Probable Optimization of Reactive Power in distribution systems, in presence of distributed generation sources conjugated to network and islanding

Transcription

1 IOSR Journal of Electrcal and Electroncs Engneerng (IOSR-JEEE) e-issn: ,p-ISSN: , Volume 11, Issue 5 Ver. II (Sep - Oct 2016), PP Probable Optmzaton of Reactve Power n dstrbuton systems, n presence of dstrbuted generaton sources conjugated to network and slandng QolamrezaNematollah, EsmalKhallzadeh, MohammadrezaAghae, MojtabaBorhan 12Azad Unversty Branch of Marvdasht 34 AzerbajanPower Consultng Engneerng Company Abstract: In ths artcle Probable Optmzaton of Reactve Power n dstrbuton systems, n presence of dstrbuted generaton sources conjugated to network and slandng has been presented.for probable assessment of success operaton of mcro grds based on actve and reactve power qualty and consderaton of voltage lmtaton, a systematc and optmzed strategy for optmum placement of generaton resources has been presented. The problem has been solved by the taboo search optmzaton algorthm and consderaton of two objectve functon. Frst objectve functon s the annual energy loss and second one s a novel ndex n mcro grd successon assessment based on actve and reactve power qualty and voltage lmtatons. Then the reactve optmum power s desgned for reducton of annual energy loss of systems conjugated to network. By a case study n ths artcle t wll be shown that placement locaton and the amount of dstrbuted generaton resources have the consderable effect on mcro grds operatons. Keywords: Energy loss, mcro grd operaton, optmum placement, reactve power desgnng, taboo search. I. Introducton Dstrbuted Reactve Sources (DRS) used n dstrbuton networks for some reasons such as decreasng the energy loss, mprovng the voltage profle, rectfyng the power coeffcent and ncreasng the Capactance of system. Optmum desgnng of reactve power or especally the optmum locatng of capactor n dstrbuton networks, performed many tmes by the power engneers for dfferent ams such as peak power, energy loss, cost decreasng and mprovng the dstrbuton network relablty (Etemad and Fotuh-Fruzabad 2008). Dfferent methods suggested n ths doman, from the prevous analytcal methods (Mekhamer et al. 2002) to nnovatve and ntellgent methods (Dadkhah and Venkatesh 2012).Changng the dstrbuton system structure because of the mcro-grds entrance, probable nature of added dstrbuted productons and ther ablty n reactve power provdng, caused to understand the necessty of desgnng and new strategy n extracton for current dstrbuton systems.in present study, optmum locatng and measurng of old reactve sources wll revewed, consderng the novel system s crcumstances. For ths am, the problems of reactve power desgnng n dstrbuton systems wll formulated, consderng the dfferent sdes of modern dstrbuton systems. These propertes for reactve power desgnng have not been stated by Etemad and Fotuh-Fruzabad 2008; Mekhamer et al. 2002; Dadkhah and Venkatesh 2012n lterature.two objectve functon consdered n ths artcle. Frst am s mnmzng the annual energy loss n lnk to the network condton and second am s maxmzng the novel defned ndex for success performance n slandng stuaton at feedng-losng condton. The man part of ths artcle s about the researches of optmzed reactve power desgnng n modern dstrbuton systems. Problem Mathematcal Explanaton In ths secton, the objectve functon equatons wll explaned. In ths artcle, the DRS s reactve power presented n general form but they can consdered as the capactor banks that are so affordable. In stuatons that dstrbuted productons should be used along wth power electronc convertors or n stuaton that capactors are changeable the same equatons, algorthms and analyzes used. Only dfferences s that the DRS capactes for all load-producton states wll be varable. Fnally, the maxmum calculated capacty of DRSs n each stuaton assumed the optmum capacty of t. A. Energy Losses It s possble to mnmze the annual energy losses va the DRSs locaton optmzng n dstrbuton system. Mnmzng the energy losses can be an mportant goal for network desgners. In ths artcle, some cases such as varable characterstc of dstrbuted generatons and load profle wll be consdered for mnmzng the annual energy losses. Objectve functon defned as below: DOI: / Page

2 F 1 = Annual losses of energy N P h n1 Lossn n n Where N s the number of states n a year, P Loss n s the system losses for a dstnct tme from a year, ρ n s the related probablty state and h n s a tme part of related state that s consdered one hour here. Va the optmum locatng and measurng of DRSs, t s possble to reduce the total energy losses. B. Mcro Grd Success Index Mcro grds can used n slandng form to provde crtcal loads at outage of man feedng source and mprove the dstrbuton system relablty. One crcumstance for success performance n slandng state s as below: )1( P P P DG Load Loss )2( Where P DG s the generated power by the mcro grd DGs, P load s the load power n sland and P loss s the power losses ofthe sland and assumed that s 5% of current load (Jan, Sngh, and Srvastava 2013). It assumed that bomass dstrbuted generaton unts are programmable and s the second crcumstance of success operaton of mcro grd (Plo, Psano, and Soma 2011). P BM 0.6 P DG Where P BM s the generated power by the Bomass DGs n mcro grd. Equaton 2 and 3 are the necessary crcumstances for the success performance of mcro grd. There are 365*3456 state wth dfferent probablty for each state. Calculatng the defned ndex for each mcro grd, each state calculate separately and zero or one gven to each one. )4( 1 PG PL & QG Q L G & Vmn V Vmax 0 otherwse Where G shows the total generaton and L s the total actve and reactve power consumpton n state, V shows the all mcro grd buses n state. after the ndex calculaton for each state, usng the each state probablty, they can sum as below: N G G 1 Where ρ s the state probablty. Calculated ndex n equaton 5 s just for a mcro grd. For a system wth many mcro grds the success ndex obtaned by equaton 6. )3( )5( F 2 NoM Gk k 1 NoM k 1 N N Lk Lk Where NoM and N LK are the number of mcro grds and amount of mcro grd load respectvely. )6( C. Compound Objectve Functon Two objectve functon can be compounded and generate an optmzaton objectve problem: mn( F), F K F +K (1-F ) It should consdered that F 1 and F 2 should be unt and F 2 modfed to 1-F 2. K1 and K2 determne that objectve functon can be annual energy losses, success ndex of a mcro grd or a compound of them. )7( DOI: / Page

3 Constrants of problem Constrants of problem contans the below cases: 1) load part equatons P P P P Subt DGt DESRs Loadt nbus V V Y cos( ) j, t 1 1 t, t, j, j j t, j t, Q Q Q Q Subt DGt DESRs Loadt nbus V V Y sn( ) j, t t, t, j, j j t, j t, 2) Voltage lmtatons V V V 1 mn t, max 3) Feeders capactance lmtatons P P subt rated 4) dscrete values Q =k Q DRS_ Q_ Step 5) Penetraton amount of DRS n each Bus Q Q DRSs_ max 6) Total reactve power capacty (Q DRS ) Q Q DRSs_ DRS )8( )9( )10( )11( )12( )13( )14( Problem soluton algorthm Taboo search TS (Guerrero et al and Nunna ; Doolla 2013) s an algorthm based on teraton that uses dfferent memory structures for the optmzaton problem solvng. Optmzaton soluton s vector. Its row conssts of consdered buses for DRS nstallaton as below: Q )15( DRS QDRS_ 1 QDRS_ k Q DRS_ N C TS starts by an expermental soluton and follows towards the neghborhood space. Test soluton s a vector smlar to Q DRS, whch conssts of DRS measurement n each system bus. Changng the parameters of ths vector, a set of vectors generated called neghborhood. Solve procedure s as below consderng the K 1 and K 2 amounts. 1) IfK 1 0, F 1 should calculated n connected mode to the network. 2) IfK 2 0, F 2 calculated for system n slandng state, then the success ndex calculated va the 4 to 6 equatons. Calculatng the F 1 and F 2, objectve functon F, calculated for all test solutons n neghborhood and then the search process keep gong untl the fndng best neghborhood. Implementaton and Senstvty Analyss Studes PG&E 69 bus dstrbuton system used for the algorthm mplementaton and senstvty analyss (Tent et al. 2012). Actve and reactve modfed power shown n fgure 1. In general state the energy losses s equal to MWh. The optmum locaton and named capacty of DGs presented n table 1. DOI: / Page

4 Fg1. The 69 Bus Dstrbuton System Loads Table1. Optmum selected Buses for the DGs nstallaton DG Type Basses Nomnal Capacty(kW) Wnd Turbne PVModule Producton of bomass dspersed Mcrogrd Table2. Mcro grds wth optmzed desgn Mcrogrd Basses )kva( Peak Tmes /0+j304/ /6+j178/ /0+j /8+j171/ /9+j280/ /6+j129/0 Producton )kva( 225/0+j41/1 125/0+j24/7 225/0+j41/1 125/0+j24/6 200/0+j41/1 100/0+j24/6 Addng the DGs to the system, the energy losses reduced to 90.58MWh. Then system s dvded nto sx vrtual mcro grd to mnmze the actve and reactve power unbalance (Chaouach et al. 2013).Table 2 shows these sx mcro grds and ther related propertes. Fgure 2 shows the system mono lnear dagram. Consderng the annual energy losses and success ndex as the objectve functon, the DRSs located n system as optmzed form. Fg2. The 69 Bus dstrbuton system wth locaton of dstrbuted generatons DOI: / Page

5 Total Capacty )kvar( Tabl3. DRS, ther locatons and capactes n stat A. Basses DRS Nomnal Capacty(kVAr) Annual losses of )MWh(energy 78/93 72/43 66/88 62/28 58/59 55/49 52/07 51/07 A. Objectve functon of Annual energy losses (ndvdually) In ths part, the annual energy losses consdered as the objectve functon only to fnd the optmum locaton of DRSs n dstrbuted system. It assumed that total capactydetermned from 100 KVAr to 450 KVAr wth 25 KVAr steps as the optmzed form to mnmze the annual energy losses. DRSs amount and ther locaton shown n table three. It reveals that annual energy loss decrease from MWh to MWh n response to DRS ncreasng.fgure 3 shows the DRS locaton n stuaton that total amount of DRS s equal 350 KVAr. In ths state, the Buses voltages modfed as P.U. [ ] for all states, addng the DRSs to the system. After one year, t changed to [ ] P.U. Fg3. DRSs Locaton n State A. B. objectve functon of mcro grd success ndex (ndvdually) In ths part, only the success ndex of mcro grd consdered as the objectve functon to fnd the optmum locaton of DRSs n dstrbuton system. In ths part assumed that the total determned capacty, KVAr, used n dstrbuton system for optmzaton to success ndex be maxmum. Table 4 shows the Success Index n the entre system n stuaton that DRSs located n system as optmum form. It can be seen that Success Index ncreases from 23.88% to 94.9%. Fgure 4 the DRS locaton has been shown n stuaton that total amount of them s 350 KVAr. Total Capacty )kvar( Table4. DRSs, ther locaton and capactes n stuaton B. Basses DRS Nomnal Capacty(kVAr) Annual losses of )MWh(energy 23/88 39/47 54/12 65/98 71/49 DOI: / Page 80/79 88/11

6 /90 Fg4. Locaton of DRSs n stuaton B. C. consderng both objectve functon In ths part both objectve functon are consdered together. For ths am, assumed that entre amount of DRSs equal to 350 KVAr. Ths works progressvely wth K 1 ncreasng and K 2 decreasng. The objectve functon n ths stuaton wll be as equaton 16: F F K1 F1 Opt F K2 1 1Opt 1 F2 1 F2Opt 1 F2 Opt 100 Table 5 shows the results of optmum locatng probablty of DRS n power system. Fgure 5 shows the DRS locaton n stuaton that total amount of them s 350 KVAr. )16( Fg5. DRSs locaton, consderng the both objectve functon DOI: / Page

7 K 0 0/1 0/3 0/4 0/5 0/6 0/7 0/ /9 0/7 0/6 0/5 0/4 0/3 0/1 0 K Table5. DRSs, ther locaton and capactes n stuaton C. Basses DRS Nomnal Capacty(kVAr) The objectve functon(%) 0/0 0/0225 0/1076 0/0927 0/1582 0/1761 0/1502 0/1495 0/0 II. Concluson In ths artcle, a systematc and optmzed strategy presented for DRSs optmum locatng n mcro grds.the TS optmzaton algorthm, consderng the two objectve functon, has solved the problem. The frst one s the annual energy loss and second one s the success ndex based on actve and reactve power qualty n addton to voltage lmtatons. In ths artcle, assessng the some case studes, t has been shown that DRS locaton and amount have the sgnfcant effect on mcro grd successful operaton. Therefore, ths problem should be revewed for temporary part tme dstrbuted generaton unts. References: [1]. Chaouach, Aymen, Rashad M Kamel, Rdha Andouls, and Ken Nagasaka 'Multobjectve ntellgent energy management for a mcrogrd', IEEE Transactons on Industral Electroncs, 60: [2]. Dadkhah, Maryam, and Bala Venkatesh 'Cumulant based stochastc reactve power plannng method for dstrbuton systems wth wnd generators', IEEE Transactons on Power Systems, 27: [3]. Etemad, AH, and M Fotuh-Fruzabad 'Dstrbuton system relablty enhancement usng optmal capactor placement', IET generaton, transmsson & dstrbuton, 2: [4]. Guerrero, Josep M, Juan C Vasquez, José Matas, Lus García De Vcuña, and Mguel Castlla 'Herarchcal control of droopcontrolled AC and DC mcrogrds A general approach toward standardzaton', IEEE Transactons on Industral Electroncs, 58: [5]. Jan, Naveen, SN Sngh, and SC Srvastava 'A generalzed approach for DG plannng and vablty analyss under market scenaro', IEEE Transactons on Industral Electroncs, 60: [6]. Mekhamer, SF, ME El-Hawary, SA Solman, MA Moustafa, and MM Mansour 'New heurstc strateges for reactve power compensaton of radal dstrbuton feeders', IEEE Transactons on Power Delvery, 17: [7]. Nunna, HSVS Kumar, and Suryanarayana Doolla 'Multagent-based dstrbuted-energy-resource management for ntellgent mcrogrds', IEEE Transactons on Industral Electroncs, 60: [8]. Plo, Fabrzo, Gudtta Psano, and Gan Guseppe Soma 'Optmal coordnaton of energy resources wth a two-stage onlne actve management', IEEE Transactons on Industral Electroncs, 58: [9]. Tent, Paolo, Alessandro Costabeber, Paolo Mattavell, and Danela Trombett 'Dstrbuton loss mnmzaton by token rng control of power electronc nterfaces n resdental mcrogrds', IEEE Transactons on Industral Electroncs, 59: DOI: / Page