Short Term Generation Scheduling of Thermal Units with Emission Limitation in Deregulation Environment

Transcription

1 Short Term Generaton Schedulng of Thermal Unts wth Emsson Lmtaton n Deregulaton Envronment C.Sakthvel 1, S.Pavthra 2, S.Pradeep kumar 3, R.Guruprasath 4 1Assstant Professor, Department of EEE, JCT College of Engneerng and Technology 2Assstant Professor, Department of EEE, JCT College of Engneerng and Technology 3Assstant Professor, Department of EEE, JCT College of Engneerng and Technology 4Assstant Professor, Department of EEE, SNS college of Technology, Combatore Abstract -Ths paper presents a generaton schedulng of thermal unts consderng startup and shutdown ramp lmts by usng Shuffled Frog Leapng Algorthm. The Startup and shutdown ramp lmts are does not consder n the conventonal method of unt commtment (UC. Wthout consderng the ramp lmts n generaton schedulng problem, soluton for large power system does not gve practcal value. The objectve of proposed work s to determne the optmal commtted of thermal generatng unts at mnmum operatng whle consderng load demand, spnnng reserve and other equalty and nequalty constrants at each hour tme nterval. The soluton obtaned from the proposed Shuffled Frog Leapng Algorthm (SFLA for four unt system s compared wth other conventonal methods. Keywords -Unt Commtment, Shuffled Frog Leapng Algorthm (SFLA, Startup Ramp lmt and Shutdown Ramp lmt. I. INTRODUCTION The plannng and operaton of generaton schedulng of thermal unts n electrc power system s based on the total load demand on the system. The daly load demand of the system vares tme to tme and hour to hour. So the commtment of thermal unts gets mportant to ON/OFF the thermal generatng unts [1]. The commtted unts optmally dstrbute the forecasted load. Unt commtment s a mxed nteger nonlnear optmzaton problem used to schedule the thermal generatng unts n order to satsfyng the load demand and reserve requrements of mnmum [2]. The thermal unt commtment problem s solved by varous optmzaton methods [3].The methods are lke Prorty lst (PL, Forward Dynamc programmng(fdp, Lagrangan Relaxaton (LR, Genetc algorthm(ga, Smulated Annealng (SA, Partcle Swarm optmzaton (PSO and Ant colony Optmzaton (ACO. The prorty lst method comparatvely gves hgh producton and hgh computatonal tme. The dynamc programmng method has more mathematcal complexty compare to other methods for solvng unt commtment problem [4]. The nherence sub optmalty s the man drawback n the Lagrangan relaxaton based unt commtment problems [5]. In genetc algorthm convergence does not guaranty to produced optmal soluton compare to other methods [6]. The smulated annealng also has some mathematcal complexty to mplement the schedulng problem [7]. Partcle swarm Optmzaton s based on brds flockng behavor whch produces the effcent output [8],[9]. In ant colony Optmzaton, the colony of ndvduals s adoptng by decson makng polcy [10]. In ths paper we propose a nteger coded Shuffled Frog Leapng Algorthm (SFLA for solvng the short term thermal generaton schedulng problems. The four thermal generatng unts wth eght hour load demand s consder as a case study for our proposed work. The organzaton of paper s follows. Secton I descrbes the ntroducton of unt commtment; secton II a detaled mathematcal formulaton of problem wth consderng the varous constrants. The proposed nteger coded SFLA wth flow chart s explaned n secton III. Secton IV and V gves the result and dscusson and concluson of our proposed work. II. PROBLEM FORMULATION The Unt commtment problem can be mathematcally formulated by the followng equaton. 2015, IRJET ISO 9001:2008 Certfed Journal Page 1100

2 The objectve functon subjected to followng constrants. II.1. Equalty Constrants: A. Power Balance Constrant (5 1 Tme duraton curve Fg. II.2. Inequalty Constrants: A. Generaton lmt constrant P,mn P,t P,max B. Reserve lmt constrant 0 R,t P,max P,t (7 C. Mnmum up tme constrant T ON UT D. Mnmum down tme constrant T OFF DT E. Spnnng reserve constrants F. Ramp up & down constrant (11 (12 Where, K=60 mn s the UC schedulng tme step. G. Startup Tme Ramp lmt constrant H. Shut down Tme Ramp lmt constrant (8 (9 (10 (13 (14 The followng Fg. 1 shows the tme duraton curve wth startup ramp lmt (SR u and shutdown ramp lmt (SR D. III. Shuffled Frog Leapng Algorthm: The Shuffled Frog Leapng Algorthm s one of the Meta heurstc optmzaton methods to fnd a soluton of combnatonal optmzaton problem [11]. It s Developed B.Y. Ensuff and Lansey n SFLA s based on the behavor of group of frogs searchng for the maxmum amount of avalable food locaton. The populaton of group of frogs s parttoned nto several parallel communtes called memeplex. Wthn each memeplex each frog holdng some dea and t can be nfluenced by the other frog s dea. The deas are passed between the parallel communtes n a shufflng process[12]-[15]. The local search and the shufflng process are contnung up to the convergence crtera of the tolerance value or number of total teratons met. The shufflng enhances the qualty of mems after beng nfected frogs from all memeplexes. In the proposed method ntally P numbers of populatons are created randomly and t s dvded by number of memeplexes. The worst frog poston s updated by the followng two equatons. The shufflng processes of frogs n memeplexes are contnued untl the convergence crtera s are satsfed [16],[17]. ( ( The specfed no of shufflng process s reached. The relatve change n output n the two shufflng s less than the specfed tolerance value. D = rand.(x b X w (15 X new W X D (D mn <D <D max (16 current W, 2015, IRJET ISO 9001:2008 Certfed Journal Page 1101

3 S. N o Table. I : Unt operator data of a 4-unt base problem Table II : Proposed method output wth and wthout consderng startup, shutdown ramp lmt Load (MW Wthout consderng Reserve & Ramp constrant Wth consderng Reserve constrant Power Generaton of Product Power Generaton of Product Cost ($ Cost ($ on Cost ($ Cost ($ on Cost Cost ($ ($ s Un t no. Pma x (M W Pmn (MW Fuel a b c UT (hr DT (hr Ramp Lmt (MW/hr Shut Down ($/hr Hot Star t Up ($/h r Cold Start Up ($/hr Cold Start Tme (hr Unt Condt on (hr , IRJET ISO 9001:2008 Certfed Journal Page 1102

4 Table.III : Optmal solutons of the proposed method for consderng ramp and startup, shutdown ramp lmt Wth consderng Reserve &normal Ramp constrant Wth consderng Start up and Shut down Ramp lmt S. Power Generaton of Power Generaton of Load N Product (MW Producto o Cost ($ Cost ($ Cost ($ Cost ($ on Cost n Cost ($ ($ Table IV : Load demand of the 4-unt base problem HOUR LOAD HOUR LOAD Table V : Comparson of smulaton results S.No. Method Constrants Taken operatng operatng ($ / day ($ / year 1. DP Wthout Ramp & Reserve Wthout Ramp & Reserve SFLA Wth Reserve Wth normal Ramp & Reserve Wth Startup and Shutdown Ramp , IRJET ISO 9001:2008 Certfed Journal Page 1103

5 IV. Result and Dscusson The proposed method has been appled to four unt system over a schedulng perod of 24 hour. Table I shows the operator data for 4 unt system. The table IV shows the load demand value for 24 hour schedulng horzon. The ntal populaton of hundred frogs are dvded nto twenty parallel combnaton are taken as an nput parameters. The shuffled Frog Leapng Algorthm (SFLA s mplemented usng MATLAB software package and the system confguraton s Intel core 3 wth 2.7GHZ speed and 4GB RAM. The spnnng reserve s assumed to be 10% of the systems load demand of that hour. The proposed method s tested wth three dfferent set of nstances. Table II shows the proposed method output wth and wthout consderng startup, shutdown ramp lmt. The normal ramp lmt constrant and startup, shutdown ramp lmt consdered output s gven n table III. Fnally the comparson results of the varous set of nstances are gven n table V. V. Concluson Ths paper has proposed a new nteger coded Shuffled Frog Leapng Algorthm for the soluton of unt commtment problem wth startup and shutdown ramp lmts. Four unt systems are consdered to gve the effectveness of the proposed Shuffled Frog Leapng Algorthm and smulaton results were compared wth results obtaned by conventonal dynamc programmng method. The smulaton results show that the proposed work has less conventonal tme and producton compare to other conventonal methods. The results acheved n proposed method for solvng unt commtment problem s qut engagng compare to other methods REFERENCES [1] J. Wood and B. F. Wollenberg, Power Generaton Operaton and Control. New York: Wley, [2] S.M. Shahdehpour, H.Y. Yamn, Z. L, Market Operatons n Electrc Power Systems, John Wley and Sons, [3] N. P. Padhy, Unt commtment A bblographcal survey, IEEE Trans. Power Syst., vol. 19, no. 2, 2004, pp [4] C. S. Boopath, SubhransuSekhar Dash, K. Selvakumar, A. Venkadesan, C. Subraman and D. VamsKrshna, Unt Commtment Problem wth POZ Constrant Usng Dynamc Programmng Method, Internatonal Revew of Electrcal Engneerng (IREE Vol.9 Issue.1. Pp , July-August [5] C.-P. Cheng, C.-W. Lu, and C.-C. Lu, Unt commtment by Lagrangan Relaxaton and genetc algorthm, IEEE Trans. Power Syst., vol. 15, pp , May [6] K. S. Swarup and S. Yamashro, Unt commtment soluton methodology usng genetc algorthm, IEEE Trans. Power Syst., vol. 17, pp , Feb [7] K. P.Wong and Y.W.Wong, Short term hydro thermal schedulng part I: smmulated annealng approach, Proc. Inst. Elect. Eng., Gen. Transm. Dst., vol. 141, pp , [8] A.Y. Saber, T. Senjyu, N. Urasak, and T. Funabash, Unt commtment computaton A novel fuzzy adaptve partcle swarm optmzaton approach, n Proc. IEEE PSCE, Nov. 2006, pp [9] T. O. Tng, M. V. C. Rao, and C. K. Loo, A novel approach for unt commtment problem va an effectve hybrd partcle swarm optmzaton, IEEE Trans. Power Syst., vol. 21, no. 1, pp , Feb [10] Sureshkumar. C, Selvakumar. K and Santhoshkumar. G, Untcommtment problem solved by Ant colony search algorthm, IJAIR, Vol.1, No.7, pp , [11] Javad Ebrah, Seyed Hossen Hossenan, and Gevorg B. Gharehpetan, UC Problem Soluton Usng Shuffled FrogLeapng Algorthm, IEEE Trans. Power Syst., vol. 26, no. 2, pp May [12] M.M.Eusuff, K.E.Lansey, and F.Pasha, Shuffled Frog- Leapng Algorthm: A Memetc Meta-Heurstc For Dscrete Optmzaton Eng.Optmz, Vol.38, No.2, Pp , [13] Selvakumar.K, Vgnesh.R.S and Vjayabalan.R, shuffled frog leapng algorthm for solvng Proft Based Unt Commtment problem, n proc. IEEE ICCCA, pp.1-6, [14] Selvakumar.K, Venkatesan.T and Sanavullah.M.Y., Prce Based Unt Commtment problem soluton usng shuffled frog leapng algorthm, n proc. IEEE ICAESM, pp , [15] M.M.Eusuff and K.E.Lansey, Optmzaton of water dstrbuton network desgn usng the shuffled frog leapng algorthm, J. Water Resources Plannng &.Management, vol.129,no.3,pp ,2003 [16] T. Venkatesan and M.Y. Sanavullah, Shuffled Frog Leapng Algorthm Based Soluton to the Proft Based Unt Commtment Problem, Archves Des Scences, Vol. 65, No. 4, pp , Apr [17] E.Ebeltag,T.Hegazy,andD.Grerson, Comparson among fve evolutonary based optmzaton algorthms, Adv. 2015, IRJET ISO 9001:2008 Certfed Journal Page 1104