Voltage security constrained reactive power optimization incorporating wind generation

Transcription

1 Unversty of Wollongong Research Onlne Faculty of Engneerng and Informaton Scences - Papers: Part A Faculty of Engneerng and Informaton Scences 2012 Voltage securty constraned reactve power optmzaton ncorporatng wnd generaton L G. Meegahapola Unversty of Wollongong, lasantha.meegahapola@rmt.edu.au E Vttal Unversty College Dubln, eknath.vttal@ucd.e A Keane Unversty College Dubln, andrew.keane@ucd.e D Flynn Unversty College Dubln Publcaton Detals L. G. Meegahapola, E. Vttal, A. Keane & D. Flynn, "Voltage securty constraned reactve power optmzaton ncorporatng wnd generaton," n IEEE Internatonal Conference on Power System Technology (POWERCON), 2012, pp Research Onlne s the open access nsttutonal repostory for the Unversty of Wollongong. For further nformaton contact the UOW Lbrary: research-pubs@uow.edu.au

2 Voltage securty constraned reactve power optmzaton ncorporatng wnd generaton Abstract Ths paper presents a comparatve analyss between conventonal optmal power flow (OPF) and voltage constraned OPF strateges wth wnd generaton. The study has been performed usng the New England 39 bus system wth 12 doublyfed nducton generator (DFIG) based wnd farms nstalled across the network. A voltage securty assessment s carred out to determne the crtcal wnd farms for voltage stablty enhancement. The power losses and ndvdual wnd farm reactve power generaton have been compared wth and wthout voltage stablty constrants mposed on the OPF smulaton. It s shown that voltage constraned OPF leads to much greater actve power losses n the network. Furthermore, the reactve power contrbuton of each wnd farm s determned and a selectve optmzaton completed to evaluate ndvdual wnd farm contrbutons towards system actve power losses. Moreover, number of reactve power optmzed wnd farms can be reduced by only usng those wnd farms whch contrbute least to system actve power losses. In addton, selectve voltage constraned OPF can also be performed to mnmze the adverse effect on system losses. Ultmately, the system operator should select the optmal wnd farms for both voltage stablty and loss mnmzaton consderng the trade-off between energy savngs and voltage securty. Keywords securty, constraned, reactve, power, optmzaton, voltage, ncorporatng, generaton, wnd Dscplnes Engneerng Scence and Technology Studes Publcaton Detals L. G. Meegahapola, E. Vttal, A. Keane & D. Flynn, "Voltage securty constraned reactve power optmzaton ncorporatng wnd generaton," n IEEE Internatonal Conference on Power System Technology (POWERCON), 2012, pp Ths conference paper s avalable at Research Onlne:

3 1 Voltage Securty Constraned Reactve Power Optmzaton Incorporatng Wnd Generaton L.G. Meegahapola, Member, IEEE, and E. Vttal, Member, IEEE, A. Keane, Member, IEEE, D. Flynn, Senor Member, IEEE Abstract Ths paper presents a comparatve analyss between conventonal optmal power flow (OPF) and voltage constraned OPF strateges wth wnd generaton. The study has been performed usng the New England 39 bus system wth 12 doublyfed nducton generator (DFIG) based wnd farms nstalled across the network. A voltage securty assessment s carred out to determne the crtcal wnd farms for voltage stablty enhancement. The power losses and ndvdual wnd farm reactve power generaton have been compared wth and wthout voltage stablty constrants mposed on the OPF smulaton. It s shown that voltage constraned OPF leads to much greater actve power losses n the network. Furthermore, the reactve power contrbuton of each wnd farm s determned and a selectve optmzaton completed to evaluate ndvdual wnd farm contrbutons towards system actve power losses. Moreover, number of reactve power optmzed wnd farms can be reduced by only usng those wnd farms whch contrbute least to system actve power losses. In addton, selectve voltage constraned OPF can also be performed to mnmze the adverse effect on system losses. Ultmately, the system operator should select the optmal wnd farms for both voltage stablty and loss mnmzaton consderng the trade-off between energy savngs and voltage securty. Index Terms doubly-fed nducton generator (DFIG), loss mnmzaton, Prony analyss, reactve power optmzaton, voltage securty. I. INTRODUCTION OLTAGE securty and reactve power are two major Vconcerns for renewable energy domnant power networks. An nablty to meet reactve power demand wll result n addtonal actve power losses and voltage securty may be threatened due to nadequate voltage control mechansms n the power networks. However, both voltage and reactve power are nterrelated; therefore maxmzaton of one objectve may result n an adverse mpact on the other. In terms of wnd generaton, most power electroncs based wnd generators (e.g. doubly-fed nducton generator (DFIG), fullconverter wnd generator (FCWG)) offer control flexblty over ther reactve power generaton, and these generators possess consderable reactve power capablty wthn ther generator and converter system. In the publshed lterature wnd farm reactve power capablty s utlzed to mnmze losses n the transmsson system [1-2]. Ths s acheved by Lasantha Meegahapola s wth the Endeavour Energy Power Qualty and Relablty Centre, Unversty of Wollongong, Wollongong, 2500, Australa. (e-mal: lasantha@uow.edu.au). Eknath Vttal, Andrew Keane and Daman Flynn are wth the Unversty College Dubln, Ireland (e-mal addresses: eknath.vttal@ucd.e, andrew.keane@ucd.e, daman.flynn@ucd.e). optmal power flow (OPF) analyss constraned to network capablty and grd-codes [1]. A number of studes have been publshed on securty constraned OPF methods [3-8]. In these research studes the voltage stablty and other securty constrants are employed wthn the optmzaton algorthm usng securty ndces, hence nfluence of ndvdual generators on system securty haven t explctly consdered. In addton, these studes are lmted to conventonal power networks wthout any wnd generaton. Moreover, a number of OPF studes have been performed wth wnd generaton [1-2], however n these studes voltage securty constrants haven t been explctly consdered wthn the optmzaton algorthm. The wnd farms used n ths study are based on DFIGs and they can be operated ether n voltage control or power factor control mode. In the termnal voltage control strategy, reactve power producton s controlled to acheve a target voltage at the desgnated bus, whle for fxed power factor control reactve power s produced n proporton to the actve power output. In systems wth hgh penetratons of wnd generaton, voltage securty s also a crtcal ssue that must be addressed. Snce wnd generaton s hghly varable, applcaton of reactve power control from the wnd farms can play an mportant role n mprovng system voltage securty [9]. In ths presented study crtcal wnd farms are screened consderng ther partcpaton factor to determne the wnd farms whch requre voltage control strategy n order to mprove voltage stablty. Comparatve analyss s then conducted between conventonal OPF and voltage constraned OPF methods, crtcally analyzng the reactve power dspatch and network actve power losses. Fnally, selectve optmzaton s carred out n order to determne the most sutable wnd farms for optmzaton. Ths paper s structured as follows: the optmzaton algorthm and network confguraton are descrbed n Secton II. Voltage securty assessment for wnd farms s presented n Secton III. A comparatve analyss between conventonal OPF and voltage securty constraned OPF s presented n Secton IV. In secton V selectve optmzaton s carred out to determne the wnd farms wth sgnfcant contrbuton to system loss reducton. Dscusson and conclusons are presented n Secton VI and VII respectvely. II. OPTIMIZATION ALGORITHM AND NETWORK CONFIGURATION A. Optmal Power Flow for Loss Mnmzaton The man objectve of OPF s to utlze the reactve power capablty of the DFIG wnd farms for system actve power

4 2 loss mnmzaton. The objectve here s to mnmze the actve power loss n the network, subject to the followng objectve functon for loss mnmzaton. F = = k k N b k (, j) 2 g ( V + V 2 j 2V V j cosθ ) where, V, V j g k, N b, and θ denote the voltage at bus, the voltage at bus j, the conductance of branch k, the number of branches n the system, and the voltage angle dfference between bus and bus j respectvely. Regardless of the objectve functon, however, an OPF must ensure that the entre set of voltage and power constrants are satsfed. Varous categores of constrants exst, and these dstnct categores are descrbed below. 1) Equalty constrants The transmsson network s modelled by a power balance equaton at each node. The algebrac sum of the actve and reactve powers njected nto each node must equal zero: (1) the mnmum and maxmum reactve power lmts of the generator at bus respectvely. Voltage lmts constran bus voltages (V ) to reman wthn an allowable range. Our assumpton here s that node voltages are mantaned between 0.95 and 1.05 pu. 0.95pu V 1.05 pu, N B (6) The formulated OPF problem was then solved usng the optmal power flow faclty of DIgSILENT Power Factory [10]. B. Network Confguraton The New England 39 bus system was modfed by ntroducng DFIG based wnd farms across the network whle creatng wnd-rch hgh demand regons. The wnd farm locatons were chosen based on ther proxmty to synchronous generaton and system loads as llustrated n Fg. 1. P V Q V N B j= 1 N B j= 1 V ( G cosθ + B snθ ) = 0 j j V ( G cosθ B snθ ) = 0, N where P, Q, N B, G, and B denote the actve power njected at bus, the reactve power njecton at bus, the number of buses n the system, the mutual conductance between bus and bus j, and the mutual susceptance between bus and bus j respectvely. Each wnd farm s actve power output s also consdered to be at a fxed value for a specfc wnd condton durng the optmzaton process. B (2) N W PWG = P, wnd N (3) W = 1 where P WG, P wnd and N W denote the actve power output of wnd farm, the total wnd power generaton, and the total number of wnd farms n the system respectvely. 2) Inequalty constrants The DFIG reactve power output (Q DFIG ) can be controlled, and the followng nequalty constrant can be ncluded wthn the OPF framework. The DFIG reactve power capablty (Q DFIG ) for a power factor range of 0.95 leadng to 0.95 laggng can be expressed as follows: 0.328pu Q pu, (4) DFIG N W The conventonal generatng unts have maxmum and mnmum generatng lmts, both for real and reactve power, beyond whch t s not feasble to generate for techncal or economc reasons. mn max P P P g g g mn max (5) Q Q Q N g g g G where P g mn, P g max Q g mn, and Q g max denote the mnmum and maxmum actve power lmts of the generator at bus, and Fg. 1. New England 39 test system There are 12 wnd farms nstalled n the test system. For each case the maxmum capacty of the farm was vared to create dfferent penetraton levels n the system. In the frst case, the nstalled capacty of each of the 12 farms was 50 MW resultng n a penetraton level of 10.4% (600 MW). The second case had a penetraton of 20.7% (1,200 MW) wth farms of 100 MW nstalled capacty. Fnally, the last case had a penetraton of 31.1% (1800 MW) wth farms havng a capacty of 150 MW each. The New England 39 bus system was adopted as the test network for ths study wth 12 DFIG based wnd farms nstalled across the network. Each ndvdual wnd farm s capable of delverng 150 MW actve power output wth an assumed (49.2 MVAr) ±0.328 pu reactve power capablty across ts operatng range. III. VOLTAGE SECURITY ASSESSMENT The voltage securty assessment was conducted usng the DSATools smulaton package [11]. In order to dentfy the crtcal wnd farm locatons from a voltage stablty perspectve a small-sgnal analyss was conducted for the New England 39 bus system. However, DFIGs are mechancally decoupled from the power system; therefore a new methodology was developed n order to carry out the small-sgnal stablty study. In ths methodology wnd farms are ntally modelled as conventonal synchronous generators

5 3 wth exctaton systems, governors, and stablzers, dentfed as synchronous wnd farms. Followng ths, a small-sgnal analyss s completed and the partcpaton factors of the synchronous wnd farms are dentfed for varyng system condtons. The analysed system scenaros are lsted n Table I. Table I: System Scenaros for Small-Sgnal Analyss Scenaro A B C D Wnd 2400 MW 2400 MW 1200 MW 1200 MW Generaton Load 5811 MW 6811 MW 5811 MW Each of the orgnal ten synchronous unts n the New England 39 bust test system was ndvdually dsplaced n order to represent dfferent ntal dspatch condtons. The modes of the nne orgnal unts that remaned onlne were examned and the partcpaton factors that belonged to the synchronous wnd farms were recorded. Over the four scenaros, the partcpaton factors of the synchronous wnd farms were averaged and those farms that dsplayed consstently hgh partcpaton factors were deemed to be crtcal. The average partcpaton factors are shown n Fg. 2. Partcpaton factors Crtcal Wnd Farms Scenaro A Scenaro C Scenaro B Scenaro D 6811 MW Bus No: Fg. 2. Recorded partcpaton factors for wnd farms Accordng to Fg. 2 wnd farms nstalled at buses 2, 24, 26, 32, 33 and 37 ndcate hgh partcpaton factors (.e. partcpaton factors above 0.1) despte dfferent wnd and load condtons; and hence they were regarded as crtcal wnd farms for stablty enhancement. Therefore, they should be operated n a voltage control mode for voltage stablty enhancement. In order to valdate the above outcome the synchronous generators at the wnd farm locatons were replaced wth DFIGs. Those farms that were dentfed as crtcal had termnal voltage control mplemented, whle the remanng farms were operated usng a fxed power factor control scheme at 0.95 capactve. Ths control confguraton was called the crtcal control case. To confrm the effectveness of the crtcal control case, a transent smulaton was also completed. The transent smulaton was carred out for Scenaro A. Here, the crtcal control case was compared to a scenaro where all of the wnd farms were modelled as DFIGs wth full termnal voltage control enabled, known as the full control case. The rotor angle traces of the generator at bus 34 followng the loss of the generator at bus 32 s gven n Fg. 3. Fg. 3. Rotor angle traces for the generator at bus 34 followng the loss of the generator at bus 32. From Fg. 3 t can be seen that the two control strateges llustrate the same behavour followng the dsturbance. To further nvestgate the two control condtons Prony analyss was carred out and Table II llustrates the results. Table II: Prony Analyss Results for both Control Cases Crtcal Control Case Magntude Phase Frequency (Hz) Dampng (%) Full Control Case Accordng to Table II t s furtherr evdent that both control condtons llustrate the same characterstcs, and thus t confrms the valdty of the methodology used for voltage securty assessment. IV. COMPARATIVE ANALYSIS BETWEEN CONVENTIONAL OPF AND VOLTAGE CONSTRAINED OPF A. Optmzaton Strateges Two optmzaton strateges were developed n order to nvestgate the nfluence of voltage securty constrants on system loss mnmzaton. Therefore, followng two strateges have been analyzed: Strategy 1 Reactve power was optmzed at twelve wnd farms wth the objectve of mnmzng the system losses. The bus voltages were mantaned between 0.95 pu to 1.05 pu, wth an assumed ±0.328 pu reactve power capablty for each DFIG wnd farm durng optmzaton. Strategy 2: Optmzaton was carred out to mnmse system losses wth voltage constrants appled at certan wnd farms to mantan voltage stablty n the network based on the conclusons drawn from the voltage stablty study. Therefore, the bus voltages for the wnd farms sted at buses 2, 24, 26, 32, 33, and 37 were mantaned at a constant value durng the optmzaton. The remanng wnd farms were optmsed

6 4 based on the reactve power capablty and voltage lmts outlned n strategy 1. B. Comparson of Losses It s assumed that all the wnd farms experence the same wnd condtons; and hence generate the same actve power output. The system load was assumed to be constant throughout the smulaton. The varaton n system losses wth wnd farm actve power output for both scenaros s llustrated n Fg. 4. System losses (MW) Actve power output (pu) Fg. 4: Actve power loss varaton wth wnd farm power output Accordng to Fg. 4, when wnd farms are operated n voltage control mode the actve power losses have ncreased from 13% to 31% when the wnd farms power output ncreases from 0 pu to 1 pu. Therefore, strategy 2 ndcates a sgnfcant ncrease n actve power losses n the network. The trend n system actve power loss reducton remans the same for both strateges despte dfferences n ther mnmum actve power losses n the system. The average reactve power dspatch as a percentage of total reactve power capablty can be expressed as follows: Reactve Power Contrbuton = Q Q The reactve power contrbuton from the wnd farms for strategy 1 was calculated accordng to (7) and llustrated n Fg. 5. Reactve power contrbuton B2 B6 B8 B12 B16 B18 B24 B26 B32 B33 B35 B37 Wnd Farm Fg. 5: Reactve power contrbuton of wnd farms for optmzaton strategy 1 ds cap Strategy 1 Strategy 2 (7) Accordng to Fg. 5, wnd farms sted at buses 2, 6, 8, 12, 16, 18 and 33 sgnfcantly contrbute to the system loss mnmzaton. The reactve power contrbuton from each wnd farm for optmzaton strategy 2 s llustrated n Fg. 6. Reactve power contrbuton B2 B6 B8 B12 B16 B18 B24 B26 B32 B33 B35 B37 Wnd Farm Fg. 6: Reactve power contrbuton of wnd farms for optmzaton strategy 2 Accordng to Fg. 6, most wnd farms ncrease ther reactve power output to mantan the bus voltages at a constant value. Ths has resulted n an ncrease n actve power losses n the network. C. Comparson of Reactve Power Contrbuton from Wnd Farms Accordng to the optmzaton results two dstnct wnd farm groups can be dentfed for both voltage stablty mprovement and loss mnmzaton. Therefore, a sgnfcant dfference n reactve power output from the wnd farms can be observed between both strateges. In partcular, the system losses have ncreased for the voltage control strategy due to reactve power njecton / absorpton by wnd farms to control the termnal voltage. The wnd farm reactve power dspatch dfference for both strateges as a fracton of the reactve power capablty of the DFIG s llustrated n Fg Reactve power devaton B2 B6 B8 B12 B16 B18 B24 B26 B32 B33 B35 B37 Fg. 7: Reactve power dspatch dfference between two optmzaton strateges It can be seen that wnd farms sted at buses 2, 12, and 37 show an nsgnfcant dfference n reactve power dspatch between both strateges whle wnd farms sted at buses 18,

7 5 26, 32, 6, 16, 24 and 33 ndcate a reactve power dspatch dfference exceedng 20% of the total reactve power capablty. Therefore, n addton to voltage control wnd farms some other wnd farms also show sgnfcant devaton n ther reactve power output. V. SELECTIVE REACTIVE POWER OPTIMIZATION Selectve optmzaton was carred out n order to determne the most crtcal wnd farms for loss mnmzaton. Ths has been analyzed under two scenaros, consderng wnd farms for actve power loss mnmzaton and voltage securty enhancement. A. Selectve Reactve Power Optmzaton In selectve optmsaton, wnd farms were ranked accordng to ther reactve power contrbuton towards the loss optmzaton (consderng strategy 1 Fg. 5), applyng voltage control for those wnd farms contrbutng most towards loss optmzaton whle operatng the remanng wnd farms at a fxed power factor. The bus voltages are allowed to vary between 0.95 pu to 1.05 pu. The selectve optmzaton scenaros outlned n Table III were carred out for system loss optmzaton. Table III: Wnd farms for selectve optmzaton Optmzaton Strategy Wnd Farms Selectve optmzaton 6 WFs 2, 18, 12, 8, 16, 33 Selectve optmzaton 4 WFs 18, 2, 8, 33 Selectve optmzaton 2 WFs 2, 8 System losses were analyzed by consderng the selectve optmzaton scenaros n Table III wth the same load demand n the network. Fg. 8 llustrates the system losses for selectve optmzaton under dfferent wnd farm loadng condtons. System losses (MW) Strategy 1 Selectve Optmzaton wth 6 WF Selectve Optmzaton wth 4 WFs Selectve Optmzaton wth 2 WFs Wnd Farm Output (pu) Fg. 8: System losses wth selectve optmzaton. Accordng to Fg. 8, t can be seen that near maxmum beneft can be acheved by optmzng the reactve power only at wnd farms wth a sgnfcant contrbuton towards system loss reducton. As an example, when only two wnd farms were optmzed for loss reducton the system losses ncreased by 0.07 MW whch represents only a 0.23% ncrease compared to strategy 1. Therefore, selectve optmzaton can reduce the reactve power dspatch burden for the system operator whle mantanng the system losses close to the optmal levels. B. Selectve Voltage Control In ths secton, reactve power was optmzed whle operatng only one wnd farm n the voltage control mode. Therefore, each ndvdual wnd farm n strategy 2 (crtcal wnd farms dentfed n Fg. 2) was operated n voltage control mode durng optmzaton to analyze the contrbuton of each wnd farm to system losses. Fg. 9 llustrates the average system loss ncrease from optmal (.e. from actve power losses ndcated n OPF strategy 1) when voltage control was mplemented at dfferent wnd farm locatons. System losses (MW) Wnd Farm Fg. 9: System loss varaton wth voltage control at each wnd farm locaton Accordng to Fg. 9, the maxmum devaton occurs when voltage control s mplemented at the wnd farm at bus 2. However, when the wnd farm at bus 32 s operated n voltage control mode t has the least mpact on network losses. The wnd farms were then ranked based on ther mpact towards system optmal losses (.e. strategy 1 losses) and then voltage control was mplemented selectvely at the wnd farm locatons. Fg. 10 llustrates the system losses wth selectve optmzaton. System losses (MW) Actve Power: 0 pu Actve power: 0.67 pu Actve power: 0.33 pu Actve power: 1 pu 0 32/33 32/33/24 32/33/24/26 32/33/24/26/2 Voltage controlled wnd farms Fg. 10: System losses wth selectve voltage control

8 6 It can be seen that system losses ncrease when more and more wnd farms are operated n voltage control mode. As an example, at 0 pu actve power output, when only two wnd farms (e.g. 32 and 33) are operated n voltage control mode the system losses ncrease by 0.4 MW compared to strategy 1, however when fve wnd farms (32, 33, 24, 26 and 2) are operated n voltage control mode the system losses have ncreased by 4 MW compared to strategy 1. VI. DISCUSSION Tradtonally, wnd farms are not requred to provde ancllary servces for network support; hence they do not partcpate n ether voltage control or reactve power provson. However, wth the ncreased penetraton levels of wnd generaton, voltage control and reactve power support have been dentfed as two necessary requrements for future wnd farms. Ths study has shown that two dstnct wnd farm categores exst based on ther contrbuton towards voltage stablty enhancement and system actve power loss mnmzaton. However, certan wnd farms have been dentfed crtcal for both voltage enhancement and system loss mnmzaton. Therefore, the system operator must determne the tradeoff between two wnd farm control strateges based on the network requrements. VII. CONCLUSIONS Ths study has nvestgated two reactve power optmzaton strateges ncorporatng wnd generaton. It has llustrated that actve power losses ncrease when wnd farms are operated n voltage control mode compared to the optmzed reactve power dspatch strategy mplemented at all wnd farms. The study has further shown that by optmzng the reactve power dspatch only at the most crtcal (n terms of loss reducton) wnd farm locatons results n only a margnal ncrease n actve power losses and t s sgnfcantly less than the losses ncurred durng the voltage control mode (strategy 2). Reducng the number of wnd farms operatng n voltage control mode can reduce the system losses. In partcular, by rankng the voltage controlled wnd farms based on ther ndvdual contrbuton towards the actve power loss ncrease and by selectve operaton of wnd farms n voltage control mode the system losses can be notceably reduced. Therefore, the tradeoff between voltage stablty mprovement and system loss reducton can be acheved by selectve operaton of wnd farms based on ther contrbuton towards loss reducton and voltage stablty enhancement. Further studes are requred to analyze the system dynamc performance (.e. voltage stablty and losses) under varable load and wnd condtons n order to determne the crtcal wnd farms for dfferent system condtons. VIII. REFERENCES [1] L. Meegahapola, S. Duraraj, D. Flynn, and B. Fox, Coordnated utlsaton of wnd farm reactve power capablty for system loss optmsaton, European Transactons of Electrcal Power, Vol. 21(1), 2011, pp [2] R. J. Konopnsk, P. Vayan, V. Ajjarapu, Extended reactve capablty of DFIG wnd parks for enhanced system performance, IEEE Trans. Power Syst., vol. 24, no. 3, pp , Aug [3] N. Mo, Z.Y. Zou, K.W. Chan and T.Y.G. Pong, Transent stablty constraned optmal power flow usng partcle swarm optmsaton, IET Gener. Transm. Dstrb., 2007, vol. 1, no. 3, pp [4] R. J. Avalos, C. A. Canzares, and M. F. Anjos, A practcal voltagestablty-constraned optmal power flow, IEEE PES GM 2008, Pttsburgh, USA, Jul [5] X. Zhang, R. W. Dunn, and F. L, Stablty constraned optmal power flow n a practcal balancng market, IEEE PES GM 2007, Florda, USA, Jul [6] P. Duttal, and A. K. Snha, Voltage stablty constraned multobjectve optmal power flow usng partcle swarm optmzaton, ICIIS 2006, Sr Lanka, Aug [7] Y. Xu, Z. Y. Dong, K. Meng, J. H. Zhao, and Kt Po Wong, A hybrd method for transent stablty-constraned optmal power flow computaton, IEEE Trans. Power Syst. (n press) [8] W. Rosehart, C. Cazares, V. Quntana, Optmal power flow ncorporatng voltage collapse constrants, PES Sumer Meetng 1999, Edmonton, Canada, Jul [9] E. Vttal, M. O Malley, and A. Keane, A steady-state voltage stablty analyss of power systems wth hgh penetratons of wnd, IEEE Trans. Power Syst., vol. 25, no. 1, pp , Feb [10] Power Factory manual, DIgSILENT Power Factory Verson 13.2, GmbH, Germany, [11] DSATools User Manual, Powertech Labs Inc., Surrey, Brtsh Columba, Canada. IX. BIOGRAPHIES Lasantha Meegahapola s a lecturer n power engneerng at Unversty of Wollongong, Australa. He receved a BSc. Eng. degree (Frst Class) from the Unversty of Moratuwa, Sr Lanka n 2006, and a PhD from The Queen's Unversty of Belfast, UK n Hs research nterests nclude renewable power generaton, power system stablty, and ntellgent approaches n power systems. Eknath Vttal receved hs B.E. from the Unversty of Illnos Urbana- Champagn and hs M.S. from Iowa State Unversty n Electrcal Engneerng n 2005 and 2007 respectvely. He completed hs Ph.D. at Unversty College Dubln n He s currently a senor research engneer at Unversty College Dubln wth research nterests n power systems operaton, plannng and stablty. Daman Flynn s a senor lecturer n power system operaton & control n the School of Electrcal, Electronc & Communcatons Engneerng, Unversty College Dubln. Hs research nterests nvolve the ntegraton of renewables nto power systems, and advanced modelng and control technques appled to power plant. Andrew Keane receved B.E. and Ph.D. degrees n Electrcal Engneerng from Unversty College Dubln n 2003 and 2007 respectvely. He s a lecturer wth the School of Electrcal, Electronc & Communcatons Engneerng, and Unversty College Dubln wth research nterests n power systems plannng and operaton, dstrbuted energy resources and dstrbuton networks.