OPTIMAL LOCATION OF STATCOM USING PSO IN IEEE 30 BUS SYSTEM

Transcription

1 OPTIMAL LOCATION OF STATCOM USING PSO IN IEEE 30 BUS SYSTEM 1 Seraj Ahamad, 2 Prof. Azz Ahmed, 3 Prof. M.A Khan, 4 Sharad Kumar Pandey 1 M.Tech-Power System, Department of Electrcal & Electroncs Engg., Al-Falah School Of Engneerng & Technology, (An Autonomous Insttuton) Dhauj Fardabad Haryana, Pn Department of Electrcal & Electroncs Engg., Al-Falah School Of Engneerng & Technology,(An Autonomous Insttuton) Dhauj Fardabad Haryana Pn Department of Electrcal Engneerng.,Jama Mlla Islama 4 Assstant Professor, Department of Electrcal Engneerng., Greater Noda Insttute Of Technology, Knowledge Park-II, Greater Noda UP Pn ABSTRACT In ths paper we fnd the optmal locaton of STATCOM n the IEEE 30 bus to mnmze the value of voltage devaton. For fndng the optmal locaton of STATCOM we use the evolutonary computatonal technque called partcle swarm optmzaton (PSO) technque. We use the Newton-Raphson method to fnd the value of voltage at each and every bus. Voltage devaton s calculated by a objectve functon whch s mnmze to fnd the mnmum value of voltage devaton.frst we fnd the optmal allocaton of one STATCOM n IEEE 30 bus by smple mnmzng the value of objectve functon and after that we wll use two STATCOM smultaneously and fnd the optmal allocaton of them by usng partcle swarm optmzaton technque.. Keyword- FACTS, STATCOM, evolutonary technque, partcle swarm optmzaton technque (PSO), Newton-Raphson method. I INTRODUCTION Reactve power compensaton s a very mportant ssue n electrcal power systems and by the use of flexble ac transmsson system (FACTS) devces we can control the reactve power flow to the power network and hence the system voltage fluctuatons and stablty[1]. Voltage collapse problems n power systems have been permanent concern snce several major blackouts throughout the world have been drectly assocated wth ths voltage collapse problem. The collapse ponts are also known as maxmum loadablty ponts. The power flow control and statc stablty lmts of power system can be consderably modfed usng the new reactve compensaton equpments[2]. Newton-Raphson method: the Newton Raphson method s a powerful method of solvng nonlnear algebrac equatons. It works faster and surely converge n many cases but t requre large computer memory. II STATCOM The STATCOM (Statc synchronous Compensator) s a shunt connected reactve compensaton equpment whch s capable of generatng and absorbng reactve power that ts output can be vared so as to mantan control of specfc 156 P a g e

2 parameters of electrc power system[3],[4].statcom s normally consstng of voltage source nverter wth dc-lnk capactor. It s normally nterfaced to a system through a transformer. The transformer serves the purpose of steppng up the STATCOM voltage and also ts leakage reactance helps prevent short crcut of the dc-lnk capactor. Hence when usng multlevel STATCOM where by the STATCOM voltage could be bult from a number of dc-lnk to that of the system voltage, nterface reactor s requred to serve as the transformer leakage reactance. Bascally STATCOM can be modeled as a synchronous generator that absorbs/njects reactve power but does not generate any real power rather absorbed real power to cater for ts nternal and nterface losses. Z STATCOM represents the mpedance of the STATCOM. The magnary part represents the transformer leakage reactance/nterface reactor and the real part forms the ohmc losses of the STATCOM. Assumng multlevel Inverter STATCOM, n ths model, the ohmc resstance s excluded from the admttance matrx. Ths s due to the fact that, varous swtchng technques can be employed[6][7]. Man swtchng technques are PWM and fundamental frequency swtchng (FFS) where the swtchng losses are dfferent. The real power losses of the STATCOM are computed before the load flow dependng on the reactve powerspecfed at the STATCOM bus. The bus ntroduced four parameters, the same number as the system buses. The bus the STATCOM s connected to remans a load bus and so s the STATCOM bus. In the laggng power factor, the sgn of the STATCOM reactve power s the same as that of the load but opposte n leadng power factor. III. PARTICLE SWARM OPTIMIZATION TECHNIQUE There are many evolutonary technque n whch by whch we can fnd the optmal locaton of STATCOM such as generc algorthm (GA) evolutonary programmng Partcle swarm optmzaton (PSO). Partcle swarm optmzaton (PSO) s a populaton based stochastc optmzaton technque developed by Dr.Eberhart, Dr.Kennedy and Dr.sh n 1995, nspred by socal behavor of brd flockng or fsh schoolng. Partcle swarm adaptaton s an optmzaton paradgm that smulates the ablty of human socetes to process knowledge. The PSO s a s an effcent global optmzer for contnuous varable problems The algorthm models the exploraton of a problem space by a populaton of ndvduals; ndvduals' successes nfluence ther searches and those of ther peers. The algorthm s relevant to cognton, n partcular the representaton of schematc knowledge n neural networks. Partcle swarm optmzaton successfully optmzes network weghts, smulatng the adaptve sharng of representatons among socal collaborators. Partcle swarm optmzaton has roots n two man component methodologes. Perhaps more obvous are ts tes to artfcal lfe (A-lfe) n general, and to brd flockng, fsh schoolng, and swarmng theory n partcular. It s also related, however, to evolutonary computaton, and has tes to both genetc algorthms and evolutonary programmng. These relatonshps are brefly revewed n the paper. Basc algorthm as proposed by Kennedy and Eberhat Poston of ndvdual partcles updated as follows: x k - Partcle poston v k - Partcle velocty 157 P a g e

3 p - Best "remembered" ndvdual partcle p k p g k - Best "remembered" swarm poston c, 1 c - Cogntve and socal parameters 2 r, - Random numbers between 0 and 1 1 r2 Poston of ndvdual partcles updated as follows: x xk vk k1 1 wth the velocty calculated as follows: v v c r p k1 k 1 1 k k 2 2 x c r p g k xk IV. BUS SYSTEM For our research work we have taken the IEEE-30 bus system.the lne data, bus data s taken from the reference. Fg-1 Sngle lne dagram of the IEEE 30-bus test system 158 P a g e

4 Lne data of IEEE 30 bus system s gven n table -1. We use these data to calculate the YBUS of IEEE 30 test bus system whch s further use n Newton-Raphson load flow analyss of IEEE 30 bus system. As we see from the standard IEEE30 bus data the bus system has a) 1 slack bus b) 5 generator or PV buses c) 24 load or PQ buses By usng the load data and bus data we calculate the voltage at each bus by NewtonRaphson load flow method. The result s gven n followng table-3. Bus no. Voltage at bus OBJECTIVE FUNCTION: objectve functon (J) s a RMS value of voltage devaton. It s gven by j 30 V P a g e

5 Our man objectve s to mnmze the objectve functon because less objectve functon means less voltage devaton and good stablty. Wthout usng STATCOM the objectve functon of the IEEE 30 bus s V. ALGORITHM OF PSO VI. PSO PARAMETER We take the followng PSO parameter by applyng ht and tral method. 6.1 Inerta weght: we can take nerta weght n three manner 160 P a g e

6 a) Constant nerta weght- the nerta weght can be taken constant throughout the process but t seems somewhat unconvncng because as the partcle reach near the fnal soluton the mpact of pbest and gbest should be ncrease to avod to trap n local mnma. b) Lnearly decreasng nerta weght we can tan take nerta weght as lnearly decreasng but t has also problem to trap n local mnma. c) Randomly decrease nerta weght- n our thess we take nerta weght as randomly decreasng nerta weght as t has not any problem to trap n local mnma. So for our study we take nerta weght as a functon whch decrease randomly from.9 to.1. It s gven by followng: w ter * max_ ter No. of of partcle: we have taken 5 partcle wth the random ntal STACOM poston. 6.3 Acceleraton constant: there are two acceleratons constant we use. a) Indvdual acceleraton constant- ths acceleraton constant s used for show the mpact of ndvdual partcle best poston gan by partcle t self on the next value of velocty of that partcle. It s denoted by c 1. For our study we take c 1 =2.6 b) Socal acceleraton constant- the socal acceleraton constant s used to show the effect of best poston ganed by any partcle(gbest) on the next value of velocty of that partcle. It s denoted by c 2. For our study we take c 2 = No of teraton: 10 teratons s suffcent whle we deal wth sngle STATCOM and when we use 2 STATCOM we need mnmum 30 teratons. VII. RESULT 7.1 for one STATCOM: a) Wthout usng PSO- wthout usng PSO we have to calculate the objectve functon whle apply the STATCOM on every bus one by one and then we can choose the optmal bus locaton wth the mnmum objectve functon. whle we use only sngle STATCOM t s very easy to fnd the optmal locaton because we have to check only 30 bus but when we use two STATCOM at a tme then t s very dffcult because we have to check almost 800(30*29) buses whch s very dffcult so t s easy to use PSO method. In the followng table the objectve functon correspondng to STATCOM connected bus s gven. As we see that when we connected the STATCOM to the bus no. 12 then we get the mnmum value of objectve functon whch s.1364 whch s very less n comparson of other prevous objectve functon wthout usng the STATCOM P a g e

7 STATCOM OBJECTIVE b) By usng PSO: by usng PSO we get the same result as we get wthout the use of PSO. We get the optmal locaton n only 10 teratons when two STATCOM connected: When two STATCOM s connected then t s very dffcult to fnd optmal locaton wthout usng of PSO. So we get the optmal locaton of STATCOM by usng PSO.By usng ths algorthm we fnd a optmum bus locaton of (12 24) n whch the frst STATCOM s connected to the BUS NO 12 and the second STATCOM s connected to the bus no. 24. On ths optmal locaton we found the mnmum value of objectve functon whch s.1227 whch s very less n the comparson. As we see that when we connect the one STATCOM we found the mnmum value of objectve functon whch s.1364 whle usng two STATCOM we found the mnmum objectve functon Ths ndcate that f we ncrease the no. of STATCOM then t s possble to fnd the lesser value of objectve functon. 162 P a g e

8 ACKNOWLEDGMENT The author gratefully acknowledges the contrbuton of my senors Mr. V.K. Chopra, Md. Dansh Equbal, Mr. K.M Raf, Mr.Naqu Anwer, Mr. Sharqu Ahmed, Mr. Sanjay Rawat, Mr. Anl Gupta and my frends Mr. P.K bhardwaj Mr.Shakr, Mr. Amt Rajput, Mr. Pawan Kumar, Mr. Sanjeev Ojha, Mr. Rahul Prasad, Mr. Sonam Sngh and Mr. Abhay kumar Shah, who helped develop the methods reported n ths paper REFERENCES 1. N.G. Hngoran, and L. Gyugy, "Understandng FACTS; Concepts and Technology of Flexble AC Transmsson Systems," IEEE Press, New York, Modelng of Power Electroncs Equpment (FACTS) n Load Flow and Stablty Programs: A Representaton Gude for Power System Plannng and Analyss, Techncal brochure no. 145, TF , CIGRE (Aug. 1999). 3. C. A. Canzares, Modelng and Implementaton of TCR and VSI Based FACTS Controllers, Internal report, ENEL and Poltecnco d Mlano,Mlan, Italy (Oct.1999). 4. H. Mor, and Y. Goto, A parallel tabu search based method for determnng optmal allocaton of FACTS n power systems, Proc. Of the Internatonal Conference on Power System Technology (PowerCon 2000), vol. 2, 2000, pp Y. del Valle, Student Member, IEEE, J. C. Hernandez, Student Member, IEEE,G. K. Venayagamoorthy, Senor Member, IEEE, and R. G. Harley, Fellow, IEEE Optmal STATCOM szng and placement Usng Partcle Swarm Optmzaton Accepted by the IEEE PES Transmsson and Dstrbuton Conference and Exposton Latn Amerca 2006, Caracas, Venezuela, Nabae, A, Takahash, I, and Akag, H, A new neutral clamped PWM nverters, IEEE Trans. Ind. Applcat., pp , Meynard, T.A, and Foch, H, mult-level converson: hgh voltage chopper and voltage-source nverter, IEEE-PESC Conf. Rec, pp , Rodrguez, J, La, J. S, and Peng, F. Z, Multlevel nverters: a survey of topologes, controls, and applcatons, IEEE Trans. Ind. Electron., pp , J. Kennedy, and R.C. Eberhart, "Swarm ntellgence," Morgan Kaufmann, San Francsco, J. Kennedy, R. Eberhart, Partcle swarm optmzaton Proceedngs of IEEE Internatonal Conference on Neural Networks (lcnn 95), vol. IV, pp , Perth. Australa R. Eberhart, and J. Kennedy, "A new optmzer usng partcle swarm theory," n Proc. 6th Int. Symp. Mcro Machne and Human Scence (MHS'95), 1995, pp P a g e

9 12. J. Kennedy, "The partcle swarm: socal adaptaton of knowledge," n Proc. IEEE Int. Conf. Evolutonary Computaton, 1997, pp L.J. Ca, I. Erlch, and G. Stamtss, "Optmal choce and allocaton of FACTS devces n deregulated electrcty market usng genetc algorthms," Proc. of the IEEE PES Power Systems Conference and Exposton, vol. 1, 2004, pp W. Ongsakul, and P. Jrapong, " Optmal allocaton of FACTS devces to enhance total transfer capablty usng evolutonary programmng," Proc. of the IEEE Internatonal Symposum on Crcuts and Systems (ISCAS 2005), vol. 5, 2005, pp Venayagamoorthy GK, del Valle Y, Qao W, Mohaghegh S, Ray S, Harley RG, "Effects of a STATCOM, a SSSC and a UPFC on the Dynamc Behavor of a 45 Bus Brazlan Power System", IEEE PES Inaugural 2005 Conference and Exposton n Afrca, Durban, South Afrca, July 11-14, 2005, pp S. Gerbex, R. Cherkaou, A.J. Germond, Optmal locaton of multtype FACTS devces n a power system by means of genetc algorthms, IEEE Transactons on Power Systems, Volume 16, Issue 3, Aug Page(s): W. Ongsakul, P. Jrapong, Optmal allocaton of FACTS devces to enhance total transfer capablty usng evolutonary programmng, IEEE Internatonal Symposum on Crcuts and Systems, 2005.ISCAS May 2005, Vol. 5 Page(s): E. Nasr Azadan, S. H. Hossenan, M.Janat and P.Hasanpor, Optmal Placement of Multple STATCOM, Power System Conference, MEPCON th Internatonal Mddle-East 19. F.F.Syed, A.H.M.A.Rahm, SM, IEEE and J.M.Ba-Khashwan, RobustSTATCOM Controller Desgn Usng PSO Based Automatc Loop-Shapng Procedure Control Applcatons, CCA Proceedngs of 2005 IEEE Conference on 20. M. Clerc, J. Kennedy, The partcle swarm exploson, stablty and convergence n a multdmensonal complex space, IEEE Trans, Evolutonary Computaton, vol. 6(1), pp , Feb 2002 AUTHORS: Seraj Ahamad receved the B.Tech (Electrcal Engg.) Degree from Uttar Pradesh Techncal Unversty, Lucknow, Inda. He s currently pursung the M.Tech ( Power System ) Degree n the Department of Electrcal & Electroncs Engneerng, AFSET, MDU Rohtak, Haryana Inda. Hs research nterests are Power System, Specal Electrcal Machnes, Swtch Gear & Protecton. Prof. Azz Ahmed receved the B.Tech (Electrcal Engg.) Degree from Jama Mla Islama, New Delh,.and M.Tech(Control System) Degree from Delh Techncal Unvesty. He s currently pursung the Phd Degree n the Department of Electrcal Engneerng, from Jama Mla Islama, New Delh Inda. He s currently workng as Professor & HOD n the Department of Electrcal & Electroncs Engneerng, AFSET Fardabad Haryana. Hs research nterests are Facts, Power System, Specal Electrcal Machnes, Swtch Gear & Protecton. 164 P a g e

10 Prof. M.A Khan receved the B.E, M.E & Phd (Electrcal Engg.) Degree from Algarh Muslm Unversty, Inda. He s currently workng as Professor & HOD n the Department of Electrcal Engneerng(Polytechnc),Jama Mlla Islama, New Delh Inda. Hs research nterests are Deregulaton of electrcty, Power System, Specal Electrcal Machnes,Swtch Gear & Protecton. Mr.Sharad Kumar Pandey receved the B.Tech (Electrcal Engg.)Degree from Uttar Pradesh Techncal Unversty, Lucknow,.and M.Tech (Power System ) Degree from Punjab Engneerng Collge Chandgarh Inda. Hs researchnterests are Power System, Specal Electrcal Machnes, Swtch Gear & Protecton. 165 P a g e