Capacitance based Reliability Indices of a Real Time Radial Distribution Feeder

Transcription

1 IOSR Journal of Electrcal and Electroncs Engneerng (IOSR-JEEE) e-issn: ,p-ISSN: , Volume 7, Issue 2 (Jul. - Aug. 23), PP -2 Capactance based Relablty Indces of a Real Tme Radal Dstrbuton Feeder D. Mural, 2 Mr.A.Hema sekhar PG Student (EPS) Department of EEE Sr Venkatesa Perumal College of Engg & Tech, puttur 2 Assocate Professor & HOD Department of EEE Sr Venkatesa Perumal College of Eng. & Tech,Puttur Abstract: Assessment of customer power supply relablty s an mportant part of dstrbuton system operaton and plannng. Dstrbuton system relablty assessment s a measure of contnuty and qualty of power supply to the consumers, whch manly depends on nterrupton profle, based on system topology and component relablty data. The paper manly descrbes about the radal Dstrbuton system relablty s evaluated n two methods. One by placng capactor at weak voltage nodes for mprovement of voltage profles, reducng the total losses. Second way by mprovng relablty ndces by placng protectve equpment (solators) n the feeder. Ths paper present an effectve approach for real tme evaluaton of dstrbuton power flow solutons wth an objectve of determnng the voltage profles and total losses. To mprove the voltage profles and reducng losses by placng capactors at weak voltage profle nodes usng Partcle Swarm Optmzaton (PSO) technque. The Dstrbuton System Relablty Indces are also calculated for the exstng radal dstrbuton system before and after placement of solator. In ths paper we have consdered the load dversty factor for analyss of load data for real tme system. Two matrces the bus-njecton to branch-current matrx (BIBC), the branch-current to bus voltage matrx (BCBV) and a smple matrx multplcaton are used to obtan power flow solutons. Ths paper also presents an approach that determnes optmal locaton and sze of capactors on exstng radal dstrbuton systems to mprove the voltage profles and reduce the actve power loss. The performance of the method was nvestgated on an kv real tme Upadyayanagar radal dstrbuton feeder as system of case study. A matlab program was developed and results are presented. Keywords: BIBC, BCBV, Dversty Factor, BVSI, Relablty Indces, Dstrbuton Load Flows, PSO. I. Introducton The demand for electrcal energy s ever ncreasng. Today over 2% (theft apart!!) of the total electrcal energy generated n Inda s lost n Transmsson (5-7%) and Dstrbuton (5-8%). The electrcal power defct n the country s currently about 35%.Clearly, reducton n losses can reduce ths defct sgnfcantly. It s possble to brng down the dstrbuton losses to 6-8% level n Inda wth the help of newer technologcal optons (ncludng nformaton technology) n the Electrcal Power Dstrbuton Sector whch wll enable better montorng and control. The electrc utlty system s usually dvded nto three subsystems whch are Generaton, Transmsson, and Dstrbuton. A fourth dvson, whch sometmes made s Sub-Transmsson. Electrcty dstrbuton s the fnal stage n the delvery of electrcty to end users. A Dstrbuton Network carres electrcty from the transmsson system and delvers t to consumers. Typcally, the network would nclude medumvoltage (<5kV) power lnes, electrcal substatons and pole-mounted transformers, low-voltage (less than V) dstrbuton wrng and sometmes electrcty meters. Electrc power s normally generated at -25kV n a power staton. To transmt over long dstances, t s then stepped-up to 4kV, 22kV or 32kV as necessary. Power s carred through a transmsson network of hgh voltage lnes. Usually, these lnes run nto hundreds of klometers and delver the power nto a common power pool called the grd. The grd s connected to load centers through a sub-transmsson network of normally 33kV (or sometmes 66kV) lnes. These lnes termnate nto a 33kV (or 66kV) substaton, where the voltage s stepped-down to kv for power dstrbuton to load ponts through a dstrbuton network of lnes at kv and lower. The power network, whch generally concerns the common man s the dstrbuton network of kv lnes or feeders downstream of the 33kV substaton. Each kv feeder whch emanates from the 33kV substaton branches further nto several subsdary kv feeders to carry power close to the load ponts (localtes, ndustral areas, vllages,etc.,).at these load ponts, a transformer further reduces the voltage from kv to 45V to provde the last mle connecton through 45V feeders (Low Tenson (LT) feeders) to ndvdual customers, ether at 24V (as ph. supply) or at 45V (as 3ph. supply).a feeder could be ether an overhead lne or an underground cable. In urban areas, owng to the densty of customers, the length of an kv feeder s generally up to 3 km. On the other hand, n rural areas, the feeder length s much larger (up to 2 km). A 45V feeder should normally be restrcted to about.5-. km unduly long feeder s lead to low voltage at the consumer end. Page

2 II Dversty Factor The probablty that a partcular pece of equpment wll come on at the tme of the faclty's peak load. It s the rato of the sum of the ndvdual non-concdent maxmum demands of varous subdvsons of the system to the maxmum demand of the complete system[6]. The dversty factor s always greater than. The (unoffcal) term dversty, as dstngushed from dversty factor refers to the percent of tme avalable that a machne, pece of equpment, or faclty has ts maxmum or nomnal load or demand (a 7% dversty means that the devce n queston operates at ts nomnal or maxmum load level 7% of the tme that t s connected and turned on). Ths dversty factor s used to estmate the load of a partcular node n the system. III Load Flow Studes The load-flow study n a power system has great mportance because t s the only system whch shows the electrcal performance and power flow of the system operatng under steady state [-3]. A load-flow study calculates the voltage drop on each feeder, the voltage at each bus, and the power flow n all branch and feeder crcuts. Losses n each branch and total system power losses are also calculated. Load-Flow studes are used to determne the reman wthn specfed lmts, under varous contngency condtons only. Load-flow studes are often used to dentfy the need for addtonal Generaton, Capactve/Inductve VAR support or the placement of capactors and/or reactors to mantan system voltages wthn specfed lmts. An effcent load-flow study plays vtal role durng plannng of the system and also for the stablty analyss of the system. Usually the dstrbuton networks are ll-condtoned n nature. Therefore, the varables for the load-flow analyss of dstrbuton systems are dfferent from those of transmsson systems. Many modfed versons of the conventonal load-flow methods have been suggested for solvng power networks wth hgh R/X rato. The followng are the effectve load flow technques used n the dstrbuton networks are Sngle-Lne Equvalent Method, Very Fast Decoupled Method, Ladder Technque, Power Summaton Method, Backward and Forward Sweepng Method. The proposed algorthm s tested on a Real Tme system. Formulaton of Load Flow Model (a) Algorthm Development: The technque s based on two matrces, the bus-njecton to branch-current matrx and the branch current to bus-voltage matrx, and equvalent current njectons. In ths secton, the development procedure wll be descrbed to develop BCBV and BIBC for radal dstrbuton feeder. For bus, the complex load S s expressed by S =P +jq () Where =, 2,... N And the correspondng equvalent current njecton at the k th teraton of soluton s I k =(P +jq /V k )* (2) Where V k and I k are the bus voltages and equvalent current njecton of bus at kth teraton respectvely. (b) Relatonshp Matrx Development A smple dstrbuton network shown n fgure s used to fnd the current equatons are obtaned from the equaton (2). The relatonshp between bus currents and branch currents can be obtaned by applyng Krchhoff s current law (KCL) to the dstrbuton network. Usng the algorthm of fndng the nodes beyond all branches proposed by Gosh et al. The branch currents then are formulated as functons of equvalent current njectons for example branch currents B, B 3 and B 5 can be expressed as Fgure. Smple dstrbuton system B = I 2 +I 3 +I 4 +I 5 +I 6 B 3 =I 4 +I (3) B 5 = I 6 Therefore the relatonshp between the bus current njectons and branch currents can be expressed as B B 2 B 3 B4 B 5 I I I I I (4a) Page

3 Eq (4a) can be expressed n general form as [B]= [BIBC] [I] (4b) The constant BIBC matrx s an upper trangular matrx and contans values of and only. The relatonshp between branch currents and bus voltages as shown n Fg.. For example, the voltages of bus 2, 3, and 4 are V 2 =V -B (5a) V 3 =V 2 -B (5b) V 4 =V 3 -B (5c) where V s the voltage of bus, and j s the lne mpedance between bus and bus j. Substtutng (5a) and (5b) nto (5c), (5c) can be rewrtten as V 4 =V -B 2 -B B (6) V V 2 V V 3 V V 4 V V5 V V B B 2 B 3 (7a) B5 B 6 From (6), t can be seen that the bus voltage can be expressed as a functon of branch currents, lne parameters, and the substaton voltage. Smlar procedures can be performed on other buses, therefore, the relatonshp between branch currents and bus voltages can be expressed as [ V]= [BCBV] [B] (7b) Where BCBV s the branch current to bus voltage matrx. (C) Buldng Formulaton Development: Observng (4), a buldng algorthm for BIBC matrx can be developed as follows: Step : For a dstrbuton system wth m-branch secton and n bus, The dmenson of the BIBC matrx s m (n- ). Step 2: If a lne secton (B) s located between bus and bus j, copy the column of the I th bus of the BIBC matrx to the column of the j th bus and fll a to the poston of the k th row and the j th bus column. Step 3: Repeat Step (2) untl all lne sectons s ncluded n the BIBC matrx. From equaton (7a and 7b), a buldng algorthm for BCBV matrx can be developed as follows. Step 4: For a dstrbuton system wth m-branch secton and n-k bus, the dmenson of the BCBV matrx s m x (n-). Step 5: If a lne secton s located between bus and bus j, copy the row of the th bus of the BCBV matrx to the row of the j th bus and fll the lne mpedance ( ) to the poston of the j th bus row and the k th column. Step 6: Repeat procedure (5) untl all lne sectons s ncluded n the BCBV matrx. The algorthm can easly be expanded to a mult phase lne secton or bus. D. Soluton Technque Developments: The BIBC and BCBV matrces are developed based on the topologcal structure of dstrbuton systems. The BIBC matrx represents the relatonshp between bus current njectons and branch currents. The correspondng varatons at branch currents, generated by the varatons at bus current njectons, can be calculated drectly by the BIBC matrx. The BCBV matrx represents the relatonshp between branch currents and bus voltages. The correspondng varatons at bus voltages, generated by the varatons at branch currents, can be calculated drectly by the BCBV matrx. Combnng (4b) and (7a), the relatonshp between bus current njectons and bus voltages can be expressed as [ V]=[BCBV][BIBC][I]=[DLF][I] (8) I k =I r (V k )+ji (V k )=((P +jq )/V k )* (9a) [ V k+ ]=[DLF][I k ] (9b) [V k+ ] = [V ] + [ V k+ ] (9c) And the soluton for dstrbuton power flow can be obtaned by solvng teratvely. Accordng to the research, the arthmetc operaton number of LU factorzaton s approxmately proportonal to N 3. For a large value of N, the LU factorzaton wll occupy a large porton of the computatonal tme. Therefore, f the LU factorzaton can be avoded, the power flow method can save tremendous computatonal resource. From the soluton technques descrbed before, the LU decomposton and forward/backward substtuton of the Jacoban matrx or the Y admttance matrx are no longer necessary for the proposed method. Only the DLF matrx s necessary n solvng power flow problem. Therefore, the proposed method can save consderable computaton resources and ths feature makes the proposed method sutable for onlne operaton. E. Losses Calculaton: The Real power loss of the lne secton connectng between buses and +s computed as P Q P (, ) R () RLOSS 2 2, 2 V 2 Page

4 The Reactve power loss of the lne secton between buses and +s computed as 2 2 P Q PXLOSS (, ) X (), 2 V The total Real and Reactve power loss of the feeder P FRLOSS s determned by summng up the losses of all sectons of the feeder, whch s gven by: N PFRLOSS (, ) PRLOSS (, ) (2) N PFXLOSS (, ) PX LOSS (, ) (3) IV Partcle Swarm Optmzaton Partcle Swarm Optmzaton (PSO) s a Meta heurstc parallel search technque used for optmzaton of contnues nonlnear problems. PSO has roots n two man component methodologes perhaps more obvous are tes to artfcal lfe. It s also related, however to evolutonary computaton and has tes to both genetc algorthms and evolutonary programmng. It requres only prmtve mathematcal operators, and s computatonally nexpensve n terms of both memory requrements and speed. It conducts searches usng a populaton of partcles, correspondng to ndvduals. Each partcle represents a Canddate soluton to the capactor szng problem. In a PSO system, partcles change ther postons by flyng around a mult-dmensonal search space untl a relatvely unchanged poston has been encountered, or untl computatonal lmts are exceeded. The general elements of the PSO are brefly explaned as follows: Partcle X(t): It s a k-dmensonal real valued vector whch represents the canddate soluton. For an th partcle at a tme t, the partcle s descrbed as X (t)={x,(t), X,2(t),...X,k(t)}. Populaton: It s a set of n number of partcles at a tme t descrbed as {X (t), X 2 (t) X n (t)}. Swarm: It s an apparently dsorganzed populaton of movng partcles that tend to cluster together whle each partcle seems to be movng n random drecton. Partcle Velocty V(t): It s the velocty of the movng partcle represented by a k-dmensonal real valued vector V (t)= {v,(t), v,2(t) v,k(t)}. Inerta weght W(t): It s a control parameter that s used to control the mpact of the prevous velocty on the current velocty. Partcle Best (pbest): Conceptually pbest resembles autobographcal memory, as each partcle remembers ts own experence. When a partcle moves through the search space, t compares ts ftness value at the current poston to the best value t has ever attaned at any tme up to the current tme. The best poston that s assocated wth the best ftness arrved so far s termed as ndvdual best or Partcle best. For each Partcle n the swarm ts pbest can be determned and updated durng the search. Global Best (gbest): It s the best poston among all the ndvdual pbest of the partcles acheved so far. Velocty Updaton: Usng the global best and ndvdual best, the th partcle velocty n kth dmenson s updated accordng to the followng equaton. V[][j]=K*(w*v[][j]+c*rand*(pbestX[][j]- X[][j])+ c2*rand2*(gbestx[j]-x[][j])). Where, K constrcton factor, c, c2 weght factors, w Inerta weght parameter, partcle number, j control varable, rand, rand2 random numbers between and Stoppng crtera: Ths s the condton to termnate the search process. It can be acheved ether of the two followng methods:. The number of the teratons snce the last change of the best soluton s greater than a pre-specfed number.. The number of teratons reaches a pre specfed maxmum value. V. Algorthm for Pso Step: Run the base case dstrbuton load flow and determne the actve power loss. Step2: Identfy the canddate buses for placement capactor. Step 3:Generate randomly n number of partcles where each partcle s represented as partcle[][7]{qc,qc2,..qcj} Step 4: Run the load flow by placng a partcle at the canddate bus for reactve power compensaton and store the actve power loss (TLP). Step 5: Evaluate the ftness value. If the current ftness value s greater than the ts pbest value, then assgn the pbest value to the current value. Step6: Determne the current global best (g_best_partcles) mnmum among the partcles ndvdual best (pbest) values. 3 Page

5 Step 7: Compare the global poston wth prevous. If the current poston s greater than the prevous, then set the global poston to the current global poston. Step 8: update the partcle velocty by usng V[][j]=K*(w*v[][j]+c*rand*(pbestX[][j]-X[][j])+ c2*rand2*(gbestx[j]-x[][j])). Step 9: Update the poston of partcle by addng the velocty v[][j]. Step : Now run the load flow and determne the actve power loss (pl) wth the updated partcle. Step : Repeat step 5 to 7 Step 2: Repeat the same procedure for each partcle from step 4 to step 7. VI Relablty Indces System Average Interrupton Duraton Index (SAIDI) The most often used performance measurement for a sustaned nterrupton s the System Average Interrupton Duraton Index (SAIDI). Ths ndex measures the total duraton of an nterrupton for the average customer durng a gven perod. SAIDI s normally calculated on ether monthly or yearly bass; however, t can also be calculated daly, or for any other perod. Sumofcustomer nt erruptonduraton SAIDI Tota ln umberofcustomers U * N N Where U =Annual outage tme, Mnutes, N =Total Number of customers of load pont. SAIDI s measured n unts of tme, often mnutes or hours. It s usually measured over the course of a year, and accordng to IEEE Standard the medan value for North Amercan utltes s approxmately.5 hours. Customer Average Interrupton Duraton Index (CAIDI) Once an outage occurs the average tme to restore servce s found from the Customer Average Interrupton Duraton Index (CAIDI). CAIDI s calculated smlar to SAIDI except that the denomnator s the number of customers nterrupted versus the total number of utlty customers. CAIDI s, Sum of customer nterruptons duratons CAIDI= = U N.. (5) Total number of customers nterrupptons N I Where U =Annual outage tme, Mnutes, N = Total Number of customers of load pont., λ =Falure Rate. CAIDI s measured n unts of tme, often mnutes or hours. It s usually measured over the course of a year, and accordng to IEEE Standard the medan value for North Amercan utltes s approxmately.36 hours System Average Interrupton Frequency Index (SAIFI) The System Average Interrupton Frequency Index (SAIFI) s the average number of tme that a system customer experences an outage durng the year (or tme perod under study). It s usually measured over the course of a year, and accordng to IEEE Standard the medan value for North Amercan utltes s approxmately. nterruptons per customer. Total number of customer nteruptons SAIFI= = N (6) Total number of customers served NI SAIDI SAIFI. (7) CAIDI Where N =Total Number of customers nterrupted. λ =Falure Rate. Average Servce Avalablty Index (ASAI) Ths s sometmes called the servce relablty ndex. The ASAI s usually calculated on ether a monthly bass (73 hours) or a yearly bass (8,76 hours), but can be calculated for any tme perod. The ASAI s found as,. (8) ( r* N) ASAI [ ( )]* ( N * T) T (4) ASUI ASAI. (9) Where T= Tme perod under study, hours. r =Restoraton Tme, Mnutes, N =Total Number of customers nterrupted. N T =Total Customers served. 4 Page

6 Average Energy Not Suppled (AENS) Ths s also called as Average System Curtalment Index (ASCI) Totalenergynot sup pled AENS Tota ln umberofcustomersserved L * ( ) a () U N. (2) VII. Investgated REAL TIME SYSTEM &RESULTS In ths paper real tme radal feeder s consdered, UPADHYA NAGAR urban feeder located at 33Kv MANGALAM substaton n Trupat, Chttoor (Dt.), Andhra Pradesh, Inda. It s an fast growng resdental area shown n fgure 2. Real tme radal feeder system data The radal dstrbuton systems have followng characterstcs Base Voltage = KV.Base MVA=. Conductor type = All Alumnum Alloy Conductor (AAAC) Resstance =.55 ohm/km., Reactance =.35 ohm/km. A software program was developed n MATLAB for Load flow soluton and PSO s used for placement of capactor to analyze the results for Radal Dstrbuton feeder. To understand the effectves of the method, a 42-node kv Upadhayan urban feeder s selected. Lne data for ths feeder s shown n Table I. Throughout day Load s not constant; t vares from tme to tme. By consderng the terms Dversty factor and Power Factor, fve deferent condtons are consdered.. Average DF Good PF, 2.Hgh DF Hgh PF, 3.Hgh DF Low PF, 4. Low DF Hgh PF, 5.Low DF low PF 6. Average DF Poor PF 7. Unty DF Low PF.Generally a feeder that occurs wth Average DF Good PF where Average DF s.4 and Good PF s.93. When the load s hgh (Hgh DF) and the PF s also hgh (Hgh PF), ths condton does not occur n the day but for the analyss only t consdered. When the load s hgh (Hgh DF), the PF decreases (Low PF), ths condton occurs durng the peak demand. When the load n Low (Low DF) then the PF s hgh (Hgh PF), ths condton occurs durng the lght load condtons. Low DF and Low PF condton does not occur n the day. Ths condton s assumed for analyss purpose only. The load flow soluton obtaned s used to know bus voltages profles for 7 condtons whch are shown n below fgure 3 and losses n Table II. Ths Upadhayanagar feeder s not nstalled by any capactor bank at LT sde. Wthout nstallaton also there are no nodes havng voltages less than.95 p.u. So there s no need of capactor placement for above fve condtons. Fgure 2: Upadhayanagar Radal feeder, Trupat as per standard system 5 Page

7 Table I Lne data of Upadhayanagar Feeder, Trupat Bus No From Node To Node Dstance (KM) R Ω X Ω Fgure 3: bus voltage for dfferent condtons by BIBC & BCBV method 6 Page

8 Bus Voltage Senstvty Index (BVSI): Load flow wth capactor capacty of 5% of the total feeder loadng capacty s carred out to fnd BVSI at varous buses usng (2). Fgure 4 shows the varaton of VSI at varous buses. As seen from ths Fgure5.4, bus number 7 s havng the lowest BVSI value of Therefore, bus 7 s consdered as the canddate bus for the capactor placement. Fgure 4: bus Varaton of BVSI wth bus number. The results shows that followng buses are senstve bus voltages (<.95 pu) 36,37,38,39,4,4 and 42 and are nstable bus voltages, that can be mproved by placng capactor at sngle node or by placng capactor at multple nodes. By usng Partcle Swarm Optmzaton Technque (Secton V). Multple placed capactors have hgher voltages profles than the sngle placement shown fgure 4. The results for power losses are shown n table III for before and after placement of capactor. These losses are compared wth losses obtaned by usng load flow and energy consumpton method. Energy losses are computed for Updahayanagar feeder by real tme energy consumed data by the feeder from substaton. It s observed that the computed energy losses closely match wth the calculated energy (real tme data) losses Condtons Table II: Losses at dfferent condtons Real Power losses (KW) Reactve Power losses (KW) Total Losses (KW) Avg DF Poor PF AVG DF Good PF Hgh DF Good PF Hgh DF Poor DF Low DF Good PF Low DF Poor DF Unty DF Poor PF Table III: Power loss of the feeder before and after compensaton Sngle Placement of capactor Multple Placement of Capactor Q _ loss = 89.35KVAR Q _loss = KVAR P _loss= kw P _loss= KW MIN_V=.9462 MIN_V=.9462 Rank= 36, 37, 38, 39, 4 4, 42, 43 Rank= 36, 37, 38, 39, 4 4, 42, 43 After Compensaton Q _ loss =33.48 KVAR P _loss= KW MIN_V=.9634 Rank= Q _ loss = KVAR P _loss= KW MIN_V=.9634 Rank= 7 Page

9 Fgure 5: Voltage profles for before, after capactors placement usng PSO. Table IV Injected Reactve Power usng PSO at dfferent nodes Nodes Compensated 36,38 Best Node=36 Best Partcle KVAR Best Node=38 Best Partcle KVAR Total Injected Reactve Power KVAR Table IV represents the compensated nodes after placng capactor at sngle and multple nodes. Table V represents the losses calculated as per substaton and usng mat lab. Table V. Power loss calculaton by usng load flow method Upadyayanagar Feeder BIBC and BCBV Method Avg. DF Gud. PF TLP = KW TLQ = KVAR TL = KW Energy Loss =(TLP*24*3) Unts = 3.72 % Energy Loss as per PPL Sheet= 3.72% of 5233=283.3 Unts The detals of the dstrbuton system are shown n Table VI. There are 5 nterrupton cases durng the year (Table VII). When the feeder was not provded wth solators, all load ponts got durng the 5 nterruptons. The Dstrbuton System Relablty Indces are calculated by usng secton VI and the results are tabulated n IX. When the feeder s provded wth solator at 3 th node, the load pont 3 wll only be and the number of load ponts s reduced from 43 to 35durng 5 nterrupton cases. Dstrbuton Relablty Indces are shown n Table IX. The percentage of ndces s represented n pe chart as shown n Fgure 7 wth and wthout solator. When the feeder s not provded wth solator the Average Energy Not Suppled (AENS) s KWh/Customer. When the feeder s provded wth solator at 3 th node the Average Energy Not Suppled (AENS) s reduced to.272 KWh/Customer Table VI Detals of Dstrbuton System S.NO No. of P Avg. Customers (KW) Load Page

10 Interrupton data Interrupton Load Pont Case Affected All load ponts get Table VII Interrupton effect n a calendar year (wthout solator) 2 All load ponts get 3 All load ponts get 4 All load ponts get 5 All load ponts get Duraton Cause of Interrupton (hrs).52 Lne clearence for Transformer mantance.35 Man supply faled due to Dstrbuton lne damage.5 Fault n dstrbuton lne.5 Lne clearance for Transformer erracton.25 man supply faled due to lne Fault 3. Trp due to envronmental condtons.2 man supply faled due to fault 3.5 Load shut for general mantance n the feeder.3 man supply faled due to fault 6. Three lne clearances for erracton of new transformers All load ponts get All load ponts get.45 man supply faled due to fault 4 no's 2.25 Lne break down due to fault 2 no's 9 Page

11 Interrupton Case Table VIII Interrupton effect n a calendar year (wth solator) Load Pont Affected Duraton (hrs) Cause of Interrupton 4,5,6,7,8,9,2, 2,22,23,24,25,26,27, 28,29,3,3 2 4,5,6,7,8,9,2, 2,22,23,24,25,26,27, 28,29,3,3 3 4,5,6,7,8,9,2, 2,22,23,24,25,26,27, 28,29,3,3 4 4,5,6,7,8,9,2, 2,22,23,24,25,26,27, 28,29,3,3 5 4,5,6,7,8,9,2, 2,22,23,24,25,26,27, 28,29,3,32 4,5,6,7,8,9,2, 2,22,23,24,25,26,27, 28,29,3,33 4,5,6,7,8,9,2, 2,22,23,24,25,26,27, 28,29,3,34.52 Lne clearance for Transformer mantance.35 Man supply faled due to Dstrbuton lne damage.5 Fault n dstrbuton lne.5 Lne clearence for Transformer erracton.25 man supply faled due to lne Fault 3. Trp due to envronmental condtons.2 man supply faled due to fault 3.5 Load shut for general mantance n the feeder.3 man supply faled due to fault 6. Three lne clearances for erracton of new transformers.45 man supply faled due to fault 4 no's 2.25 Lne break down due to fault 2 no's Table IX Dstrbuton system Relablty Indces wth wthout Isolator ndces Wthout solator Wth solator AIFI 5. nterruptons/customer.75 nterruptons/customer AIDI 2 hrs/customer 4.57 hrs/customer AIDI 4.2 hrs/customer nterrupton 4.2 hrs/customer nterrupton SAI SUI ENS KWh/customer.272 KWh/customer Fgure 6.Indces representaton n pe chart wth and wthout solator VIII Concluson The radal dstrbuton Kv Upadyanagar feeder s appled wth load flow and the feeble voltage profles are dentfed and those nodes are proposed for capactor placement usng partcle swarm optmzaton technque. The voltages profles and losses before and after compensaton usng PSO for sngle and multple 2 Page

12 placement of capactor, the voltage profles get mproved and losses get reduced. Dstrbuton system Relablty ndces are evaluated for the feeder and results are presented. It s concluded that by provdng more solators n the radal feeder we can reduce the Average Energy Not Suppled (AENS) to the customers ther by mproves the contnuty of power supply References [] Ramanjaneyulu Reddy P, N.M.G.Kumar, P.Sangamshwara Raju, Load flow based relablty assessment of a real tme radal dstrbuton system case study UNIASCIT, Vol2(3), 22, [2] K.Prakash, M.Sydulu, Partcal Swarm Optmzaton Based Capactor Placement on Radal Dstrbuton System,Power Engneerng Socety, IEEE General Meetng - PES, pp. -5, 27 [3] S.Svanagaraju, J.Vswanatha Rao, M.Grdhar, A Loop based load flow method for weakly meshed dstrbuton network, ARPN Journal of Engg. and Appled Scences, Vol 3 No 4, 28. [4] A.A.A. Esmn and G. Lambert-Torres, A Partcle Swarm Optmzaton Appled to Loss Power Mnmzaton, IEEE Transactons on Power Systems, USA, vol. 2, no. 2, pp , 25. [5] S. Ghosh and D. Das, Method for load-flow soluton of radal dstrbuton Networks, IEEE Proceedngs on Generaton, Transmsson & Dstrbuton, Vol.46, No. 6, pp , 999 [6] Turan Gonen Electrc power dstrbuton system Engneerng 2 ed edton, CRC press by 28 [7] K. A. Brt, J. J. Graffy, J. D. McDonald, and A. H. El-Abad, Three phase load flow program, IEEE Trans. Power Apparat. Syst., vol. PAS.95, pp , Jan./Feb [8] D. Shrmohammad, H. W. Hong,et.al A compensaton- based power flow method for weakly meshed dstrbuton and transmsson networks, IEEE Trans. Power Syst., vol. 3, pp , May 988. [9] G. X. Luo and A. Semlyen, Effcent load flow for large weakly meshed networks, IEEE Trans. Power Syst., vol. 5, pp , Nov. 99. [] C. S. Cheng and D. Shrmohammad, A three-phase power flow method for real-tme dstrbuton system analyss, IEEE Trans. Power Syst., vol., pp , May 995. [] W. M. Kerstng and L. Wlls, Radal Dstrbuton Test Systems, IEEE Trans. Power Syst., vol. 6, IEEE Dstrbuton Plannng Workng Group Rep., Aug. 99. [2] M. E Baran and F. F. Wu, Optmal Szng of Capactors Placed on a Radal Dstrbuton System, IEEE Trans. Power Delvery, vol. no., pp. 5-7, Jan [3] M. E. Baran and F. F. Wu, Optmal Capactor Placement on radal dstrbuton system, IEEE Trans. Power Delvery, vol. 4, no., pp , Jan [4] M. H. Haque, Capactor placement n radal dstrbuton systems for loss reducton, IEE Proc-Gener, Transm, Dstrb, vol, 46, No.5, Sep [5] R.Bllnton and J. E. Bllnton, Dstrbuton system relablty ndces,ieee Trans. Power Del., vol. 4, no., pp , Jan [6] IEEE Standards, IEEE Gude for Electrc Power Dstrbuton Relablty Indces, IEEE Power Engneerng Socety. [7] Roy Bllnton and Ronald N.Allan, A Text Book on Relablty Evaluaton of Power Systems 2 nd Edton, Plenum Press, New York and London. Author s detal: Mr. D.Mural Currently pursung M.Tech at Sr venkatesa Perumal College of engneerng and Technology, Puttur and Obtaned hs B.Tech n Electrcal and Electroncs Engneerng from JNTU Unversty at Audsankara College, Gudur. Hs area of nterest power systems, operaton and control, dstrbuton systems and applcaton of FACTS devces n Transmsson systems. Mr. A.Hema sekhar currently workng as Assocate Professor & Head of the Dept. of Electrcal and Electroncs Engneerng, S.V.P.C.E.T,Puttur and Obtaned hs B.Tech n Electrcal and Electroncs Engneerng from JNTU,Hyderabad, at Sree Vdyanketan Engneerng College, Rangampet; M.Tech (PSOC) from the S.V.Unversty college of Engneerng,Trupat. Hs area of nterest power systems, operaton and control, dstrbuton systems, electrcal machnes and Power System Stablty. 2 Page