EMTP Model for Analysis of Distributed Generation Impact on Voltage Sags

Transcription

1 EMTP Model or Analyi o Ditributed Generation Impact on oltae Sa Juan A. Martinez-elaco, Jacinto Martin-Arnedo Abtract Small eneration unit connected to ditribution ytem can improve end-uer power quality; e.. by providin voltae upport. However, the preence o ditributed eneration (DG) chane the radial nature o ditribution ytem and aect the perormance o the protection ytem. DG can alo aect durin- and pot-ault voltae; in addition, voltae a are hihly inluenced by the type o protective device and the coordination between them. Thi paper i aimed at explorin the impact that DG can have on voltae a characteritic at ditribution level when the main caue o voltae a can be either at a tranmiion or a ditribution network. Keyword: Ditributed Generation, Power Quality, oltae Sa, Modelin, Simulation. I. INTRODUCTION T i commonly accepted that the intallation o mall I eneration unit at ditribution level ha many advantae: economical (enery eiciency), environmental (reduction o aeou emiion), technical (voltae upport) and even political (competition) [] - [3]. Althouh cot aociated with DG technoloie are till hih, they can be a olution or thoe ituation in which hih power upply reliability i needed, or when the contruction o tranmiion line and lare power plant i not upported by end-uer. However, the intallation o DG ource raie new challene, e.. the radial topoloy o ditribution network doe not chane, but the power will no loner low in a inle direction. DG ilandin i one o the main concern [4] - [7], and althouh ilanded operation i not enerally allowed, DG may ucceully operate in iland i there i a balance between load and eneration. Characteritic o voltae a in a ditribution network that are caued at the tranmiion level can be aected by the preence o DG. Characteritic o voltae a caued rom inide the ditribution level can be aected by the preence o DG a well a by the placement o protective device and the coordination between them. Thi paper explore the impact that DG can have on the characteritic o voltae a in a Thi work i upported by the Spanih Miniterio de Invetiación y Ciencia, Reerence ENE /CON. Juan A. Martinez-elaco i with the Departament d Eninyeria Elèctrica, Univeritat Politècnica de Catalunya, Barcelona, Spain ( o correpondin author: jamv@ieee.or). Jacinto Martin-Arnedo i with ITC-, Barcelona, Spain. ( jacinto.martin@via.itc.com). Preented at the International Conerence on Power Sytem Tranient (IPST 07) in Lyon, France on June 4-7, 007 ditribution network with a hih penetration o embedded eneration, aumin that the voltae a caue can be either at the tranmiion or the ditribution level. An EMTP model o a mall ditribution network, includin protective device, ha been created to analyze the DG impact on the characteritic o voltae a. The tudy ha been carried out uin the ATP (Alternative Tranient Proram) and the library o module developed by the author or repreentin component o ditribution network. The main eature o the tet ytem are detailed in Section II. Pre-ault teady-tate condition are preented in Section III. The main part o thi work are Section I and, where reult derived rom the imulation o the tet ytem are hown and analyzed. A impliied analyi o the retained voltae at enitive equipment node durin ymmetrical ault i preented in Section I. Main concluion and uture work are ummarized in the lat ection. II. TEST SYSTEM Fi. how the diaram o the tet ytem ued in thi paper. The ubtation tranormer i rounded at the lower voltae ide by mean o a zi-za reactor, which limit the current caued by a inle-line-to-round ault to 800 A. Fi. how the time-current curve o the protective device intalled in the ytem. Modelin uideline ued to repreent the tet ytem were dicued in [8]. Ratin and electrical parameter o ynchronou enerator are hown in Table I. The block diaram o the excitation control i preented in Fi. 3. A inle-ma model repreentation will be ued or repreentation o the mechanical ytem o ynchronou enerator. The eect o primer mover will be nelected and a contant mechanical enery input will be aumed in imulation. Althouh the eect o the mechanical parameter will not be very important on voltae a characteritic, thee parameter cannot be nelected when analyzin the tranient perormance o the ytem [9]. Since a wide rane o mechanical parameter value will be ued in imulation, they are not provided. Note that all eneration unit have the ame p.u. value, no electronic interace ha been aumed or any eneration unit, the maximum load (that can be upplied rom ditribution tranormer) exceed the ubtation tranormer rated power, ditributed enerator can upply more than 50% o the maximum load and repreent 65% o the ubtation ratin. Only L load are aumed. In addition, all the tudie preented and analyzed in thi paper have been carried out with a contant impedance model.

2 DG GP 3 MA GP DG 6.0 km 5.0 km BR GP DG3 6 A B 8.0 km 7.0 km 6.0 km BR 5.5 km 0 GP DG6 7.0 km 5.0 km 4.5 km 5.0 km 5.0 km.5 km km 3 MA GP DG4 C 4.5 km 3.5 km km km 5.5 km GP DG5.5 km km 7 7 H equivalent: 0 k, 500 MA, X/R 0 Subtation tranormer: 0/5 k, 0 MA, 8%, Yd, X/R 0 Ditribution tranormer: 5/ k, 6%, Dy, X/R 0 DG tranormer: 5/6k, 8%, Yd, X/R 0 Line: / j0.39, j.56 Ω/km BR Circuit Breaker Fue GP Generator protection Fi.. Diaram o the tet ytem. 0 re 3.5 Time [] 0. Fue Breaker I cmax G E FD Fi. 3. Block diaram o the excitation control Current [A] Fi.. Time-current characteritic o protective device. TABLE I ELECTRICAL SPECIFICATION OF THE SYNCHRONOUS GENERATOR Parameter alue Rated requency 50 Hz Rated voltae 6.0 k Number o pole 4 Armature reitance: R a 04 pu Armature leakae reactance: X l 0 pu d-axi ynchronou reactance: X d.7000 pu d-axi tranient reactance: X d pu d-axi ub-tranient reactance: X d 47 pu d-axi open-circuit tranient time contant: T do d-axi open-circuit ub-tranient time contant: T do 87 ero-equence reactance: X pu III. STEADY STATE OPERATING CONDITIONS Althouh dierent teady tate operatin condition (prior to any voltae a) were analyzed, only two cenario are reported in thi paper. Table II and III how the voltae that reult with minimum and maximum load at the L load node and the ynchronou enerator terminal, repectively. The active power raction upplied by DG unit with repect to the active power upplied rom the ubtation i 9% with the maximum load and 4% with the minimum load. Note that even with DG the voltae drop at ome L node can reach up to a 0% o the rated voltae. In actual ditribution network thi can be mitiated by reulatin the ubtation tranormer and intallin voltae reulator. The voltae reulation rom the ubtation tranormer can be repreented by chanin the actual voltae upplied rom the H network equivalent. A model o the voltae reulator ha been neither developed nor applied in thi work.

3 TABLE II STEADY STATE CONDITIONS LOAD NODES Maximum load Minimum load Load Load oltae Load oltae node (MA, p) (pu) (MA, p) (pu).85, , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , () oltae value correpond to the L ide o ditribution tranormer and were obtained on a 30 bai (rm phae-to-round). () Power value correpond to a rated voltae o 400, rm phae-to-phae. (3) The power actor (p) in all load i lain. TABLE III STEADY STATE CONDITIONS SYNCHRONOUS GENERATORS Maximum load Minimum load Unit P/Q (MW/MA) oltae (pu) P/Q (MW/MA) oltae (pu).356, , , , , , , , , , , , 6.0 (PU voltae value correpond to enerator terminal and were obtained on a 6 k bai). I. OLTAGE SAGS CAUD AT THE TRANSMISSION LEEL A hih percentae o voltae a experienced at L load node are caued at the tranmiion level. However, the characteritic o voltae a (manitude, duration) meaured at L node can be very dierent rom the characteritic meaured at the H ide o the ubtation tranormer. The connection o ditribution and ubtation tranormer can iniicantly alter thee characteritic, except or three-phae voltae a. In addition, the preence o DG can alo mitiate voltae drop at ditribution node. The imulation o the eect that voltae a caued at the tranmiion network can have on load node located downtream the lowet voltae ide o the ubtation tranormer can be perormed without includin any protective device except thoe aimed at protectin ditributed eneration unit aaint abnormal voltae, and perhap unbalance loadin. Accordin to IEEE Std 547 [5], the clearin time (i.e. the time between the tart o an abnormal condition and the DG ceain to enerize the local electric power ytem) hould be baed on the (durin-ault) voltae rane, ee Table I. The tandard tate that the protection ytem hall detect the rm value o each phae-to-phae voltae, except or a rounded wye-wye tranormer connection, in which cae phae-toneutral voltae hould be detected. TABLE I INTERCONNECTION SYSTEM RESPON TO ABNORMAL OLTAGES [5] oltae rane Clearin time () () (% o bae voltae) < 50 () > 0 () Note: () Deault clearin time or DG > 30 kw () For 60Hz ytem. An apect to be conidered i the expected duration o voltae a caued by ault located at the tranmiion network. Thi value can depend on everal apect (e.. ault location or tability marin), but in eneral it will be rather hort. One can aume that very rarely it will lat or more than 0 cycle when the caue i a three-phae ault. An ATP module wa developed or imulation o voltae a uptream the ubtation tranormer. The capabilitie o thi module allow uer to peciy voltae a characteritic (retained voltae, duration, phae-anle jump, initiation o a) independently or each phae. Note that the hihet voltae at the ubtation tranormer i a ubtranmiion voltae; one hould, thereore, aume that the module i repreentin a caued at both tranmiion and ubtranmiion level. Fi. 4 depict imulation reult obtained with maximum load. Thee reult were derived without activation o DG protective device. A expected, the pu manitude o voltae a at enerator terminal and ditribution node can be very dierent rom the pu manitude meaured at the H ubtation tranormer. Subtation H Subtation M Load Node 9 DG Fi. 4. Sinle-phae voltae a caued at the tranmiion network.

4 When analyzin thee and other imulation reult, everal apect have to be conidered: It i evident the eect o the connection o ubtation, interconnect DG and ditribution tranormer. The reactive power low meaured at the lowet voltae ide o the ubtation decreae iniicantly durin a voltae a, and it can even revere the direction, lowin rom DG unit to the tranmiion network. Thi will obviouly aect the voltae drop between the ubtation terminal and the node located downtream. A third apect that can aect a manitude are the ratio between the dierent impedance involved in a voltae drop and the ratio between the rated power o the ubtation tranormer and DG unit, ee Section I. The mot important concluion rom thi tudy can be ummarized a ollow: Durin ymmetrical a, DG unit will reduce the voltae drop, but they will not avoid enitive equipment trip except with hallow a (ee Section I). On the other hand, DG protection can operate and eparate unit rom the ytem except or very hort ault duration. Durin non-ymmetrical a, the voltae drop at DG terminal and L load node can be iniicantly reduced due to the dierent tranormer connection rather than to the preence o DG unit. The inluence o the tranormer connection can alo play an important role when the voltae a i caued with a phaeanle jump. The eect o a lo o power upply rom the tranmiion level ha not been analyzed, althouh thi tudy i omehow covered in the next ection by thoe cae in which a ault condition caue a eeder breaker openin.. OLTAGE SAGS CAUD AT THE DISTRIBUTION LEEL The characteritic o voltae a caued by ault oriinated within the ditribution network depend on everal actor: ault characteritic (type, duration, reitance); ditribution tranormer connection; ubtation roundin; ratin o DG unit; dein o the dierent protection ytem; operatin condition and DG penetration level; ault and monitor location. Althouh voltae a can be either ymmetrical or aymmetrical, and everal parameter can be needed or a ull characterization, only the retained voltae and the duration are analyzed in thi work. Retained voltae at L node can be dierent rom thoe at the M node. In act, with the connection ued in thi work or ditribution tranormer, there will not be well at the L ide [8]. Simulation have been carried out aumin that a circuit breaker open the three pole, irrepectively o the hortcircuit type, while ue are o current-limitin type and open only the aulted phae. A or the coordination between thee protective device two poibilitie are analyzed: ue avin (ue are lower than breaker) and ue blowin (timecurrent characteritic are thoe depicted in Fi. ). A conequence derived rom the protective device operation i that voltae a will not be alway rectanular, ince the coordination between protective device can produce multiple event with dierent retained voltae. Characterization o multiple event (manitude and duration) become then an important iue [0]. The protection model o a mall ynchronou enerator i much more complex than the protection o a ditribution network. In addition, not only the enerator protection but alo the interconnect protection ha to be taken into account. Interconnect protection atiie the utility requirement to allow the connection o the enerator to the rid, while enerator protection provide detection o internal hortcircuit and abnormal operatin condition [6], [7], []. In the preent tudy, the protection model o mall enerator could include device aaint overcurrent, abnormal voltae, unbalance loadin and revere power. DG unit can aect voltae a manitude beore and ater the breaker o the aulted eeder ha opened. Since the impact will be maller with minimum load, the tudy will bein by aumin thi cenario. I the DG impact i not very iniicant with minimum load, it will be even le important under other operatin condition. Due to room limitation, imulation reult analyzed in thi ection correpond only to the le and the mot evere voltae a; i.e. to voltae a caued repectively by LG and 3L ault. And only three ault location (node A, B and C in Fi. ) are conidered. The ollowin ubection analyze voltae a caued repectively with ue avin and ue blowin. A. Fue avin The network i protected only by breaker, and ue model are not included in imulation. An important apect o the tudy i the clearin time a a unction o the type and location o the ault. Table how the time required by the breaker o the bottom eeder to open. Note that, with the time-current curve elected or breaker, there are ault poition or which the clearin time can be much loner than econd. In thoe cae, mechanical tranient mut be careully analyzed and the ytem model mut be improved (e.. DG prime mover hould be included). TABLE CLEARING TIMES (FAULT RESISTANCE 0) Fault type Node A Node B Node C LG 790 m 0 m 650 m 3L 70 m 400 m 780 m Fi. 5 and 6 how ome plot o voltae a caued by LG and 3L ault. Table I ummarize the main reult, or which a zero ault reitance wa aumed. Since non-zero reitance value are very uual, one hould expect dierent clearin time and dierent retained voltae.

5 TABLE I OLTAGE SAG CHARACTERISTICS Fault Upper (Unaulted) Feeder Lower (Faulted) Feeder LG 3L oltae at L node and DG terminal will never be below 90% o the rated voltae. Only when the ault poition i cloe to the ubtation (e.. node A), voltae at DG terminal can reach value a low a 65% o the rated voltae. At L load node the trend i imilar, but voltae can be even lower. Power low can revere in all DG unit. The durin-ault voltae at ome phae o L node cloe to the ault poition can be a low a 80% o the rated voltae. The voltae at DG terminal will never be below 88% o the rated voltae. When the ault poition i cloe to the ubtation (e.. node A), the durinault voltae at DG terminal can reach value a low a 0% o the rated voltae. oltae decreae ater the breaker open; the voltae reduction will depend on the load demand. Plot o Fi. 5 how voltae a o dierent duration and caued by ault located at node B and C, ee Fi.. The ault at node C will not caue breaker openin ince the clearin time or thi cae i much loner that the imulated cae, ee Table. oltae in all imulated cae reach the pre-ault value ater the eeder breaker open; that i, DG operate ucceully ater load are eparated rom the ubtation. Thee reult prove that the impact o a LG ault on the DG and L node at the unaulted eeder i not iniicant and very ew trip hould be expected or enitive equipment upplied rom L node located at thi eeder. For ome L node located at the aulted eeder, the retained voltae can drop up to the 80% o the rated voltae, but DG terminal voltae will be hardly below 90% o the rated voltae. With maximum load and ater breaker openin, voltae at DG terminal and L node can reach value below 90% o the rated voltae; the drop can be particularly important at ome end-uer node where the retained voltae can be below 80%. However, the impact at DG and L node o the unaulted eeder i almot neliible, althouh at ome phae o L node can be alo below 90%. In all cae, thi i due not only to the ault but to the pre-ault voltae, ee Table II. Remember that pre-, durin- and pot-ault voltae can be reulated to a hiher value by uin voltae reulator and voltae reulation at the ubtation; o in many intance, the eect o a ault can be eaily mitiated. Sa caued by 3L ault at the unaulted eeder can be more evere than thoe caued by LG ault, but (a expected) even with thi type o ault retained voltae above 90% o the rated voltae can appear when the ault location i ar enouh rom the ubtation (e.. at Load Node 5 DG node C), ee Section I. Plot o Fi. 6 how cae in which the aulted eeder breaker open alway; in the econd cae the voltae drop ater breaker openin and recover ater the ault clear. Since thee reult were obtained with minimum load, one hould expect wore voltae a perormance with any other operatin condition a) Fault location Node B Fault reitance 0 Ω (Duration.0) Load Node DG b) Fault location Node C Fault reitance 0 Ω (Duration.0) Fi. 5. oltae a caued by a LG ault Fue avin. Load Node DG a) Fault location Node A Fault reitance 0 Ω (Duration ) Load Node 7 DG b) Fault location Node C Fault reitance 0 Ω (Duration ) Fi. 6. oltae a caued by a 3L ault Fue avin.

6 B. Fue blowin Fi. 7 how ome imulation reult when ue are intalled and their current-time curve are a hown in Fi.. Thee reult correpond to a ault located at node C (ee Fi. ). Since thi i the urthet node rom ubtation terminal, the hort-circuit current i the mallet one while the meltin time i the lonet one. From thee reult one can conclude that in eneral only enitive equipment at node downtream the ault location will trip. A hown in Fi. 7a, only the voltae at one phae o the L node downtream the ue location will remain with the pre-ault value, while the impact on the cloet node will not be iniicant. The cae depicted in Fi. 7b how that, even or a three-phae ault, only enitive equipment on L node downtream the ue i aected. The impact on the unaulted eeder node i neliible. The perormance with a non-zero ault reitance will be imilar althouh the meltin time o ue will be loner and the impact on enitive equipment located at the aulted eeder hould not be alway nelected. Thee concluion are obviouly valid only when the ault location i in the zone protected by a ue; otherwie both reult and concluion would be thoe analyzed above. C. Dicuion Simulation reult preented and analyzed in the previou ubection were obtained by aumin that device intalled to protect DG unit and the interconnect tranormer did not operate. Thi could be the ituation when the ault i downtream a ue, but it i not alway valid i the ault location i outide a ue-protected zone. In thoe cae, DG protective device could operate accordin value hown in Table I. The main concluion rom all imulation reult can be ummarized a ollow: When ue blowin i allowed and the ault i located inide the zone protected by a ue, one can aume by deault that only enitive equipment at node downtream the ue will trip. Thi will not be alway true and depend on the ault reitance, the ault and equipment location, a well a on the voltae tolerance o enitive equipment. When the ault i located outide the zone protected by a ue or ue are aved, the eect on enitive equipment will depend on the type o ault and the repone o the DG protection ytem. Only when a ymmetrical ault i ar rom the ubtation terminal, enitive equipment at L node o the unaulted eeder will not trip. However, the durin-ault voltae at node on the aulted eeder will be uually below an acceptable value rane ater eeder breaker openin, even i DG unit are not diconnected. Senitive equipment at L node o the unaulted eeder will withtand mot voltae a caued by LG ault. Thi could be alo true or a iniicant percentae o L node o the aulted eeder i DG unit are not diconnected, ince durin-ault voltae and voltae ater DG - 4 Load Node 3 Load Node a) oltae a caued by a inle-phae-to-round ault DG - 4 Load Node 3 Load Node b) oltae a caued by a three-phae ault Fi. 7. Fue blowin. Fault location Node C, Fault reitance 0 Ω, Duration.0. eeder breaker openin will be above 90% o the rated voltae. I protection o DG unit operate (e.. due to unbalance loadin), then thi concluion will apply in eneral only to ault with a duration horter than the clearin time. I. OLTAGE SAG CALCULATION The inluence that the parameter o the main ource (network equivalent) and the mall ynchronou enerator have on the retained voltae caued by a ymmetrical ault will be analyzed. The tudy will be made by nelectin the eect o the pre-ault current. Table II how the variou topoloie that have to be conidered or the ytem under tudy. Since more than one ynchronou enerator i connected to the tet ytem, parameter involved in expreion hould be een a the parameter o the equivalent een rom the ault location. Thee expreion can be alo applied when no DG unit i connected to the tet ytem (ee Table III). oltae and impedance in the expreion hown in the table are complex quantitie (phaor), o even nelectin pre-ault current the analyi i not eay. Note that the ratio R/X i very mall in the main ource and the tranormer o the tet ytem, but it i reater than unity or line. In order to acilitate the analyi, it i aumed that main ource and DG voltae manitude are pu and their arument are the ame. Table III how the expreion that reult rom thee aumption with and without DG [], [3], [4].

7 TABLE II DURING-FAULT OLTAGES SYSTEM CONFIGURATION S DG OLTAGE SAG MAGNITUDE A The ault i at the tranmiion level G ( ) G T PCC impedance een rom the M terminal o the ubtation impedance between the ubtation and the PCC impedance between the DG unit and the PCC impedance een rom the DG tranormer terminal. S DG B The ault i at the other eeder G ) i i T ( )( i ) ( )( i i PCC impedance een rom the M terminal o the ubtation tranormer impedance een rom the DG tranormer terminal i impedance between the ubtation and the DG tranormer impedance between the ubtation and the ault location. C The ault i between the ubtation and one DG unit, at the DG unit ide S DG T PCC 3 impedance een rom the M terminal o the ubtation impedance een rom the M terminal o the DG tranormer impedance between the ubtation and the PCC impedance between the ault location and the PCC 3 impedance between the ault location and the DG tranormer D The ault i between the ubtation and one DG unit, at the ubtation ide S DG G 3 impedance een rom the M terminal o the ubtation impedance een rom the M terminal o the DG tranormer PCC 3 impedance between the ubtation and the ault location impedance between the PCC and the ault location 3 impedance between the PCC and the DG tranormer T voltae at the tranmiion level; G DG internal voltae; voltae at the enitive equipment location A urther impliication can be made by aumin that all impedance have the ame X/R ratio or all reitance are nelected. The ollowin pararaph dicu the reult derived rom thi new aumption and preent the condition to be ulilled in order to avoid the trip o enitive equipment. A. The ault location i at the tranmiion level The expreion o thi cae can be rewritten a ollow T x () x To obtain a retained voltae at the PCC equal or reater than 90% o the rated voltae, it mut be (x) 0 T. Since the value o the ratio / i maller than 0. in all ynchronou enerator, thi a manitude can be only obtained when the voltae drop at the tranmiion level i below %. That i, the inluence o DG unit when the voltae a caue i at the tranmiion level i neliible. B. The ault location i at the ditribution level The expreion that correpond to the cae with DG can be rewritten a ollow x i x () /( i ) ( ) i I 0 (i.e., the ault location i very cloe to the ubtation), then it reult x. To obtain a retained voltae at the PCC equal or reater than 90% o the rated voltae, it mut be i > 9. Given the value o or the dierent DG unit and the impedance per unit lenth o the ditribution line, the retained voltae at all node o the upper eeder will never reach a 50% o the rated voltae with a three-phae ault located on the lower eeder and cloe to the ubtation, unle wa very lare.

8 Cae A B C D TABLE III SIMPLIFIED EXPRESSIONS OF DURING-FAULT OLTAGES With DG ( T ) ( G ) Without DG ( T ) ( G 0; ) T T ( i ) ( i ) ( )( ) i 0 3 T ( T ) voltae drop at the tranmiion level I >>, >> i and i o the ame order than, then. That i, voltae at unaulted eeder node between the ubtation and a DG unit that i cloe to the ubtation can be above 90% o the rated voltae i the ault location on the lower eeder i ar enouh rom the ubtation. C. The ault location i at the ditribution level and the PCC i between the ubtation and the ault location To obtain a retained voltae at the enitive equipment equal or reater than 90% o the rated voltae, it mut be > 9( ). That i, the point o common couplin mut be ar rom the ault location and cloe to the ubtation. In act, thi cenario i the ame that reult without DG. Thereore, the analyi i well known and can be ound in many reerence, ee or intance [] or [5]. D. The ault location i at the ditribution level and the PCC i between the ault location and one DG unit To obtain a retained voltae at the enitive equipment equal or reater than 90% o the rated voltae, it mut be > 9( 3 ). That i, the point o common couplin mut be ar rom the ault location and cloe to the DG unit. Since the impedance een rom the M terminal o a DG tranormer increae a the rated power o the enerator decreae, the ditance between the ault location and the enitive equipment mut increae a the rated power o the DG unit decreae. Actually thi cenario can be analyzed by uin the reult and tudie that correpond to the previou cae by exchanin parameter o the main ource and the DG unit. Note that thi can be een a an important eect o the preence o DG, ince the voltae at the point o common couplin conidered in thi cae would be zero without DG, a hown in Table III. II. CONCLUSIONS The work preented in thi paper ha proved that DG can have a poitive impact on the characteritic o voltae a caued at any voltae level. Simulation reult have hown that enitive equipment can withtand voltae a caued by LG ault when DG i preent ater the eeder breaker open. Thi i an important act, ince more than 60% o ault in mot ytem belon to thi type. But even with 3L ault, the retained voltae can be above the threhold voltae under ome circumtance, a dicued in Section I, i DG i preent. In any cae, a trict application o the protection ytem repone, accordin to IEEE Std. 547, could iniicantly reduce thi impact, except with hort-duration ault. The tudy ha been baed on the preence o mall ynchronou enerator only, without coniderin any power electronic interace. Future work hould conider the preence o other DG technoloie (wind, photovoltaic, uel-cell), whoe impact can be dierent rom that obtained in thi work, with pecial emphai on electronic-interaced DG. A more complete repreentation o DG unit (includin model o prime mover and any type o protection) will provide more accurate reult and expand the cae to be analyzed. In addition, dierent reult could be alo derived rom the implementation o dierent protection cheme (e.. includin recloer) [8], [6], and rom the application o more advanced load model [7]. III. REFERENCES [] H.Lee Willi and W.G. Scott, Ditributed Power Generation. Plannin and Evaluation, Marcel Dekker, 000. [] M. Godoy Simoe and F.A. Farret, Renewable Enery Sytem, CRC Pre, 004. [3] G.M. Mater, Renewable and Eicient Electric Power Sytem, John Wiley, 004. [4] IEEE Power Sytem Relay Committee, Impact o ditributed reource on ditribution relay protection, Auut 004. [5] IEEE Std , IEEE Standard or Interconnectin Ditributed Reource with Electric Power Sytem. [6] C.J. Mozina, Interconnect protection o dipered enerator, 00 IEEE PES T&D Con., pp [7] IEEE IAS WG Report, Application o ilandin protection or indutrial and commercial enerator, 59th Annual Con. or Protective Relay En., 006. [8] J.A. Martinez and J. Martin-Arnedo, oltae a tudie in ditribution network. Part I: Sytem modelin, IEEE Tran. on Power Delivery, vol., no. 3, pp , July 006. [9] R.S. Thallam, S. Suryanarayanan, G.T. Heydt and R. Ayyanar, Impact o interconnection o ditributed eneration on electric ditribution ytem A dynamic imulation perpective, IEEE PES General Meetin 006, June 3-7, Montreal. [0] IEEE P564, Recommended Practice or the Etablihment o oltae Sa Indice, Drat 6, January 004. [] D. Reimert, Protective Relayin or Power Generation Sytem, CRC Pre, 006 [] M.H.J. Bollen, Undertandin Power Quality Problem. oltae Sa and Interruption, IEEE Pre, 000, New York. [3] M.H.J. Bollen and M. Häer, Impact o increain penetration o ditributed eneration on the number o voltae dip experienced by end-cutomer, 8th CIRED, 6-9 June, Turin, 005. [4] J.C. Gómez and M.M. Morco, Coordination o voltae a and overcurrent protection in DG ytem, IEEE Tran. on Power Delivery, vol. 0, no., pp. 4-8, January 005. [5] M. Stephen Daniel, A 35-k ytem voltae a improvement, IEEE Tran. on Power Delivery, vol. 9, no., pp. 6-65, January 004. [6] J.A. Martinez and J. Martin-Arnedo, Modelin o protective device or voltae dip tudie in ditribution ytem, IEEE PES General Meetin 004, June 6-0, Denver. [7] J.A. Martinez and J. Martin-Arnedo, Advanced load model or voltae a tudie in ditribution network, IEEE PES General Meetin, June 004, Denver.