A novel method to identify critical Voltage Control Areas

Transcription

1 obert LIS 1, Grzegorz BŁAJSZCZAK 2 Poltechnka Wrocławska, Instytut Energoelektryk (1), PSE Operator S.A. (2) A novel method to dentfy crtcal Voltage Control Areas Streszczene. Systemy elektroenergetyczne całego śwata dośwadczane są awaram, które zwykle prowadzą do całkowtego zanku napęca tzw. blackout u. Ponższy artykuł przedstawa automatyczną metodę systemowego zapobegana utrace stablnośc napęcowej. Metoda VCA określa krytyczne obszary regulacj napęca, tzw. obszary VCAs w systeme (z uwzględnenem topolog, warunków ogranczeń przesyłu oraz oblcza wymagane źródła mocy bernej, potrzebne dla zapewnena bezpeczeństwa. Metoda przeszła pomyślne wstępne testy dla krajowych sec przesyłowych. (Nowa metoda dentyfkacj obszarów zagrożonych utratą stablnośc napęcowej). Abstract. The power system all over the world experence from tme to tme bg falure leadng to total lack of voltage over a large area - called blackout. Ths paper presents an automated method for preventng the voltage collapse. Ths method called VCA (Voltage Control Area) allows to calculate reactve power necessary to mtgate the grd and pont out the weak knots where the reactve power should be njected. Ths method was appled and tested wth Polsh transmsson network. Słowa kluczowe: stablność napęcowa systemu elektroenergetycznego, symulacja pracy, nezawodność pracy, regulacja mocy bernej. Keywords: Power system voltage stablty, Power system smulaton, Power system relablty, eactve power control. Introducton Assessng and mtgatng problems assocated wth voltage securty remans a crtcal concern for many power system planners and operators. Snce t s well understood that voltage securty s drven by the balance of reactve power n a system, t s of partcular nterest to fnd out what areas n a system may suffer reactve power defcences under some condtons. If those areas prone to voltage securty problems, often called Voltage Control Areas (VCAs), can be dentfed, then the reactve power reserve requrements for them can also be establshed to ensure system secure operaton under all condtons. A number of attempts have been made n the past to dentfy those areas, ncludng a wde range of academc research and efforts toward commercal applcatons. A bref revew of methods for determnng VCA groups s presented n [1], where the author developed the Voltage Stablty Securty Assessment and Dagnostc (VSSAD) method. The VSSAD method breaks up any power system nto non-overlappng set of coherent bus groups (VCAs), wth unque voltage stablty problems. There s a eactve eserve Basn (B) assocated wth each VCA, whch s composed of the reactve resources on generators, synchronous condensers, and other reactve power compensatng devces, such that ts exhauston results n voltage nstablty ntated n ths VCA. The VCA bus group acts lke a sngle bus and can t obtan reactve power supply at the same level of reactve power load no matter how t s dstrbuted among the buses n that group. Fndng VCAs and ther assocated B s n VSSAD method s based on QV curve analyss performed at each test VCA. It nvolves the placement of a synchronous condenser wth nfnte lmts at VCA buses and observng the reactve power generaton requred for dfferent set pont voltages. QV curve analyss can be tme consumng f curves have to be found for every bus n the system. Fndng VCA s and ther assocated B s n VSSAD method s based on QV curve analyss performed at each test VCA. It nvolves the placement of a synchronous condenser wth nfnte lmts at VCA buses and observng the reactve power generaton requred for dfferent set pont voltages. QV curve analyss can be tme consumng f curves have to be found for every bus n the system. Thus another method has been proposed by Schlueter [2,3], whch reduces the number of QV curves that need to be found for determnng system s Bs. Coherent bus groups can be found by ths method that have smlar QV curve mnma s and share a smlar set of exhausted generators at these mnma s. Ths method, however, nvolves a farly hgh degree of tral and error and requres the computaton of QV curves at hgher voltage buses before the QV curves for each ndvdual bus group can be found. A modal analyss technque has been appled to evaluate voltage stablty of large power systems [4]. Although t has proven, when combned wth PV analyss, to be an effectve tool for determnng areas prone to voltage nstablty for ndvdual selected system scenaros, t has not been used drectly as an approach to automatcally determne VCAs when numerous contngences or system scenaros are nvolved. In summary, the exstng methods have had only a lmted success n commercal applcaton because they cannot produce satsfactory results for practcal systems. Ths, n general, s because of the followng dffcultes: - The problem s hghly nonlnear: To examne the effects of contngences the system s repeatedly stressed n some manner by ncreasng system load and generaton. The process of stressng the system normally ntroduces a myrad of non-lneartes and dscontnutes between the base case operatng pont and the ultmate nstablty pont - The VCAs must be establshed for all expected system condtons and contngences: Fndng VCAs s a large dmensoned problem because many system condtons and contngences need to be consdered. It may not be possble to dentfy a small number of unque VCAs under all such condtons. The VCAs may also change n shape and sze for dfferent condtons and contngences. To deal wth those ssues, a more practcal approach s needed that can clearly establsh the VCAs for a gven system and all possble system condtons. The approach s based on a QV Curve method combned wth Modal Analyss [5,7]. Typcally, a QV curve s created by ncreasngly stressng the system and solvng a power flow at each new loadng pont. When the power flow fals to converge, the nose of the QV curve has been reached and ths pont corresponds to the stablty lmt for that partcular mposed stress. Contngences can also be appled at ponts along the QV curve to generate post-contngency. Techncal approach The proposed approach s based on a PV curve method combned wth Modal Analyss. The general approach s as follows: - A system operatng space s defned based on a wde range of system load condtons, dspatch condtons, and defned transactons (source-to-snk transfers). 212 PZEGLĄD ELEKTOTECHNICZNY (Electrcal evew), ISSN ,. 88 N 2/2012

2 - A large set of contngences s defned whch spans the range of credble contngences. - Usng PV curve methods, the system s pushed through every condton, under all contngences untl the voltage nstablty pont s found for each condton. - To dentfy the VCA for each case usng modal analyss: At the pont of nstablty for each case (nose of the PV curve) modal analyss s performed to determne the crtcal mode of nstablty as defned by a set of bus partcpaton factors correspondng to the zero egenvalue (bfurcaton pont). - The results of the modal analyss wll s placed n a database for analyss usng data mnng methods to dentfy the VCAs and track them throughout the range of system changes. - The reactve reserve requrements for selected VCA wll then be establshed. Whle the concept of V-Q senstvty s a famlar one (the effect on voltage of a reactve njecton at a bus), the concept of modal analyss, as used to determne area prone to voltage nstablty, s less wdely understood. Therefore, t s useful to relate the two concepts to classfy the meanng of modal analyss results. The network constrants are expressed n the followng lnearzed model around the gven operatng pont [6,8]: P J P J PV (1) Q JQ JQV where: ΔP ncremental change n bus real power, ΔQ ncremental change n bus reactve power, Δθ ncremental change n bus voltage angle, ΔV ncremental change n bus voltage magntude, J Pθ, J PV, J Qθ, J QV are Jacoban submatrces. The elements of the Jacoban matrx gve the senstvty between power flow and bus voltage changes. Whle t s true that both P and Q affect system voltage stablty to some degree, we are prmarly nterested n the domnant relatonshp between Q and V. Therefore, at each operatng pont, we may keep P constant and evaluate voltage stablty by consderng the ncremental relatonshp between Q and V. Ths s not to say that we neglect the relatonshp between P and V, but rather we establsh a gven P for the system and evaluate, usng modal analyss, the Q-V relatonshp at that pont. Based on the above consderaton the ncremental relatonshp between Q and V can be derved from Equ.1 by lettng ΔP=0: (2) Q J V where J s the reduced Q-V Jacoban sub-matrx: (3) J J J J J From Equ. 2 we can wrte: QV Q P (4) V J Q where the nverse matrx (5) V Q J The th dagonal element of matrx PV J s the V-Q senstvty matrx: J s the V-Q senstvty at bus, whch represents the slope of the V-Q curve at the gven operatng pont. A postve V-Q senstvty s ndcatve of stable operaton: the smaller the senstvty the more stable the system. The senstvty becomes nfnte at the stablty lmt. Senstvty matrx J s a full matrx whose elements reflect the propagaton of voltage varaton through the system followng a reactve power njecton n a bus. V-Q senstvty Analyss V-Q senstvtes provde nformaton regardng the combned effects of all modes of voltage reactve power varatons. The relatonshp between bus V-Q senstvtes and egenvalues can be derved from the general Equ. 4. Usng the egenvalues and egenvectors of the reduced Jacoban matrx J we can wrte: (6) J where: 1, 2,..., N J ;,..., s the rght egenvector matrx of 1, 2 N s the left egenvector matrx of J and Λ s the egenvalue matrx of J. Snce we can also wrte: (7) J Substtutng Equ. 6 n Equ. 4 gves: (8) V 1 Q or: (9) V Q where λ s the th egenvalue of J and and are ts correspondng rght and left egenvectors. Bus V-Q senstvtes can be derved from Equ. 9 as follows. Let ΔQ = e k where e k has all zero elements except for the k th element that s equal to 1. The V-Q senstvty at bus k s then gven by: Vk (10) k k Q where and k k k are the k th elements of the rght and left egenvectors respectvely correspondng to egenvalue λ. The V-Q senstvtes provde nformaton regardng the combned effects of all modes on voltage-reactve power varaton. The magntudes of the egenvalues can provde a relatve measure of the proxmty to voltage nstablty. When the system reaches the voltage stablty crtcal pont, the modal analyss s helpful n dentfyng the voltage stablty crtcal areas and buses, whch partcpate n each mode. The relatve partcpaton of bus k n mode s gven by the bus partcpaton factor: P ξ (11) k k k From Equaton 10 we could see that bus partcpaton factor P k determnes the contrbuton of egenvalue λ to the V-Q senstvty at bus k. VCA Identfcaton Procedure In the proposed approach, the power system s stressed to ts stablty lmt for varous system condtons under all credble contngences. At the pont of nstablty (nose of the PV curve) modal analyss s performed to determne the crtcal mode of voltage nstablty for whch a set of bus η PZEGLĄD ELEKTOTECHNICZNY (Electrcal evew), ISSN ,. 88 N 2/

3 partcpaton factors (PFs) correspondng to the zero egenvalue (bfurcaton pont) s calculated. Based on these PFs, the proposed method dentfes the sets of buses and generators that form the varous VCAs n a gven power system. It s assumed that for a gven contngency case, buses wth hgh PFs ncludng generator termnal buses, form a VCA. Ths suggests that each contngency case mght produce ts own VCA. In practce, however, the large number of credble contngency cases generally wll produce only a small number of VCAs because several contngences are usually related to the same VCA. The proposed dentfcaton procedure apples heurstc rules to (a) group contngences that are related to the same VCA; and (b) dentfy the specfc buses and generators that form each VCA (see Fgure 1). Determne X=number of generators at ther lmt n contngency case I; Select the buses wth PFs >= PF_T; Denote the selected buses as set SFA; Include the correspondng X generator buses, f any, nto SFA; End. Note: A SFA set conssts of: a. buses wth PFs >= PF_T, selected for analyss, b. X generator buses that have exhausted ther reactve reserves. Fg. 1. VCA dentfcaton n a Power System The VCAs are dentfed based on the results of the analyss of all credble contngences and dfferent power system condtons. Each VCA dentfed s related to a cluster of contngences; these cases are the so-called support of that VCA. Ths means that frst smlar contngency cases are clustered and then the specfc buses and generators that form the VCAs are dentfed. Before clusterng contngency cases, however, a prelmnary selecton of buses and generators s done at an earler stage of the VCA dentfcaton process as ndcated n Fgure 2. Step 1: Selecton of Buses for VCA Identfcaton From each contngency modal analyss results, a subset of buses wth hgh PF s selected for further analyss (remanng buses are dscarded). Several strateges to select such subset can be appled. For nstance, one could predefne a PF threshold and then select the buses wth PFs above ths threshold. Such approach would assume that t s meanngful to compare PFs values among varous contngences. However, such assumpton may be false because the PFs calculated for each contngency are normalzed wth respect to the maxmum PF value of each mode. Therefore, because dfferent contngences use dfferent references for ther PFs, they cannot be compared. Snce each contngency case s unque, a better approach to select the Set for Further Analyss (SFA) buses s to base t on the characterstcs of each contngency. Generator termnal buses are PV type buses and thus are not ncluded n the reduced Jacoban matrx. Therefore, PF cannot be calculated for a generator termnal bus untl the generator exhausts ts reactve reserves, whch s marked as a Q- lmted (QL) bus, and t becomes a PQ type bus. The number of QL buses, characterstc for each contngency, determnes the selecton of SFA buses. The selecton for SFA buses ncludes all generators QL buses and a subset of buses wth the hghest PFs. The pseudo-code for ths step s as follows: Set PF_threshold=PF_T For each contngency case : Fg. 2. Data Flow Dagram for VCA Identfcaton Procedure Step 2: Clusterng of Contngency Cases based on SFAs. In ths step only the buses havng hgh PFs are used for comparson (generators whch are at ther reactve power lmt are not consdered at ths stage). Several contngency clusters Ck are constructed n ths step. Later on, these clusters wll be used to dentfy the VCAs n the power system (Steps 6 and 7). The frst step n a clusterng process s the selecton of a partcular SFAx as the base for the cluster (heurstc rules for selectng the base are gven n next secton). Then every SFA s compared aganst ths base set. If predetermned percentage of SFA buses are members of the SFAx set, then those sets are consder beng smlar and are grouped together. After groupng the SFAs smlar to SFAx base set a new base set SFAz s selected for the remanng SFAs. Then the process s repeated untl all SFAs are grouped (groups of a sngle SFA are allowed). The pseudo-code for ths step s as follows: Set k=1 (counter for number of clusters Ck) epeat untl all SFAs are grouped Create empty cluster Ck; From SFAs not yet grouped select base set SFAx.; Include SFAx n Ck (SFAx Ck) - For every SFA not yet grouped: If buses n SFA are smlar to buses n SFAx then nclude SFA n Ck - End for every SFA not yet grouped If every SFA has been grouped then STOP; otherwse ncrease k and repeat the procedure. End 214 PZEGLĄD ELEKTOTECHNICZNY (Electrcal evew), ISSN ,. 88 N 2/2012

4 Step 3: Normalzaton of Generator Buses PFs. For every SFA n cluster Ck, the generator buses PFs are normalzed. If a gven SFA contans X generator buses then the maxmum PF value of those X buses s used as a normalzaton factor. Then, a subset of Y generator buses wth the hghest normalzed PFs s selected for further analyss (remanng generator buses are elmnated from SFA). The pseudo-code for ths step s as follows: For each cluster Ck For each SFA n Ck Normalzed the PFs of the X generator-buses; Select the Y generator buses wth normalzed PFs>=β; In SFA: replace set X by set Y; End for each SFA n Ck; End for each cluster Ck. Note: The β factor s used to select only the most sgnfcant generator buses; β s a threshold for the generator buses normalzed PFs below whch the generator buses are excluded from SFAs. Step 4: Selecton of Generators n Cluster Ck. For each cluster Ck, the frequency of generator bus partcpatons n ths Ck s calculated as the number of SFAs n whch a gven generator bus s present. The generator buses wth the hghest frequences are selected to represent the cluster Ck reactve reserves and are denoted as GENk. The pseudo-code for ths step s as follows: For each cluster Ck For each generator-bus-z n Ck Compute frequency Fz for generator bus z: Fz=number of SFAs where generator bus z s present; End for each generator bus z; Select set of gen. buses wth Fz>= δ; Denote ths set of gen. bus set as GENk; emove generator buses from each SFA n cluster Ck; End for each cluster Ck. Note: The factor δ s a frequency threshold used for the selecton of generator buses. The hgher the frequency of a generator bus, the hgher the possblty of selectng the generator bus. The value for δ depends on the number of SFAs n a gven Ck. Step 5: Clusterng of Ck based on GENs. In ths step, Ck are grouped together f ther correspondng GEN sets are smlar. Two GENs are consdered smlar f certan percentage of generator buses are matched. If GEN (from C) and GENj (from Cj) are smlar, then C and Cj are grouped together nto a prelmnary VCA, say VCAm. Ths VCAm s assocated wth a set of generator buses GENm that conssts of the generator buses of the combned GEN and GENj. The pseudo code for ths step s as follows: Set m=1 (counter for number of prelmnary VCAs); epeat untl all clusters Ck have been grouped: Create empty prelmnary VCAm Create empty GENm From Ck not yet grouped select base set GENx. Include all SFAs, from correspondng Cx, nto the VCAm: SFA(Cx) VCAm Update GENm = GENx GENm I For each C not yet grouped: If correspondng GEN s smlar to GENx then SFA(C) VCAm Update GENm = GEN GENm I End for each C not yet grouped If all Ck have been grouped then STOP; otherwse ncrease m and repeat the procedure; End. After ths step, a set of prelmnary VCAs s establshed. Each prelmnary VCAm relates to a unque set of generator buses GENm. Step 6: Selecton of buses. For each prelmnary VCAm, compute the frequency of each bus. Then select the buses wth a frequency greater than 50% the number of SFAs n that VCAm. These are the buses that form VCAm of the gven power system. Step 7. Selecton of generators. For each GENm, get the frequency of each generator bus. Then select the generator buses wth a frequency greater than 50% the number of SFAs n the correspondng VCAm. The generators assocated wth these generator buses are the ones that form controllng generators assocated wth VCAm of the gven power system Heurstc ules for Base Selecton and Smlarty Measurement (a) Selecton of a base for clusterng process From the VCA dentfcaton process, as mentoned earler, we can observe that clusterng s carred out twce: Clusterng contngency cases based on SFAs (Step 2), Clusterng Ck based on GENs (Step 5). Each clusterng process starts wth the selecton of a base set for the cluster. Then any other set s compared to ths base to evaluate whether they are smlar. In Step 2, two dfferent crtera for the selecton of a base SFAx set were tested: Largest contngency (SFA). After the SFAs are found n Step 1, the number of buses n each SFA s counted. The SFA wth the hghest number of buses s selected as the SFAx base for a cluster and then smlar SFAs are grouped together. Most severe contngency (SFA). As part of the voltage stablty assessment of the system, we also compute the margn for each contngency case. The SFA correspondng to the contngency wth the smallest margn s selected as the base of the cluster. Then smlar SFAs are grouped together. The second crteron was found more sutable and therefore t s appled n the VCA dentfcaton process. For clusterng n Step 5, the GEN set wth the hghest number of generator-buses s selected as the base GENx of a cluster. Then smlar GENs are grouped together. (b) Measure of smlarty between sets Whether we are dealng wth SFAs or GENs the measure of smlarty s the same. Frst the numbers of buses n the base sets SFAx or GENx as well as the SFAs or GENs sets for all cases are counted. Then the elements of set- (ether SFA or GEN) are compared wth the elements of the base set (ether SFAx or GENx). The number of common elements C s counted and compared wth the smlarty threshold T. If the number of common PZEGLĄD ELEKTOTECHNICZNY (Electrcal evew), ISSN ,. 88 N 2/

5 elements C s greater than the threshold T, then set- and the base set are consdered beng smlar. The smlarty threshold T s set as a percentage of the number of elements of n the largest set (set- or the base set). If all elements of the smaller set (base or set-) are ncluded n the larger set then those sets are consdered beng smlar. Analyss of VCA Buses Let s assume that the total number of buses n the system equals four (n=4). Let s also assume that two contngency cases are consdered and that ther PFs are those gven by: Bus PFs: CntgA=[B1 B2 B3 B4] =[ ] CntgB=[B1 B2 B3 B4] =[ ] To rank the buses lsted above one proceeds as follows. For a gven contngency, the bus wth the hghest PF s mapped/ranked nto n=4. Then the bus wth the second hghest value s mapped nto (n-1), then the next one nto (n-2) and so on. Buses wth PFs=0 are mapped nto 1 (mnmum rankng value). That s, the buses lsted above are ranked as follows. Bus ankng: CntgA=[B1 B2 B3 B4] =[ ] CntgB=[B1 B2 B3 B4] =[ ] have been exhausted. That s, VCA generators are the locaton where reactve power reserves should be kept so that voltage nstablty s avoded. In the prevous secton, buses are ranked n order to dentfy how mportant they are. Such an approach s not sutable for the generators snce we are not nterested on how mportant (rank) they are, but rather how effectve they are n preventng voltage nstablty. Fg. 4. ankng of Non-VCA Buses (30 Buses and 50 contngency cases) For a sngle contngency case, for nstance, a generator that s ranked n second place mght not be as effectve n avodng voltage nstablty as the generator ranked n the frst place. In other words, reactve power reserves n the generator ranked second wll not produce the same system mprovement as f these reserves were allocated to the generator ranked frst nstead. A per-contngency generator- PF-normalzaton metrc can measure how effectve generators are. It s expected that VCA generators have hgher normalzed PFs than those of non-vca generators. Fg. 3. ankng values VCA-2 Buses (30 Buses and 50 contngency cases) Then, rankng values are normalzed wth respect to n: Normalzed ankng: CntgA=[B1 B2 B3 B4] =[ ] CntgB=[B1 B2 B3 B4] =[ ] The normalzed rankng values are not the same as the PFs. For nstance, the bus wth the second hghest PF s always ranked to the same normalzed rankng value (0.75 n the example gven);.e., ths rankng s ndependent of how dfferent the PFs of these buses are for the dfferent contngences. These normalzed rankng values are used to evaluate the dentfed VCA buses. Fgure 3 shows the rankng values of a set of 30 buses across 50 contngency cases; these buses and contngences are related to VCA-2. On the other hand, Fgure 4 shows the rankng values of a set of non-vca buses. Comparng Fgure 3 versus Fgure 4, one can observe that the rankng values of the VCA buses are hgher than those of the non-vca buses. That s, the dentfed VCA buses are ndeed the most mportant buses prone to voltage nstablty problems. Analyss of VCA Generators VCA generators are those generators that ntate the nstablty of the VCA once ther reactve power reserves Fg. 5. Normalzed PFs of VCA-1 Generators (6 Generators and 50 contngency cases) An example of how to normalzed generators PFs follows. Consder the followng contngences cases and generators PFs. PFs of QL generators (generators that are at ther reactve power lmt): 216 PZEGLĄD ELEKTOTECHNICZNY (Electrcal evew), ISSN ,. 88 N 2/2012

6 CntgA=[G1 G2 G3 G4] = [ ] (max=0.4) CntgB=[G1 G2 G3 G4] = [ ] (max=0.3) Then, one normalzes the PFs wth respect the hghest PF of the correspondng contngency; that s, Normalzed PFs of QL generators: CntgA=[G1 G2 G3 G4] =[ ] CntgB=[G1 G2 G3 G4] =[ ]. Fgure 5 shows the normalzed PFs of the VCA-1 generators; these values are hgher than those of the non- VCA generators. That s, the set of dentfed VCA generators are the most effectve to avod voltage nstablty f reactve reserves are kept n. Performance of VCA-1 Generators when reactve power reserves are ncreased In order to evaluate the effectveness of the dentfed VCA generators, for VCA-1, the followng test was carred out. An addtonal 50 MVA reserve was unformly dstrbuted on the set of the VCA-1 generators. Then the ponts of voltage nstablty of the assocated contngences were computed. These voltage nstablty ponts were compared aganst those when there s no ncrease n reserves. The objectve of ths test s to measure how the power transfer ncreases, for the varous contngences consdered, when the reactve reserves n ths set of VCA-1 generators s ncreased. The above ncrement n power transfer was compared aganst that obtaned when a 50 MVA reserve s dstrbuted on each of two other sets of generators. Based on experence, these two other sets of generators were dentfed as the most promsng for a hgh power transfer ncrement. Fgure 6 shows the MW-Transfer ncrease obtaned when an addtonal 50 MVA reactve power reserve s dstrbuted on varous sets of generators. The mean MW transfer ncrease (M-MW-Inc) s hgher for the set of VCA-1 generators than that for the other two sets. That s, the dentfed VCA-1 generators are the most effectve n securng voltage stablty snce the ponts of voltage nstablty, for the varous assocated contngences, occurs farther ahead than that at the other two sets of generators tested. Concluson In normal operatng condtons the control system manage to adjust generaton of reactve power n power plants and to control the voltages. If the locaton of reactve power sources s nadequate, the control system may lead the power grd to a blackout. For ensurng approprate relablty of supply n electrcal energy are responsble network-companes, whch usually do not own any generaton. Thus, they have to acheve ther task by control of contracted reserves and generaton-load balancng. Snce customers are free to change ther power demand at any tme, the balancng conssts n requestng a change n generaton. In market economy all contracts are proft orented. A term of the common good or wealth s not appealng to companes desperately needed ncome. The relablty of supply n market envronment s on one hand more mportant due to customer expectaton and on the other hand more dffcult and more expensve to acheve. An overvew of the state-of-the-art of on-lne VSA has been presented and a new method for the dentfcaton of voltage control areas s descrbed. For a wde range of system condtons and contngences, the technque can dentfy the buses n each VCA and dentfy VCAs whch are common for a set of contngences and/or condtons. In addton, the method dentfes the generators whch are crtcal to mantanng stablty for a gven VCA. Voltage control areas descrbe the regons n a power system that under specfc condtons are prone to voltage nstablty. Intellgent systems hold promse to mprove VSA speed, provde adaptve learnng capabltes and offer the ablty to dentfy key system parameters. Lterature [1] Voltage Stablty Assessment: Concepts: Concepts, Practces and Tools, IEEE Power Engneerng Socety, Power System Stablty Subcommttee Specal Publcaton, August [2]. A. Schlueter, A Voltage Stablty Securty Assessment Method, IEEE Trans. Power Syst., Vol. 13, pp , Nov [3]. A. Schlueter, S. Lu, K. Ben-Klan, Justfcaton of the Voltage Stablty Securty Assessment and Dagnostc Procedure Usng a Bfurcaton Subsystem Method, IEEE Trans. Power Syst., Vol. 15, pp , Aug [4] C. A. Aumuller, T. K. Saha, Determnaton of Power System Coherent Bus Groups by Novel Senstvty-Based Method for Voltage Stablty Assessment, IEEE Trans. Power Syst., Vol. 18, No.3, pp , Aug [5] G. E. Tovar, J. G. Calderon, V. E. de la Torre, F. I. Neva, eactve eserve Determnaton Usng Coherent Clusterng Appled to the Mexcan Norwest Control Area, Power Systems Conference and Exposton, PSCE 2004, New York Cty, Oct.10-13, 2004 [6] J. Zhong, E. Noble, A. Bose and K. Bhattacharya Localzed eactve Power Markets Usng the Concept of Voltage Control Areas, IEEE Trans. Power Syst., Vol. 19, No.3, pp , Aug [7] M.K. Verma, S.C. Srvastava, Approach to Determne Voltage Control Areas Consderng Impact of Contngences, IEE Proc.-Gener. Trans. Dstrb. Vol. 152, No. 3, May 2005 [8] H. Lu, A. Bose, V. Vencatasubramanan, A Fast Voltage Securty Assessment Method Usng Adaptve Boundng, IEEE Trans. Power Syst., Vol. 15, No.3, pp , Aug Fg. 6. Transfer Increase Comparson: Set of VCA-Generators Versus Other Sets. MW Transfer-Increment vs. eactve eserve Increment (50MVars) Authors: dr. nż. obert Ls, Insttute of Electrc Power Engneerng Wroclaw Unversty of Technology, Wybrzeze Wyspanskego 27, Wroclaw, Poland, E-mal: robert.ls@pwr.wroc.pl; dr nż. Grzegorz Błajszczak, PSE-Operator (Polsh Transmsson System Operator), ul. Warszawska 165, Konstancn- Jezorna, Poland, E-mal: grzegorz. blajszczak@pse-operator.pl PZEGLĄD ELEKTOTECHNICZNY (Electrcal evew), ISSN ,. 88 N 2/