 7, while the impedance scans for each scenariodepict a prominent dip, there is no clear indicator that separatesthe scenarios in terms of SSCI potential.Figs. 8–12 depict the results of the driving-point reactance

scans of the system from the point of interest. As evident fromFigs. 8–12, the reactance scans have been performed using eachof the four methods outlined in Section III. As evident fromFigs. 8–12, the reactance scans for Scenarios 1 through 5 pro-vide more insight into the potential for SSCI concerns relativeto the impedance scans. While the dips in the impedance scans




make it difficult to utilize the information to screen for prob-lematic scenarios that does not seem to be the case for reactancescans, especially when using the crossover approach.As evident from the results of the reactance scans, Method

1 and Method 2 results seem to line up very closely. WhileMethods 3 and 4 provide certain variation in terms of the mag-nitude and point of reactance crossover, it is interesting to notethe following.• Reactance scans derived from all four methods are con-sistent in terms of the scenarios that result in reactancecrossover (i.e., Scenarios 3–5).



Fig. 13. EMT simulation results, Case 1).


• For the scenarios exhibiting reactance crossover, the sub-synchronous frequency range of interest is also very closeacross all five scenarios.

• Finally, the subsynchronous frequency range at whichreactance crossover is observed is within the range overwhich the DFIG turbine is observed to exhibit negativedamping.

Based on the aforementioned items and the observationsmade from the turbine-side scans, Scenarios 3–5 are indicativeof potential SSCI concerns. For Scenarios 3–5, the system-side

Fig. 15. EMT Simulation Results, Case 3).

reactance scans are indicative of a crossover range across thefollowing subsynchronous frequencies:• Scenario 3: 23–28 Hz;• Scenario 4: 16–26 Hz;• Scenario 5: 0–25 Hz.In addition, for these scenarios, the subsynchronous fre-

quency range over which reactance crossover occurs coincideswith the region wherein the turbine-side scans indicate neg-ative resistance and/or damping. This underlines that whilethe system and turbine-side scans are necessary conditionsfor evaluating SSCI conditions associated with DFIG-basedwind farms, each of them, by itself, is not a sufficient conditionto make conclusive judgments. The combination of turbineand system-side scans provides enough information to makeconclusive observations regarding conditions poising potentialSSCI concerns.It is important to note that Scenarios 3 and 4 are nonradial

conditions that have also been shortlisted as those exhibitingSSCI concerns. The results of the inferences drawn from the re-actance scans have been further corroborated by means of EMTsimulations presented in the ensuing subsection.

C. EMT Simulations

In order to corroborate the inferences drawn from the system-side reactance, crossover points, in conjunction with the turbine-side scans, showed the following scenarios of EMT simulations:Case 1) EMT simulation associated with Scenario 1;Case 2) EMT simulation associated with faultless outage of

Scenario 2;Case 3) EMT simulation associated with faultless outage of

Scenario 3;Case 4) EMT simulation associated with fault based outage

of Scenario 3;Case 5) EMT simulation associated with faultless outage of

Scenario 4;Case 6) EMT simulation associated with faultless outage of

Scenario 5.Unlike fault-based outages, faultless outages involve the loss

of lines comprising the contingency without a system fault pre-ceding the event. In case of fault-based outages, a three-phasefault is applied at the point of common coupling associated with




the wind farm under study for a duration of four cycles (typicalfor 345-kV facilities). Both three-phase fault and line outagesoccur at 5 s. As mentioned earlier, the manufacturer-specificDFIG turbine model was utilized to model the 500-MW WGRat Bus 5. Figs. 13–18 depict the results of the EMT simulationsassociated with Cases 1)–6), respectively.As evident from the results of the EMT simulations presented

in Figs. 13 and 14, EMT cases corresponding to Scenarios 1 and2 (for the reactance scans) do not indicate any SSCI concerns.Moreover, even a system disturbance in the form of a faultlessoutage for Case 2) does not trigger any SSCI concerns for Sce-nario 2. The results associated with EMT simulations for Cases3)–6), as depicted in Figs. 15–18, respectively, indicate the pres-ence of the SSCI concern for Scenarios 3–5 corresponding toweakly nonradial and radial conditions. The magnitude of theSSCI concern becoming exaggerated from Case 3) to Case 6)is evident from the results depicted in Figs. 15–18. Specificallyspeaking, minor but sustained subsynchronous oscillations areobserved for Case 3) involving the faultless outage of the doublecircuit between Buses 5 and 12. However, further investiga-tion is performed for the same outage albeit initiated througha fault-based outage in the form of Case 4. As obvious fromFig. 16, the fault-based outage of the double circuit between


Fig. 19. EMT simulation results FFT analysis, Case 5 (Iwf, the total wind farmcurrent output).

Bus 5 and Bus 12 results is significant SSCI concerns. Expect-edly, the SSCI concern is observed to be further exacerbatedunder near radial conditions as exhibited from Fig. 17 outliningresults of EMT simulations for Case 5. The exacerbated SSCIconcerns are observed to result in the turbine model to trip atabout 8.5 s. Finally, the worst case in terms of SSCI concerns,as also depicted from the reactance scans, is Case 6) (i.e., com-pletely radial condition). In case of Case 6), the extreme SSCIconcerns are observed to result in an almost immediate trip ofthe turbine model from the grid.In order to corroborate the findings of the reactance scans

with the EMT simulation results, FFT analysis was performedon the turbine current signal under a condition exhibiting SSCIconcern. Fig. 19 exhibits the results of the FFT analysis on theturbine current signal for Case 5). As evident from Fig. 19,a dominant subsynchronous oscillation frequency of 23 Hz isobserved via the FFT analysis for the EMT simulation resultsfor Case 5). The dominant mode of oscillation observed duringEMT simulation is observed to line up very closely to the reac-tance crossover approach for Scenario 4 in Fig. 11.


TABLE IISYSTEM DATA

TABLE IIITURBINE DATA

V. CONCLUSION

The discussion presented in this paper has endeavored to uti-lize a reactance crossover-based approach in conjunction withturbine-side scans for investigating potential SSCI concerns as-sociated with DFIG-basedWGRs. System-side frequency scansare augmented with turbine-side frequency scans. A dynamicfrequency-scanning method for the turbine side is developedwhich takes the turbine nonlinearities and its active behaviorinto account. A portion of the ERCOT grid model has been uti-lized to demonstrate the practical feasibility and application ofthe approach. Comprehensive EMT simulations have been per-formed to corroborate the findings outlined via reactance scancrossover and turbine-side scans.The key contributions of the research documented in the

paper include the use of reactance crossover techniques inconjunction with turbine-side scans for evaluating an emergingand complex issue (i.e., SSCI). In addition, the investigationsdocumented in this paper underlie the importance of systemstrength in relation to the potential for SSCI conditions es-pecially under nonradial conditions. Finally, comprehensiveEMT simulations and postprocessing are indicative of highcorrelation between the results of the reactance scans and thedominant mode of subsynchronous oscillations.The turbine-side scans presented in this paper are specific to a

particular manufacturer model. Additional research is necessaryto specify generic guidelines associated with SSCI concerns forDFIG technology. Moreover, the dynamic frequency-scanningapproach utilized to perform the turbine-side scan has to betuned when using different turbine models. Additional work isnecessary to standardize the dynamic frequency-scanning ap-proach to be independent of the turbine model being subjectedto the scan.

APPENDIXSYSTEM AND TURBINE DATA

System and turbine and data can be see in Tables II and III.

REFERENCES[1] P. Belkin, “Event of 10-22-09,” presented at the CREZ Tech. Conf.,

Taylor, TX, Jan. 26, 2010. [Online]. Available: www.ercot.com/con-tent/meetings/rpg-crez/keydocs/2010/0126/8Belkin_Event%20of%2010.ppt

[2] “Terms, definitions and symbols for subsynchronous oscillations,”IEEE Trans. Power App. Syst., vol. PAS-104, no. 6, pp. 1326–1334,Jun. 1985.

[3] S. R. Puchalapalli, R. G. Farmer, G. G. Karady, J. Hernandez, and Y.Suh, “z-bus based frequency scanning program for sub-synchronousresonance screening,” in Proc. IEEE Power Tech., Lausanne, Switzer-land, 2007, pp. 149–154.

[4] M. El-Marsafawy, “Use of frequency-scan techniques for subsyn-chronous-resonance analysis of a practical series-capacitor compen-sated AC network,” Proc. Inst. Elect. Eng., Gen., Transm. Distrib.,vol. 130, no. 1, pp. 28–40, Jan. 1983.

[5] R. G. Farmer, A. L. Schwalb, and E. Katz, “Navajo project report onsubsynchronous resonance analysis and solutions,” IEEE Trans. PowerApp. Syst., vol. PAS-96, no. 4, pp. 1226–1232, Jul. 1977.

[6] M. Elfayoumy and C. G. Moran, “A comprehensive approach forsubsynchronous resonance screening analysis using frequency scan-ning technique,” presented at the IEEE Bologna Power Tech Conf.,Bologna, Italy, 2003.

[7] A. K. Jindal, G. D. Irwin, and D. A. Woodford, “Sub-Synchronousinteractions with wind farms connected near series compensated AClines,” in Proc. 9th Int. Workshop Large Scale Integr. Wind, QuebecCity, QC, Canada, Oct. 18–19, 2011, pp. 559–564.

[8] G. D. Irwin, A. K. Jindal, and D. A. Woodford, “Sub-synchronous con-trol interactions between type 3 wind turbines and series compensatedAC transmission systems,” presented at the IEEE Power Eng. Soc. Gen.Meeting, Detroit, MI, Jul. 2011.

[9] C. Karawita and U. D. Annakkage, “Control block diagram represen-tation of an HVDC system for sub-synchronous frequency interactionstudies,” presented at the 9th Inst. Eng. Technol. Int. Conf. AC DCPower Transm., London, U.K., 2010.

[10] Y. Tang and R. Yu, “Impacts of large-scale wind power integrationon subsynchronous resonance,” presented at the Power Energy Eng.Conf., Wuhan, China, Mar. 25–28, 2011.

[11] L. Fan, R. Kavasseri, Z. L. Miao, and C. Zhu, “Modeling of dfig-basedwind farms for ssr analysis,” IEEE Trans. Power Del., vol. 25, no. 4,pp. 2073–2082, Oct. 2010.

[12] R. K. Varma, S. Auddy, and Y. Semsedini, “Mitigation of subsyn-chronous resonance in a series-compensated wind farm using FACTScontrollers,” IEEE Trans. Power Del., vol. 23, no. 3, pp. 1645–1654,Jul. 2008.

[13] S. Golshannavaz, M. Mokhtari, and D. Nazarpour, “SSR suppressionvia STATCOM in series compensated wind farm integrations,” pre-sented at the 19th Iranian Conf. Elect. Eng., Tehran, Iran, May 17–19,2011.

[14] M. S. El-Moursi, B. Bak-JenseM, and H. Abdel-Rahman, “NovelSTATCOM controller for mitigating SSR and damping power systemoscillations in a series compensated wind park,” IEEE Trans. PowerElectron., vol. 25, no. 2, pp. 429–441, Feb. 2010.

[15] X. Zheng, Z. Xu, and J. Zhang, “A supplementary damping controllerof TCSC for mitigating SSR,” presented at the Power Energy Soc. Gen.Meeting, Calgary, AB, Canada, Jul. 26–30, 2009.

[16] Y. Cheng, M. Sahni, J. Conto, S.-H. Huang, and J. Schmall, “Voltageprofile based approach for developing collection system aggregatedmodels for wind generation resources for grid voltage ride-throughstudies,” Inst. Eng. Technol. Renew. Power Gen. J., vol. 5, no. 5, pp.332–346, 2011.

[17] T. Rauhala, P. Järventausta, and H. Kuisti, “Frequency scanning pro-gram for SSR studies implemented to function in connection of PSS/Epower flow analysis program,” presented at the 15th PSCC, Liege, Bel-gium, Aug. 2005.

[18] “Second benchmarkmodel for computer simulation of subsynchronousresonance,” IEEE Trans. Power App. Syst., vol. PAS-104, no. 5, pp.1057–1066, May 1985.

[19] “E-TRAN User’s Manual,” Electranix Corp., Winnipeg, MB,Canada. [Online]. Available: http://www.electranix.com/E-TRAN/E-TRAN_Chap9_Theory.pdf, Ch. 9


Yunzhi Cheng (M’12) received the B.S. and M.S.degrees in electrical engineering from Shanghai Jiao-tong University, Shanghai, China, and the Ph.D. de-gree in electrical engineering from the University ofTexas, Arlington.During 2003 to 2006, he was a Transmission and

Generation Planning Engineer at the East ChinaElectric Power Design Institute, focusing on powersystem planning for East China Power Grid. Since2009, he has been a Senior Research Engineer ofGeneration & Transmission Planning at PWR Solu-

tions, Dallas TX. His research interests include renewable energy modeldingand integration, dynamic parameter identification, state estimation, hydrgoeneconomy, and power markets.

Mandhir Sahni (M’12) received the B.S. degree inelectrical engineering from Bangalore University,Bangalore, India, and the M.S. and Ph.D. degrees inelectrical engineering from the University of Texas,Arlington.He has more than ten years of experience in power

market analysis; generation and transmission plan-ning in deregulated energy markets in the U.S; powerquality; as well as stability and protection coordina-tion analysis with an extensive background in powersystem analysis. He has served as theTechnical Lead

on the Electrical Reliability Council of Texas Voltage Ride-Through (VRT)study effort. He is currently leading all transmission planning activities asso-ciated with numerous CREZ utilities in ERCOT. He has also served as expertwitness on behalf of the staff of the West Virginia Public Service Commissionfor the hearing associated with the 765 kV Potomac Appalachian TransmissionHighway (PATH) project in Mid-Atlantic PJM. His research has been in pub-lished in numerous referred journals, IEEE Power and Energy Magazine andconference proceedings. His research interests include power system planning,integration of renewable generation into power system grids, power quality, andpower system reliability.

Dharshana Muthumuni received the Ph.D. degreein electrical engineering from the University of Man-itoba, Winnipeg, MB, Canada, in 2001.He joined the Manitoba HVDC Research Centre

(MHRC), Winnipeg, in 2001 and is currently theManager of the Engineering Simulations and StudiesDepartment. He has more than 17 years of experiencein engineering studies using a variety of simulationproducts during his career, including PSCAD andPSS/E./EMTDC. He has led the MHRC technicalteam to solve challenging engineering problems,

including wind farm modeling and integration, SSR, black start restoration, andinterconnection issues. He is the MHRC’s machines and Transformer Modelingand Simulation Specialist and plays an active role in developing new modelsof power system apparatus for transient simulation studies, working closelywith equipment manufacturers to develop detailed simulation models. Some ofhis work includes the development and validation testing of detailed machinemodels, transformer and current-transformer saturation routines, developmentof models for wind power and solar applications, and complex magnetic faultcurrent limiter models.

Babak Badrzadeh (S’03– M’07–SM’12) receivedthe B.Sc. and M.Sc. degrees in electrical power engi-neering from the IranUniversity of Science and Tech-nology, Tehran, Iran, in 1999 and 2002, respectively,and the Ph.D. degree in electrical power engineeringfrom Robert Gordon University, Aberdeen, U.K., in2007.After spending a short period as an Assistant pro-

fessor at the Technical University of Denmark, KgsLyngby, he joined Mott MacDonald, Transmissionand Distribution Division, Glasgow/Brighton U.K.,

as a System Analysis and Network Planning Engineer. Since 2010, he has beenwith Power Plant Solutions, Vestas Technology R&D, Aarhus, Denmark, wherehe is a Lead Engineer in the area of advanced wind powerplant simulation andanalysis. He has published several articles and presented tutorials in IEEE or-ganized conference on different areas of power systems and power electronics.He has prepared two two-part educational courses for the IEEE e-learninglibrary on high-power variable speed drives, and HVDC transmission systems.He is a Guest Editor for the special issue of IEEE Industry ApplicationsMagazine on high-power variable speed drives. His areas of interest includespower system electromechanical and electromagnetic transients, application ofpower electronics in power systems, wind powerplants, as well as modelingand simulation.Dr. Badrzadeh is an active member of IEC TC 88 WG 27, WECC REMTF,

and Danish A11 standard all in the area of electrical simulation models for windpower generation.

本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP

图书馆。

图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页文献云下载图书馆入口外文数据库大全疑难文献辅助工具

http://www.xuebalib.com/cloud/

http://www.xuebalib.com/

http://www.xuebalib.com/cloud/


http://www.xuebalib.com/vip.html

http://www.xuebalib.com/db.php

http://www.xuebalib.com/zixun/2014-08-15/44.html


reactance scan crossover-based approach for investigating...

Documents