Fuzzy control of a doubly fed induction generator based wind power system

Size: px

Start display at page:

Download "Fuzzy control of a doubly fed induction generator based wind power system"

Edgar Doyle
6 years ago
Views:

1 Maouane El Azzaoui, Haane Mahoudi conol of a doubly fed inducion geneao baed wind powe ye Maouane El Azzaoui and Haane Mahoudi Powe eleconic and conol ea, Depaen of elecical engineeing, Mohaadia School of Enginee, Mohaed V Univeiy Agdal, Raba, Moocco aouane.elazzaoui@eeach.ei.ac.a Abac: - Thi pape deal wih he veco conol baed on fuzzy logic of acive and eacive powe of a doubly fed inducion geneao (DFIG) inegaed in a wind ubine ye. The powe anfe beween he ao and he gid i achieved by acing on he oo ignal via a bidiecional convee. Fo a copaaive udy, he independen conol of acive and eacive powe i enued in he fi ep by convenional conolle () and he econd ep by he fuzzy conolle. The pefoance and obune ae analyed and copaed by iulaion baed on Malab / Siulink envionen. Key-Wod: - Doubly fed inducion geneao, fuzzy logic, Popoional Inegal, Veco conol, wind enegy. Inoducion Duing he la yea, hee wa a ong developen of wind enegy. Renewable and clean, hi enegy of he fuue ha a ajo ole o play in eeing he cuen and fuue cliae challenge. DFIG i conideed one of he o adequae oluion o inegae i ino a wind ye. I offe any advanage, uch a educion of invee co. In hi pape, we focu on he conol aegy of he doubly fed inducion geneao (DFIG) baed wind powe ye. A fi, he odeling of he ubine and he DFIG will be peened. Then, he veco conol of acive and eacive powe i achieved by adjuing he oo agniude, while applying he and fuzzy conolle. Finally, o validae he popoed conol, we will conduc a eie of nueical iulaion uing Malab Siulink ofwae. Modeling of wind ubine Fig. how he conveion chain baed DFIG conneced o he elecical powe gid. The ubine conve he kineic enegy of wind ino elecical enegy. Thi poceing depend on he ai deniy, he aea wep by he oo and wind peed []. The aeodynaic powe appeaing a he oo of he ubine i wien [] []:... (, ). Pae = ρscp λβ v () Whee: ρ ai deniy, S ubine aea, v wind peed, C ( λβ, ) powe coefficien. p Blade Geabox DFI G convee ide oo Fig. Sucue of a wind enegy ye baed on DFIG The peed aio i defined by: Ω. R λ = () v Whee Ω epeen he oaional peed of ubine. In ou udy, he expeion of he powe coefficien i given by [4]: 5 p (, ) c 4 exp c C λβ = c c β c + c6λ λi λi () Wih: =.5, c =.576, λi λ +.8β β + c = 6, c =.4, c 4 = 5, c 5 =, c 6 =.68 Knowing he peed of he ubine, he aeodynaic oque i heefoe deeined by [5]. Pae Tae = = C p ( λβρ, ) Sv Ω. Ω convee ide gid Gid (4) E-ISSN: 4-5X 7 Volue, 7

2 Maouane El Azzaoui, Haane Mahoudi The uliplie i odeled by he following equaion: Ω = GΩ (5) T ae = GT (6) The Model popoed conide ha he oal ineia J i given by: g J J = + J g (7) G By applying he fundaenal law of dynaic, he elaionhip beween he oaion peed and oque i expeed by: dω = = Ω (8) J Tec Tg Te f Whee: T e elecoagneic oque, T g oque deived fo uliplie, f Coefficien of vicou ficion. Maheaical odel of he DFIG The geneao choen fo he wind powe ye i he DFIG. In addiion, he DFIG conolled by invee hough i oo opeae ove a ange of ±% vaiaion aound he ynchonou peed. Thi choice allow he convee dienioned o ani % of noinal powe. Theefoe, i will be econoic [6]. The odel of he DFIG i expeed in he (dq) efeence fae by he following equaion [7] [8] [9]. V V V V + ωϕ + ωϕ d d d q q q q + + ωϕ d d d d q q q q + + ωϕ d ϕ ϕ ϕ ϕ d d d q q q d d d q q q (9) () T = p( ϕ I ϕ I ) () e d q q d ω = ω pω () Whee: / ae ao/oo ubcip, V / I volage/cuen, R / R oo/ao eiance, L / Loo/ao leakage inducance, L uual inducance, p nube of pai of pole of he DFIG, ω / ω oo/ao pulaion, ϕ flux, T e elecoagneic oque, Ω echanical peed of DFIG. 4 Conol aegy To be able o conol he DFIG eaily, we will achieve independen conol of acive and eacive powe by he ao flux oienaion. The pinciple i o align he ao flux along he axi (d) of he oaing fae; hi can be wien in he fo of equaion a follow: ϕ = ϕ, ϕ = () d q The ao i aued upply by a able gid. Moeove, he ao eiance can be negleced ince i i a ealiic aupion fo high powe achine ued fo wind powe [], hee iplifying aupion allow u o wie he following equaion: V =, V = V (4) d q The ao acive and eacive powe ae wien []: P = Vd Id + VqIq (5) Q = VqId Vd Iq (6) Conideing he iplificaion enioned above, expeion of powe becoe: Q L P = V I (7) q L LV V Id L ωl = + (8) Theeafe, we eablih he equaion howing he elaionhip beween he oo volage and oo cuen: d Vd d + σl di gσlω Iq (9) diq LV Vq q + σl + gσlω Id + g () L L Whee: g Slip of he DFIG, σ = he LL dipeion coefficien of he DFIG. E-ISSN: 4-5X 74 Volue, 7

3 Maouane El Azzaoui, Haane Mahoudi Equaion (7), (8), (9) and () allow u o eablih he elecical ye of he block diaga of he DFIG o egulae (Fig.). aheaical odel of he ye, ju by incopoaing expe knowledge [5]. A peened in Fig.4, he fuzzy conol baed on Madani odel. Thi echani i divided ino hee pa. Fi, uing inpu ebehip funcion, he inpu ae fuzzified and hen baed on ule bae and infeence ye, oupu ae poduced and finally he fuzzy oupu ae defuzzified and applied o he ye. Eo and he eo change ae ae eleced a inpu. Fig. Sucue of a wind enegy ye baed on DFIG 5 conol of he DFIG The cuen and volage ae elaed by a fi ode anfe funcion. Since he lip value i weak, we will e up a conol loop on each powe wih an independen egulao () while copenaing he diubance e (Fig.). Fig.4 Block diaga of fuzzy conol Fig.5 how he block diaga whee fuzzy conolle ae inegaed ino he veco conol block of DFIG. The acive and eacive powe ae conolled independenly, each wih i own fuzzy conolle. Fig.5 conol of DFIG Fig. Diaga of conol of he DFIG by egulao The Popoional Inegal conolle () i ill he o coonly ued fo conolling he DFIG [], a well a in any induial conol ye. I i eay and quick o ipleen while poviding accepable pefoance []. Regulao each axi ole i o eliinae he gap beween he acive and eacive powe efeence and he eaued acive and eacive powe. The ynhei of conolle i peened in [4]. 6 conol of he DFIG The fuzzy conol appoache o oe exen he flexibiliy of huan eaoning; he anage o offe aifacoy pefoance wihou he need of he The deign of he fuzzy conolle i hown in he block diaga of fig.6. The gain G e, G e, and G u ae caling faco (noalizaion). We vay hee faco unil we can ge a pope conol. Indeed, i i hey who will decide he pefoance of conol. Fig.6 Deign of a fuzzy conolle Geneally, he deign of a fuzzy conolle fo he conol of elecic dive equie he choice of he E-ISSN: 4-5X 75 Volue, 7

4 Maouane El Azzaoui, Haane Mahoudi following paaee: linguiic vaiable, ebehip funcion, infeence ehod, and defuzzificaion aegy [6]. conolle inpu ae he eo and i deivaive, while he oupu i he coand ielf. Tiangula and apezoidal ebehip funcion ae ued on a univee of dicoue noalized in he ange [-; ] fo each vaiable a hown in Fig.7 and Fig.8 epecively fo he inpu (eo, eo vaiaion) and oupu (inpu poce). Degee of ebehip eo and change of eo Fig.7 Mebehip funcion of eo and change of eo Degee of ebehip BN SN AZ SP BP BN SN AZ SP BP Fig.9 how he uface geneaed by he fuzzy ye. du de Fig.9 Suface geneaed by he fuzzy ye 7 Siulaion eul The iulaion of he conol of he DFIG aociaed wih wind ye wa ade wih he Malab / Siulink envionen (Fig.). -.5 e Oupu Fig.8 Mebehip funcion of oupu The ube fuzzy ebehip wee noed a follow: BN: Big-Negaive; SN: Sall-Negaive; AZ: Abou-Zeo; SP: Sall-Poiive; BP: Big- Poiive. The fuzzy ule, fo deeining oupu vaiable of he conolle a a funcion of inpu vaiable ae gouped in he Table. Table. Infeence aix e Δu BN SN AZ SP BP BN BN BN SN SN AZ SN BN SN SN AZ SP Δe AZ BN SN AZ SP BP Fig. Block diaga of he conol in Malab/Siulink The eul ploed in he following figue how powe geneaed when he efeence ignal ae applied. Fig. diplay he eul of and fuzzy conolle uing he age in he fo of echelon. Fig. how he eul wih vaiable age of powe. The e of obune i peened in fig.. SP BN AZ SP SP BP BP AZ SP SP BP BP E-ISSN: 4-5X 76 Volue, 7

5 Maouane El Azzaoui, Haane Mahoudi Reacive powe (VAR) Acive Powe (W) (a) Sao acive powe (b) Sao eacive powe Fig. Repone wih fuzzy and conolle I i clea ha he acive and eacive powe of he ye follow pefecly he efeence level. Saic eo i alo equal o zeo. In anien ae, he ie epone of he fuzzy conolle i uch fae han he conolle (Fig.). Acive powe (W) (a) Sao acive powe (b) Sao eacive powe Fig. Repone wih fuzzy and conolle fo vaiable powe Fig. how ha he wo conolle follow hei efeence, bu he fuzzy peen bee pefoance han he in e of epone ie and aic eo. The iulaion eul in Fig. how he obune of and conolle o vaiaion in eveal paaee: Vaiaion of oo eiance noinal value. P* Q* ie () eacive powe (VAR) P* ie () Q* ie () R of +4% of i Vaiaion of he ao inducance L, oo inducance L and uual inducance L of % of hei noinal value. Acive powe (W) Reacive powe (VAR) (a) Sao acive powe (b) Sao eacive powe Fig. Repone wih fuzzy and conolle The copaion beween he wo conolle how ha he fuzzy conolle ha good pefoance. By con, fo he conolle, i pefoance i lo. 8 Concluion The conol of a wind enegy ye equipped wih a doubly fed inducion geneao ha been decibed in hi pape. A fi, a odel of he ubine and he geneao ae popoed. Then, a conol aegy baed on fuzzy logic and conolle allowing independen conol of powe ha been peened. Though he epone chaaceiic obained by he iulaion eul, he good pefoance i obeved even in he peence of vaiaion of age. Moeove, by copaing he conolle and fuzzy, i clea ha he fuzzy conol i obu again paaeic vaiaion of he achine, povide fa convegence, no affeced by noie and puiou ignal, and eay o ipleen i in a calculao. Appendix Paaee of he wind powe ye. DFIG noinal paaee V = V, 5 Hz, p =, R =.455Ω, R =.9Ω L =.7H, L =.H, ie () 4 P* Q* ie() E-ISSN: 4-5X 77 Volue, 7

6 Maouane El Azzaoui, Haane Mahoudi L =.4H Tubine noinal paaee Nube of blade =, Diaee = 6 Geabox = 7 DFIG + Tubine paaee J = Kg., f =.46 N.. d / ad Refeence: [] C. NICHITA, D. LUCA, B. DAKYO, E. CEANGA, and N. A. CUTULULIS, Modelling non-aionay wind peed fo enewable enegy ye conol, The annal of Dunaea de Jo Univeiy of Galai, Facicle III, pp. 9 4,. [] M. Aifujjaan, Veco conol of a dfig baed wind ubine, Jounal of Elecical and Eleconic Engineeing, vol., no. 9, pp , 9. [] I. Muneanu, A. I. Bacu, N. A. Cuululi, and E. Ceanga, Opial conol of wind enegy ye, Spinge, 8. [4] Z. M. Lubony and B. Anjan, Wind ubine opeaion in elecic powe ye, Spinge,. [5] M. GOSSA, K. JEMLI, M. B. MOHAMED, and M. JEMLI, Doubly fed inducion geneao (dfig) in wind ubine, odeling and powe conol, IEEE Inenaional Confeence on Induial Technology, pp , 4. [6] F. Poiie, B. Toufik, and A. Machou, Advanced conol of a doubly-fed inducion geneao fo wind enegy conveion, Elecic Powe Sye Reeach, no. 79, pp , 9. [7] H. Bekka, S. Taaf, D. Rekioua, and S. Bacha, Powe conol of a wind geneao conneced o he gid in fon of ong wind, Jounal of Elecical Sye, vol., no. 9, pp ,. [8] D. Rekioua, Wind powe elecic ye: Modeling, iulaion and conol, geen enegy and echnology, Spinge, 4. [9] A. Boyee, P. Poue, and S. Saadae, Diec and indiec conol of a doubly fed inducion geneao wind ubine including a oage uni, in h ediion of Induial Eleconic Confeence IECON 6, IEEE, Pai, Fance, Novebe 6-6, pp [] A. Tapia, G. Tapia, J. X. Oolaza, and J. R. Saenz, Modeling and conol of a wind ubine diven doubly fed inducion geneao, IEEE Tanacion on Enegy Conveion, vol. 8, no., pp. 94 4, June. [] R. PENA, J. C. CLARE, and G. M. ASHER, Doubly fed inducion geneao uing backo-back pw convee and i applicaion o vaiable- peed wind-enegy geneaion, IEE Puoc.-Elec. Powe Appl, vol. 4, no., May 996. [] M. LOPEZ, Conibuion à l opiiaion d un yèe de conveion éolien pou une unié de poducion iolée, Ph.D. dieaion, Science e Technologie de l Infoaion de Télécounicaion e de Syèe Univeié Pai Sud, Fance, 8. [] G. A. Sih and K. A. Nigi, Wind-enegy ecovey by a aic chebiu inducion geneao, in Poc. IEE, 98, pp [4] N. CHERFIA, D. KERDOUN, and A. BOUMASSATA, Coecion of he echanical peed fo he dfig wind ubine, Inenaional Jounal of Reeach in Engineeing and Technology, vol., no., pp. 9 8, Novebe 4. [5] H. K. Davijani, A. Sheikholelai, H. Livani, and M. K. Davijani, logic conol of doubly fed inducion geneao wind ubine, Wold Applied Science Jounal, vol. 6, no. 4, pp , 9. [6] M. GODOY, B. SIMGES, K. BOSE, and R. J. SEGEL, logic baed inelligen conol of a vaiable peed cage achine wind geneaing ye, in PESC, 995. E-ISSN: 4-5X 78 Volue, 7

A Novel Gain Tuning of Anti-Windup PID Controller using Ant Lion Optimization

A Novel Gain Tuning of Anti-Windup PID Controller using Ant Lion Optimization Inenaional Jounal of Conol Theoy and Applicaion ISSN : 974-557 Inenaional Science Pe Volume Numbe 6 7 A Novel Gain Tuning of Ani-Windup PID Conolle uing An Lion Opimizaion Ayyaao SLV Tummala a, Ravi Kian