Short-term Resource Scheduling for Power systems with Energy Storage Systems

Size: px

Start display at page:

Download "Short-term Resource Scheduling for Power systems with Energy Storage Systems"

Tracey Bell
5 years ago
Views:

1 1 Shor-erm Resource Scheduling for ower sysems wih Energy Sorage Sysems Se-Hwan Jang, Jong-Bae ar, Member, IEEE, Jae Hyung Roh, Member, IEEE, Sung-Yong Son, Member, IEEE, Kwang Y. Lee, Fellow, IEEE Absrac--Energy Sorage Sysems (ESS) could be widely available in power sysems, in which case he shor-erm scheduling is one of he essenial problems ha need o be addressed. This paper presens a resources scheduling mehod for he opimal operaion plan wih he energy sorage sysems, which uilizes he Mixed Ineger rogramg (MI) based on he branch bound mehod. The proposed mehod is evaluaed on wo case sudies. One sudy exaes a daily opimal operaional plan for power supply resources energy sorage sysems in an isolaed micro-grid sysem. Anoher exaes he opimal operaional plan in a convenional power sysem wih a large-scale energy sorage sysem. Index Terms Shor-erm scheduling, Energy Sorage Sysems, Renewable Energy Sysem, Mixed Ineger rogramg. E I. INTRODUCTION nergy Sorage Sysems (ESSs) are being raised as imporan resources. They have he poenial o exend opimize he operaing capabiliies of he grid since power can be sored used a a laer ime. This allows for flexibiliy in generaion disribuion, improving he economic efficiency uilizaion of he enire sysem while maing he grid more reliable robus. Addiionally, as alernaives o radiional power generaion, including variable wind solar energy echnologies may require bac-up power sorage. The mos negaive drawbac of curren ESS echnology is is high cos. To overcome i, various researches are in progress [1]. The EES could also be used for various applicaions [2, 3], such as for sabiliy enhancemen of he renewable energy sysem [4, 5], for frequency conrol in an isolaed power sysem [6, 7], for load leveling [8]. Shor-erm scheduling is one of he essenial problems for power sysem operaion. I is composed of uni commimen (UC) economic dispach (ED). I has a significan impac on secure economic operaion of power sysems. The UC problems involve scheduling he on/off saes of generaing unis, which imizes he operaing cos, sar-up cos shu-down cos for a given planning horizon under various S.-H. Jang, J.-B. ar, J. H. Roh are wih he Deparmen of Elecrical Engineering, Konu Universiy, Seoul , Korea ( {shang, baepar, hroh}@onu.ac.r). S.-Y. Son is wih he Deparmen of Energy IT, Kyungwon Universiy, Seoul, Korea ( xra@yungwon.ac.r). K. Y. Lee is wih he Deparmen of Elecrical Compuer Engineering, Baylor Universiy, Waco, TX 76798, USA ( Kwang_Y_Lee@baylor.edu). operaing consrains [9]. In he UC problem, he decisions are he selecion of he ime for each uni o be on- /or off-line (binary variables) as well as he uni s economic generaion level (coninuous variables). Thus, he UC problem can be formulaed as a nonlinear mixed ineger combinaorial opimizaion problem [10]-[15]. Over he pas decades, many salien mehods have been developed for solving he UC problems using mehods such as Lagrangian relaxaion [11], geneic algorihms [12], simulaed annealing [13], paricle swarm opimizaion [14], so on. Recenly, he UC problems have also included new elemens, such as ESSs virual power plans (Vs) [15, 16]. This paper describes a resource scheduling mehod for opimal operaion of power sysems uilizing he energy sorage sysems. The mehod employs he Mixed Ineger rogramg (MI) based on he branch bound mehod. The proposed mehod is evaluaed on wo case sudies. One sudy exaes a daily opimal operaional plan for power supply resources energy sorage sysem in an isolaed micro-grid sysem. Anoher exaes he opimal operaional plan in a convenional power sysem wih a large-scale energy sorage sysem. The simulaion is performed in he commercial MATLAB pacage program. II. MODELING FOR RESOURCES LANNING A. Obecive Funcion for Sysem Operaion The obecive he resource planning is o imize he fuel cos, sar-up cos, shu-down cos, operaing cos on all resources during a planning horizon, while saisfying all sysem consrains. The oal coss are imized by economically commiing dispaching he unis. Therefore, he obecive funcion of he resource planning problem is given by he imizaion of he following cos funcion: T N [ F (, ) u, + SU, (1 u, 1) u, ] (1) = 1 = 1 T is he number of scheduling period, N is he number of generaing unis, u, is he on/off saus of uni a hour (i.e., u, = 1 when uni is on-line, u, = 0 when uni is off-line). The main cos, F (, ), is he fuel cos funcion of uni a hour, which can usually be expressed as a second-order polynomial as follows: /12/$ IEEE

2 2 F = + (2) 2 (, ) a + b, c,, is he power generaion of uni a hour a, b, c are he cos coefficiens of uni. Sar-up cos for resaring a de-commied generaing uni, which is relaed o he emperaure of he boiler, is included in he obecive funcion. The sar-up cos is associaed wih he number of hours during which he uni has been off. Sar-up cos will be high, defined as he cold cos ( SU C, ), when down ime duraion exceeds he cold-sar hour ( T cold, ) in excess of he imum down ime; will be low, defined as as he ho cos ( SU H, ), when down ime duraion does no exceed he cold-sar hour in excess of he imum down ime. In general, he sar-up cos is described as follows: SU, SU = SU H, C, off-line a hour if if MDT TOFF TOFF,, > MDT MDT + T + T cold, cold, (3) TOFF, is he duraion for which uni has remained MDT is he imum down-ime of he -h uni. Shu-down cos is usually modeled as a consan value for each uni per shudown. In his paper, he shu-down coss have no been aen ino consideraion. ESS, is he oupu of ESS a hour, D is he oal sysem dem a hour Re, is he oal of renewable energy a hour. 2) Sysem spinning reserve consrains: Spinning reserve mus be provided so as o imize he probabiliy of load inerrupion. The spinning reserve is considered o be a pre-specified amoun or a given percenage of he forecased pea dem. Spinning reserve can be specified in erms of excess megawa capaciy, which is expressed by N = 1 ', u, D + SR (5) ' D D ESS, SESS, = + (6), is he imum oupu of -h uni u, is ' he sae of -h uni a hour ; D is he oal sysem load a hour considering he oupu of ESS SR is he required spinning reserve a hour ; S, is he sae ESS (charge/idle/discharge) of ESS a hour. 3) Generaion limi consrains: The power produced by each uni mus be wihin is limis as indicaed below: u u (7),,,,,,, are respecively he imum imum generaion limis of uni. 4) Minimum up-ime/down-ime consrains: The uni canno be urned on/off immediaely once i is commied or de-commied. The imum up/down ime consrains indicae ha a uni mus be on/off during a cerain number of hours before i becomes shu-down or sar-up, respecively. These consrains are given by 1 if 1 TON, 1 < MUT u, = 0 if 1 TOFF, 1 < MDT (8) 0 or 1 oherwise Fig. 1. Sysem configuraion. B. Sysem Resource Consrains 1) Load balance consrains: For each ime of he considered period, he oal summaion of real power of he resources in he sysem mus be equal or greaer o he elecric dem resources. N = 1, u, + ESS, = D R Re, (4) TON, is he duraion for which uni remains on-line a hour MUT is he imum up-ime of uni. 5) Generaion ramp up/down consrains: Due o he mechanical characerisics hermal sress limiaions of a generaing uni, he operaing range of all online unis is resriced by heir corresponding ramp-rae limis as follows: RD RD,,( 1) RU (9) RU are he ramp-down ramp-up limis of uni, respecively. 6) Charging/discharging consrains of energy sorage sysem: The energy sorage sysem has a imum charging or

3 3 discharging rae which is similar o he ramp rae of hermal generaion unis. The charging or discharging power canno be over his limiaion.,, 1 Ch (10) Ch is he imum charge/discharge rae of he -h ESS. I would be called he C-raes. 7) Generaion consrains of energy sorage sysem: The energy sorage sysem has a imum imum generaion consrain such as he hermal generaion unis., ESS, (11) ESS, ESS,, ESS, ESS, are respecively he imum imum oupu for he -h ESS. 8) Iniial/ending sae of charge () consrains of energy sorage sysem: To solve he uni commimen economic dispach problem wih energy sorage sysems, he iniial ending sae of charge is needed. This means ha how much of energy is in he sorage sysem when i is firs conneced o he sysem. =,, = 0 ini (12), T, fin = = (13) is he sae of charge of -h ESS,,, ini fin, are respecively he iniial final sae of charge of he -h ESS. 9) Sae of charge () consrains of energy sorage sysem: The energy limi of he energy sorage sysem is measured by he sae of charge. The ESS has a imum imum sae of charge consrain. For real circumsances, he sorage sysems canno fully charge heir capaciy for he efficiency of he sysem will lower he performance of i. The opimal range of he sae of charge for he sorage sysems are abou 90% imum 20% for he imum his consrain will be formed as bellow., (14), are respecively he imum imum sae of charge for he -h ESS. 10) Oupu consrains of renewable energy resources: The generaed oal renewable energy represens he sum of renewable resources such as phoovolaic, wind, ec. If here is a limi on he renewable oupu, i uses he imum value of he wo: Limi = Min, ) (15) Ren, ( R, R, R, pv, + w, = (16) R, is he sum of renewable oupus. pv, w, are he oupu of he phoovolaic (V), wind power, respecively. They are as follow: Ga M NOTC 20 = [ Max,0 + π ( Ta + Ga T, 0)] (17) G 800 pv, M a,0 0 Vw < Vci avw + bvw + cvw + dvw + e Vci Vw Vr w, = (18) r Vr Vw Vco 0 Vw > Vco G a, 0 is he insolaion, M Max, 0 is he imum power, T M,0 is he module emperaure a he sard condiion NOCT is he normal operaing cell emperaure of V. For wind power, r, Vci, Vco, Vr, Vw are he raed power, cu-in speed, cu-ou speed, raed speed, wind speed, respecively. III. SYSTEM ANALYSIS AND CASE STUDY Two case sudies are considered for scheduling shor-erm operaional plans for power sysems wih energy sorage sysems. One sudy exaes a daily opimal operaional plan for power supply resources energy sorage sysem in an isolaed micro-grid sysem. Anoher exaes he opimal operaional plan in a radiional power sysem wih a largescale energy sorage sysem. A. Isolaed Micro-Grid Sysem A daily operaional planning problem is considered for energy sorage sysem disribued generaors in an isolaed micro-grid sysem. The sysem has an energy sorage sysem, disribued generaors including 10 small generaors such as small gas urbines, diesel engines, micro urbines renewable energy resources such as phoovolaic sysem wind power sysem. The following ables represen he parameers on V, Wind, Energy Sorage Sysem. I is assumed ha here is no fixed charge required for he ESS. The oal capaciy of V sysem insalled is 1440W (4 360 Wp) oal capaciy of wind power insalled is 560W (4 140Wp). The spinning reserve for each hour is considered as 10% of he hourly load. TABLE 1. LOAD DEMAND FOR 24-HOUR. Hour Load [W] Hour Load [W] Hour Load [W] TABLE 2. V SYSTEM DATA. NOCT No of cells in a plan raed W μ Max Ⅹ

4 TABLE 3. WIND LANT DATA. r V ci V r V co a b c d e 140 3 15.01 17-0.01 0.33-0.9-2.1 7.1 TABLE 4. ENERGY STORAGE SYSTEM. Min [Wh] Max [Wh] C-rae iniial final 1250 2500 0.25 50% 50% TABLE 5.

4 4 TABLE 3. WIND LANT DATA. r V ci V r V co a b c d e TABLE 4. ENERGY STORAGE SYSTEM. Min [Wh] Max [Wh] C-rae iniial final % 50% TABLE 5. METROLOGICAL DATA. Hr G a T a V w Hr G a T a V w (m/s) (W/m2) (C) (m/s) (W/m2) (C) TABLE 8. OERATION LAN WITH ESS & RENEWABLES. TABLE 6. THERMAL UNITS DATA. Uni Uni1 Uni 2 Uni 3 Uni 4 Uni 5 Uni 6 Uni 7 Uni 8 Uni 9 Uni 10 (MW) (MW) a b c MUT (Hr) MDT (Hr) H C T (Hr) cold Iniial saus (Hr) In Table 7, i represens he daily opimal operaion plan of he disribued generaors. This case doesn consider he energy sorage sysem renewable sysem. Table 8 represens he daily opimal operaion plan in he isolaed micro-grid sysem wih he energy sorage sysem renewable sysem. Colored bloc indicaes he changed operaional plan due o he inroducion of energy sorage sysem renewable sysems. Green purple blocs are he effec of he energy sorage sysem renewable sysems, respecively. To compare wih exising sudies, he imum oupu of renewable energy resources is limied o 1,000W. Fig. 1 illusraes he hourly oal charge/discharge quaniy of he energy sorage sysem. Fig. 2. Hourly oal charge/discharge quaniy of ESS. TABLE 7. OERATION LAN WITHOUT ESS. Fig. 2 illusraes he hourly oal generaion according o he energy sorage sysem renewable sysems.

5 sysem, ESS he dem are given in Tables The spinning reserve requiremens are assumed o be 10% of he hourly dem. 5 TABLE 10. DEMAND DATA WITH 24-HOUR TIME HORIZON. Fig. 3. Hourly oal generaion according o ESS & renewables: Wih limi on renewable power. To he shor-erm operaional plan, energy sorage sysem has performed a load leveling generaion smoohing. The resul is compared wih previous sudies in Table 9. TABLE 9. COMARISON OF REVIOUS STUDIES. Case LR [14] GA [14] Only hermal generaors Thermal renewable Thermal, renewable ESS LRGA [14] MI 378, , , , , , , , , , , ,647 Due o he modeling difference on he energy sorage sysem, he cos level is differen in case 3. In Addiion, Fig. 4 shows he case ha he renewable oupu is no limied. Hour Dem (MW) Hour Dem (MW) Hour Dem (MW) , , , , , , , , , , , , , , , , , , TABLE 11. ENERGY STORAGE SYSTEM. Min[MWh] Max[MWh] C-rae iniial final % 30-40% TABLE 12. GENERATING UNIT DATA. Uni Uni1 Uni 2 Uni 3 Uni 4 Uni 5 Uni 6 Uni 7 Uni 8 Uni 9 Uni 10 (MW) (MW) a b c MUT (Hr) MDT (Hr) H C T (Hr) cold Iniial saus (Hr) Tables represen he opimal operaional plan of he hermal unis wihou wih he energy sorage sysem, respecively. Due o he energy sorage sysem, he operaional plan is changed as shown in green blocs in Table 14. Fig. 4. Hourly oal generaion according o ESS & renewables: Wihou limi on renewable power. B. Convenional ower Sysem A daily operaional planning is performed for he energy sorage sysem he hermal generaors in a convenional power sysem. The sysem has an energy sorage sysem, 10 large-scale generaors. The uni characerisics of he 10-uni

6 6 TABLE 13. OERATIONAL LAN WITHOUT ESS. of he energy sorage sysem, Fig. 6 illusraes he hourly oal generaion wih wihou he energy sorage sysem. Fig. 6. Hourly oal generaion according o ESS. TABLE 14. OERATIONAL LAN WITH ESS. To he shor-erm operaional plan, energy sorage sysem has performed a load leveling generaion smoohing. The resul is compared wih previous sudies only for hermal uni sysems in Table 15. TABLE 15. COMARISON WITH REVIOUS STUDIES. LR [10] GA [11] SA [12] ISO [13] MI 565, , , , ,820 IV. CONCLUSIONS This paper described a resource scheduling mehod for opimal operaional planning of power sysems wih he energy sorage sysems. Energy sorage sysem has performed a load leveling generaion smoohing o he shor-erm resources scheduling. The mehod uilized he Mixed Ineger rogramg based on he branch bound mehod. A daily opimal operaional planning was demonsraed in an isolaed micro-grid sysem wih renewable sysem energy sorage sysem in a convenional power sysem wih a large-scale energy sorage sysem. Fig. 5. Hourly oal charge/discharge quaniy of ESS. Fig. 5 illusraes he hourly oal charge/discharge quaniy V. REFERENCES [1] U.S. Deparmen of Energy Office of Elecriciy Delivery & Energy Reliabiliy, Energy Sorage: rogram lanning Documen, [2] J. Eyer G. Corey, Energy Sorage for he Elecriciy Grid: Benefis Mare oenial Assessmen Guide, Sia Naional Lab., [3] D. Rasler, Elecriciy Energy Sorage Technology Opions: A Whie aper rimer on Applicaions, Coss, Benefis, ERI, [4] H. Kanchev, Di Lu, F. Colas, V. Lazarov, B. Francois, Energy Managemen Operaional lanning of a Microgrid Wih a V-Based Acive Generaor for Smar Grid Applicaions, IEEE Trans. on Indusrial Elecronics, Vol. 58, No. 10, pp , Oc [5] D.-J. Lee L. Wang, Small-Signal Sabiliy Analysis of an Auonomous Hybrid Renewable Energy ower Generaion/Energy Sorage Sysem ar I: Time-Domain Simulaions, IEEE Trans. on Energy Conversion, Vol. 23, No. 1, pp , March [6]. Mercier, R. Cheraoui, A. Oudalov, Opimizing a Baery Energy Sorage Sysem for Frequency Conrol Applicaion in an Isolaed ower Sysem, IEEE Trans. on ower Sysems, Vol. 24, No. 3, pp , Aug [7] Bhara Bhargava Gary Dishaw, Applicaion of an Energy Source ower Sysem Sabilizer on he 10 MW Baery Energy Sorage Sysem

7 a Chino Subsaiion, IEEE Trans. on ower Sysems, Vol. 13, No. 1, pp. 145-151, Feb. 1998. [8] I.

Khederzadeh, N. Sadrmomazi, J. Olamaei, Inegraion of Wind power Energy Sorage in SCUC roblem, World Non-Grid-Conneced Wind ower Energy Conference (WNWEC), 2010 [10] A. J. Wood B. F.

on ower Apparaus Sysems, Vol. AS-102, pp. 1218-1255, 1983. [12] S. A. Kazarlis, A. G. Bairzis, V. eridis, A geneic algorihm soluion o he uni commimen problem, IEEE Trans. on ower Sysems, Vol. 11, No.

on ower Sysems, Vol. 21, No. 1, pp. 68-76, Feb. 2006. [15] T. Logenhiran D. Srinivasan, Shor erm generaion scheduling of a Microgrid, TENCON 2009-2009 IEEE Region 10 Conference, pp.1-6, Jan. 2009. [16] T.

He received his h.d. from Universiy of Michigan, Ann Arbr in 2000. He is an assisan professor in Energy IT Deparmen of Kyungwon Universiy. He is also an IEEE member.

7 7 a Chino Subsaiion, IEEE Trans. on ower Sysems, Vol. 13, No. 1, pp , Feb [8] I. apic, Simulaion Model for Discharging a Lead-Acid Baery Energy Sorage Sysem for Load Leveling, IEEE Trans. on Energy Conversion, Vol. 21, No. 2, pp , June [9] A. Daneshi, M. Khederzadeh, N. Sadrmomazi, J. Olamaei, Inegraion of Wind power Energy Sorage in SCUC roblem, World Non-Grid-Conneced Wind ower Energy Conference (WNWEC), 2010 [10] A. J. Wood B. F. Wollenberg, ower Generaion, Operaion, Conrol. New Yor: John Wiley & Sons, Inc., [11] A. Merlin. Srin, A new mehod for uni commimen a Elecricie de France, IEEE Trans. on ower Apparaus Sysems, Vol. AS-102, pp , [12] S. A. Kazarlis, A. G. Bairzis, V. eridis, A geneic algorihm soluion o he uni commimen problem, IEEE Trans. on ower Sysems, Vol. 11, No. 1, pp , Feb [13] F. Zhuang F. D. Galiana, Uni commimen by simulaed annealing, IEEE Trans. on ower Sysems, Vol. 5, No. 1, pp , Feb [14] D. N. Simopoulos, S. D. Kavaza, C. D. Vournas, Uni commimen by an enhanced simulaed annealing algorihm, IEEE Trans. on ower Sysems, Vol. 21, No. 1, pp , Feb [15] T. Logenhiran D. Srinivasan, Shor erm generaion scheduling of a Microgrid, TENCON IEEE Region 10 Conference, pp.1-6, Jan [16] T. Senyu, K. Shimabuuro, T. Funabashi, A Technique for Thermal Energy Sorage Sysem Uni Commimen, IEEE ower Engineering Sociey General Meeing, VI. BIOGRAHIES Sung-Yong Son received his B.S. M.S. degree from KAIST in , respecively. He received his h.d. from Universiy of Michigan, Ann Arbr in He is an assisan professor in Energy IT Deparmen of Kyungwon Universiy. He is also an IEEE member. He His primary research ineress include smar grid smar home relaed issues. Kwang Y. Lee (F 01) received he B.S. degree in Elecrical Engineering in 1964 from Seoul Naional Universiy, M.S. degree in Elecrical Engineering in 1968 from Norh Daoa Sae Universiy, h. D. degree in Sysems Science in 1971 from Michigan Sae Universiy. He was eleced as a Fellow of IEEE in January 2001 for his conribuions o he developmen implemenaion of inelligen sysem echniques for power plans power sysems conrol as a Life Fellow of IEEE since January He has been woring in he area of power plans power sysems conrol for over hiry years a Michigan Sae, Oregon Sae, Universiy of Houson, he ennsylvania Sae Universiy, he Baylor Universiy, he is rofessor Chairman of he Deparmen of Elecrical Compuer Engineering. His research ineress include conrol, operaion, planning of power energy sysems; compuaional inelligence, inelligen conrol heir applicaions o power energy sysems, modeling, simulaion conrol of micro-grids wih renewable disribued energy sources. Se-Hwan Jang received his B.S. M.S. degrees from usan Naional Universiy, Korea in , respecively. He is currenly woring for h.d. degree a Konu Universiy, Seoul, Korea. His research ineres includes opimal operaion planning of power sysem resources, deregulaion resrucuring of power indusry, elecriciy mare analysis, smar grid inelligen sysems applicaions o power sysems. Jong-Bae ar (M 98) received his B.S., M.S., h.d. degrees from Seoul Naional Universiy, Korea in 1987, 1989, 1998, respecively. For , he was wih Korea Elecric ower Corporaion, for he was an Assisan rofessor a Anyang Universiy, Korea. For , he was a gues researcher a ERI, USA. From 2001, he has been wih Elecrical Engineering Deparmen a Konu Universiy, he is nowa rofessor. His maor research opics include power sysem operaion, planning, economics, mares. Jae Hyung Roh received he B.S. degree in Nuclear Engineering from Seoul Naional Universiy, Korea, in 1993 he M.S. degree in Elecrical Engineering from Hongi Universiy, Korea, in He received h.d. degree in Elecrical Engineering from Illinois Insiue of Technology, Chicago, USA, in 19xx. For , he was wih Korea Elecric ower Corporaion, for , he was wih Korea ower Exchange. Since 2010, he has been wih Elecrical Engineering Deparmen a Konu Universiy, Seoul, as an Assisan rofessor. His research ineress include power sysems resrucuring, smar grid resource planning.

MULTI-OBJECTIVE OPTIMIZATION OF A BATTERY ENERGY MANAGEMENT FOR AN OFF-GRID SMART HOUSE. University of the Ryukyus, Okinawa, Japan

MULTI-OBJECTIVE OPTIMIZATION OF A BATTERY ENERGY MANAGEMENT FOR AN OFF-GRID SMART HOUSE. University of the Ryukyus, Okinawa, Japan Proceedings of BS: h Conference of Inernaional Building Performance Simulaion Associaion, Hyderabad, India, Dec. 79,. MULTIOBJECTIVE OPTIMIZATION OF A BATTERY ENERGY MANAGEMENT FOR AN OFFGRID SMART HOUSE