Reliability Based Planning Methodology for Feeder Automation

Transcription

1 Relablty Based Plannng Methodology for Feeder Automaton TB Mavuso, A Wessels, JHC Pretorus, Faculty of Engneerng and the Bult Envronment Unversty of Johannesburg Johannesburg, South Afrca jhcpretorus@uj.ac.za Abstract One of the most used measures for mprovng servce relablty s dstrbuton automaton (DA). The DA technology has been wdely used by dfferent utltes for qute some tme. Its applcablty on the Eskom dstrbuton network s an opton but t s stll n the research and development stage. Ths paper proposes a method for a network plannng engneer to motvate the use of feeder automaton (FA). Ths methodology wll serve as a gudelne to a network planner nterested n usng FA as an alternatve or addtonal soluton to mprove the relablty of a feeder. The defned method s based on lterature and exstng research on FA. Furthermore the methodology s analyzed usng the Eskom case study. A comparson s made by dentfyng a pattern of steps or methods followed n plannng for Magalesburg-Hekpoort feasblty project and comparng ths method wth the proposed methodology. Index Terms Power system relablty, power dstrbuton, feeder automaton, dstrbuton automaton. I. ITRODUCTIO In the electrc dstrbuton ndustry, servce qualty and relablty are of utmost mportance. There has been ncreased pressure on electrcty utltes to decrease outage tme due to faults. The System Average Interrupton Duraton Index (SAIDI) s commonly used as a relablty ndcator by electrc power utltes. SAIDI s the average outage duraton for each customer served, and s calculated as: CustomerInterruptonDuratons SAIDI = = TotalumberofCustomerServed Where s the number of nterrupted customers, r s the annual restoraton tme for nterrupton event and T s the total number of customers served. In other words, SAIDI s measured n unts of tme, often mnutes or hours. It s usually measured over the course of a year, and accordng to the IEEE Standard. r T (1) SAIDI forms a sound bass for measurng performance and provdes a clearer vew of performance, both on a daly bass and durng major events. SAIDI can also form a sold bass for the revew of operatonal effectveness, decson makng and polcy makng and provdes a more consstent benchmark [1]. The system average nterrupton frequency ndex s desgned to gve nformaton about the average frequency of sustaned nterruptons per customer over a predefned area [1]. SAIFI=Total number of customer nterruptons/total number of customers served SAIFI umberofcustomerinterruptons = TotalumberofCustomerServed = (2) It has been reported that most of the customer servce nterruptons are due to falures on the medum voltage (MV) electrcal dstrbuton network. In order for the electrcty delvery ndustry (EDI) to ensure a relable electrcal network, several solutons such as reducng the lengths of feeders, buldng back-up feeders and ncorporatng mproved protecton systems are appled to MV electrcal networks [2]. Currently one of the most used measures for mprovng servce relablty s dstrbuton automaton (DA). The DA technology has been wdely used by dfferent utltes for qute some tme. Its applcablty on the Eskom dstrbuton network s stll n the research and development stage [3, 4]. DA s defned as an ntegrated systems concept for MV networks and therefore entals the ntegraton of several technologes. Feeder automaton s one of the functons of DA. It entals the use of remote swtches to solate the faults and restore power to the remanng un-faulted sectons of the feeder. It s defned and chosen by most utltes as one of the best DA functons for mprovng the relablty of the feeder [2, 5]. The atonal Energy Regulator of South Afrca has also espoused a reward/penalty system to measure the effcency of EDI for outage tme [6]. Dstrbuton utltes are penalzed on T

2 tme (hrs.) when power s not avalable to the customers. Ths has resulted n Eskom beng the man dstrbuton utlty company to embark on methods amed at reducng outage tme through mprovng the relablty of the network and consequently mprovng servce delvery. Eskom has therefore prepared a busness case study to test the possblty of mplementng FA. The study ndcates that the mplementaton s possble and t can be used as a measure for mprovng the relablty of the MV network [7]. There s therefore a need to develop a method or gudelne to plan for FA. II. PROBLEM STATEMET AD OBJECTIVE A. Problem statement From a network plannng perspectve nformed decsons on solutons that wll mprove the relablty status of the network have to be made. Decsons are based on techncally sound, cost effectve and sustanable solutons that address a need. It has to ensure that the soluton to be mplemented wll reduce customer exposure to lengthy network faults and thus mprove the relablty of the network. In ths regard there are methods and gudelnes to gude a planner toward the best soluton. There s no gudelne or procedure for ths new technology to assst a network plannng engneer, to plan and motvate the use of FA technology as an alternatve soluton to the conventonal solutons a plannng engneer would use for mprovng the relablty of the network. There s therefore a need to develop a relablty based plannng methodology that wll assst and gude a network plannng engneer to plan FA, ncorporatng FA as an optmum alternatve soluton for mprovng relablty of the MV network consdered. B. Objectve The objectve of ths research s to develop a method on how a network plannng engneer should motvate for the use of FA. Ths methodology wll serve as a gudelne to a network planner nterested n usng FA as an alternatve to conventonal plannng solutons, lke constructon of a new feeder or upgradng of the conductor and convertng a feeder from 11kV to 22kV to mprove the relablty of a feeder. The objectve s also to develop a relablty based method ntended to assst a planner to best plan and nvestgate MV networks for the applcaton FA. Ths methodology wll assst network plannng engneers to acheve the followng: To know the key aspect and prncples of network and relablty plannng that a plannng engneer must consder when nvestgatng a feeder for the applcaton of FA, and how to algn these prncples n a systematc order to substantate the proposed soluton. To understand the mportance and mpact of resource selecton (selecton of technology to be used) and determne the poston of swtches, thereby motvatng the selecton of FA as the applcable soluton. III. RELIABILITY BASED PLAIG METHOD FOR FEEDER AUTOMATIO The research methodology for ths study s based on the Eskom feasblty study for FA and the lterature revews of: () FA technology and ts operatons phlosophes; () exstng plannng methods from dfferent utltes () and power system relablty theory. The Eskom feasblty study and outcomes thereof are dscussed n bref below. The proposed methodology s then presented. A. Eskom dstrbuton automaton strategy Magalesburg- Heekpot 11kV network feasblty study The objectve of the project was to desgn and mplement a FA plot system. The FA systems were commssoned and ther mpact on the network relablty and avalablty were montored to develop a busness case for FA technology n Eskom. The Magalesburg-Hekpoort 11kV feeder was proposed as the canddate ste for the FA plot system as t s a dense rural feeder wth a poorer than average performance. The proposed network was analyzed n terms of the predefned selecton crtera for FA systems and t passed all the requrements. The Magalesburg-Hekpoort FA system was successfully commssoned on 9 December The system has operated as expected and the fve sustaned faults, whch occurred durng the analyzed perod (9 December January 2012), were reported. The FA system resulted n sgnfcant techncal mprovements for the analyzed perod. The followng summary of results was reported: A SADI of 6.4 hrs. was saved over a perod of roughly 2 months. The Customer Average Interrupton Duraton Index (CAIDI) of 3.48 hrs wth FA was recorded compared to 4.8 CAIDI hrs wthout FA. 3 hrs. 39 mn was the average tme taken to restore all loads. For the study perod the sum of customers that were nterrupted wth FA was reported to be 228 compared to 429 customers wthout FA operatng on the feeder. B. etwork plannng prncples consdered from the Magalesburg-Heekpoort 11kV network feasblty study The followng are the key network plannng prncples that were consdered and used as gudng prncples for assessng the network for FA. Selecton crtera that the canddate networks had to meet n order to be consdered for the FA feasblty study. These selecton crtera consdered the followng prerequstes : etwork confguraton: - network classfed as rural/urban or sem, total load 5 MVA, possblty to nterconnect the network to enable back feedng and f the network s cable or overhead.

3 etwork locaton: - to determne how long t wll take to attend to the fault and have a safe workng envronment. etwork performance: - performance hstory and event record should be for a perod of at least 3 years and a sgnfcant number of permanent faults. Tele control and telecoms: - more than one ntegrated recloser that s remotely operated and Eskom rado coverage and cellphone coverage. Customer mnutes saved, number of customers saved due to automatc restoraton of healthy sectons of the feeder and MVA hours saved are ndcators that were consdered as factors determnng the performance of the Magalesburg-Hekpoort network wth FA technology. They were also used to set the performance targets of the network wth and wthout FA technology. preferred soluton to mprove relablty. References [2, 6, 7, 8] are some of the research papers that were revewed to defne the proposed methodology. A network plannng method can be defned n the scope of automaton only,.e. ts focus can be based on the plannng of the poston of automated swtches, te swtches and the number of swtches, as these three mportant factors play a crucal role n provng how the relablty of a dstrbuton feeder can be mproved. However, t s mportant to consder or ncorporate the exstng plannng crtera nto the proposed method as these crtera wll assst a network plannng engneer to use good judgment. Fgure 2 contans a flow dagram of the proposed methodology. C. Revew of the Magalesburg-Hekpoort FA feasblty study From the case study t was noted that there was an approach or methodology used to mplement FA. The approach was not a strategc approach or a prelmnary defned approach, snce the man am of the case study was to determne the techncal and fnancal feasblty of DA technology. It s apparent that pror to mplementng FA technology a lot of plannng had to be done, partcularly to understand the power system and operatonal dynamcs of the network. Ths ncluded developng crtera for the selecton of feeders sutable for FA mplementaton, settng targets to determne the performance of the FA soluton chosen and selectng ndcators to determne how effectve the adopted soluton s. The network plannng methodology of the Magalesburg- Hekpoort FA feasblty study s schematcally represented n Fgure 1 below. Fgure 1. Magalesburg-Hekpoort network plannng methodology for feeder automaton D. Proposed relablty based plannng method for feeder automaton The proposed plannng method s prmarly defned on exstng methods used by utltes globally. Research was done to understand how other utltes plan and motvate for FA as Fgure 2. Relablty based plannng method for feeder automaton E. Buldng blocks of the proposed method ETWORK DESCRIPTIO Ths entals the electrcal characterstcs of the network e.g. (load flow studes, back feedng capacty, load growth etc.) and physcal characterstcs e.g. (length of the feeder and

4 T-offs, radal or rng feed, accessblty of the feeder, rado and cellphone sgnal etc.). CUSTOMER IFORMATIO Ths s nformaton relatng to the type of load suppled from the feeder e.g. ndustral, resdental farmng etc. It also entals the poston of customers on the feeder and number of customers. Ths nformaton s essental for understandng that the load s dstrbuted on the feeder, settng performance targets and also to determne the power system relablty ndces lke SAIDI. COST OF ITERRUPTIO PRIOR AD RELIABILITY PRIOR The cost of nterrupton s a functon of outage duraton and load type percentages. It determnes cost of losng supply due to a faulty feeder. After automatng, the feeder cost of the nterrupton can be separated per secton and the cost of nterrupton due to an outage at a certan secton can be calculated for the entre feeder. FA REQUIREMETS AD FEEDER PRIORITISATIO It s not all feeders that wll beneft from FA and not all feeders wll show a postve nvestment. Therefore there s a need for feeders to have some characterstcs that wll ease the mplementaton and postvely reveal the mpact of FA. AUTOMATIO RESOURCE SELECTIO Due to the ablty of the feeder to meet the above requrements, t could be determned whch FA archtecture wll be the best to mprove the SAIFI and SAIDI of a feeder, as the cost savngs and nvestment returns wll also ndcate whch wll be the best archtecture to apply. AUTOMATIO RESOURCE ALLOCATIO Determnng the poston of sectonalzng swtches, automated swtches and te swtches on the feeder plays a crucal role for effectve usage of FA n locatng and solatng faults as well as restorng power to healthy sectons of the feeder. IV. THE AALYSES OF DISTRIBUTIO FA PLAIG METHODOLOGY In feeder prortzaton t was suggested that to fully come to a concluson or be able to select the sutable feeder for FA, data on the physcal and electrcal problems of the network, the number and type of customers on the network, the cost of nterrupton, the performance of the network n terms of the targeted SAIDI and SAIFI, and readness of the feeder for mplementaton of automaton need to be crtcally analyzed. Based on the fndngs one can motvate the use of FA to archve or solve related problems and also decde to what extent FA can be appled. For example, full or sem automaton of sectonalzng swtches. SELECTIO CRITERIA OF ETWORK FEEDERS suggest that to select a sutable feeder for FA, data on etwork Confguraton (whch nvolves classfcaton of customers type, electrcal load on the feeder and the ablty to nterconnect physcally or back feed from other feeders), etwork Locaton (whch concerns the geographcal poston of the feeder) etwork Performance (whch concerns the type of faults.e. permanent or temporary faults, the number of swtches to be used and as well fault dstrbuton whch wll determne the poston for swtches) and Telecommuncaton (whch entals the ease of FA functonng due to enough cell phone and rado coverage as well the fnancal ease of mplementng FA because of some exstng equpment) must be readly avalable. AUTOMATIO RESOURCES LOCATIO: Ths suggests that to fnd the optmum locaton or poston for sectonalzng swtches, automated swtches and te swtches the followng can be used as gudng factors: possble fault paths, load dstrbuton on the feeder, SAIDI and SAIFI and customer cost of nterrupton. The Eskom plannng methodology for FA, pre-defned above, suggests the use of fault occurrence.e. knowng where most faults occur on a feeder and classfcaton of faults.e. lookng specfcally at the type of faults occurrng on a feeder. Lookng at the two crtera above, t could be notced that they are completely dfferent. Therefore based on the prncples each crteron presents, t s clear that they wll reveal dfferent results f they are appled on the same feeder. Although they can be used together, t s suggested that only classfcaton of faults should be assessed to be the proposed gudng factor. Ths s the crucal part of the methodology as fndng the optmum poston for swtches wll demonstrate the beneft of usng FA technology and determne the extent to whch SAIDI and SAIFI wll be mproved. It s therefore very mportant to apply the FA plannng methodology to demonstrate ts effectveness. COMPARISO OF EVALUATIO CRITERIA. The Relablty Based Plannng Methodology (RBPM) for FA plannng methodology suggests that for evaluatng the performance of FA, cost of nterrupton wth FA, cost of automaton resources and network relablty wth FA must be evaluated. Therefore SAIDI and SAIFI wll be compared to the cost of resources used. The cost of nterruptons wth FA wll then be compared to the cost of automaton resources. From the Eskom Magalesburg-Hekpoort case study the followng were consdered for evaluatng the performance: a) Customer mnutes saved. b) umber of customer nterruptons saved due to the automatc restoraton of the healthy sectons of the feeder. c) MVA hours saved. The FA plannng methodology measures ts strategy for performance evaluaton aganst the cost of the beneft and the cost of penaltes saved, whle the Eskom case study focuses mostly on the tme t takes to restore the supply to customers and the number of customer nterruptons saved. Incorporatng the Eskom performance and evaluaton crteron nto FA plannng methodology wll gve the latter more results and mproved effectveness on evaluatng the performance, benefts and cost of the FA plannng

5 methodology, because t s mportant to consder the benefts of customers as the end users. V. DISCUSSIO AD COCLUSIO The am of the study was to develop a plannng method for dstrbuton FA ntended to assst a planner to know how to best plan or what to consder n plannng for mprovng relablty of the network usng FA. Ths was acheved through revewng the lterature on exstng plannng methods, deas and case studes for the applcaton of FA. The lterature revewed was also used to defne the proposed methodology. Furthermore the Eskom Magalesburg-Hekpoort case study was dscussed. It was noted from the case study that there were some of the network plannng aspects consdered durng the plannng stage of the case study. Ths led to the defnton of Magalesburg- Hekpoort network plannng methodology for feeder automaton. The followng key factors were noted: Hstorc Performance of the network: - ental the past performance of the network based on faults experenced. The data collected was used understand all ssues and confguraton of the network. Selecton and assessment of feeders: - ental the requrements requred for mplementng FA. These requrements were used for selectng feeders. Fault classfcaton and occurrence: - ental knowng the dstrbuton of faults on the feeder. The data was utlsed for determnng the poston of swtches. Evaluaton of mplemented FA soluton: - ental evaluatng the adopted soluton. Customer mnutes saved, number of customer nterruptons saved and MVA hours saved, were the ndcators used for assessment. The proposed method hghlghts three mportant aspects to consder when plannng for FA. These are: Selectng feeders sutable for FA technology applcaton. Knowng the prncples of network and relablty plannng that a plannng engneer must consder when nvestgatng a feeder for the applcaton of FA. Understandng the mportance and mpact of resource selecton (selecton of technology to be used) and the determnaton of the poston of swtches. The Relablty Based Plannng Method needs to be tested by performng the defned steps, so to demonstrate how practcal t s, n assstng a plannng engneer to motvate for the use of feeder automaton. Furthermore It s very mportant to understand that the resources.e. software, that wll be used to mplement some of the steps of the method wll play an mportant role n demonstratng the effectveness of the method. REFERECES [1] IEEE Standard IEEE Gude for Electrc Power Dstrbuton Relablty Indces, 1998 [2] J. Tenga & C. Lub. Value based dstrbuton feeder automaton plannng. ELSEVIER, Electrcal Power and Energy Systems 28 (2006), pg ovember [3] J.B. Bunch, Gudelnes for evaluatng dstrbuton automaton, EPRI Report EL-3728, [4] R.P. Gupta, Member IEEE. Substaton Automaton Usng IEC61850 Standard. Ffteenth atonal Power Systems Conference (PSC), IIT Bombay, December [5] M.R. Gouvea, D.P. Duarte, A.P. Cunha & I.T. Domngues Automaton as resources for dstrbuton systems plannng. CIRED 19th Internatonal Conference on Electrcty Dstrbuton. Venna, May 2007, paper 0640, pg [6] T. Kleynhans & C.G. Cater Brown. etwork Plannng Relablty Gudelne DGL Eskom Dstrbuton, February 2011, pg [7] J. van der Merwe & R.G. Coney. Eskom Dstrbuton Automaton Strategy Report. Eon Consultng, Report o: RES/RR/10/ [8] D. Gutschow & M.O. Kacheng a Makng busness sense of dstrbuton automaton systems: The case of Eskom, South Afrca. IEEE Transactons on Power Systems. Vol 20. o 1, pg , February 2005.