Procdings of th 2nd Sminar on Enginring and Information Tchnology Rotor Spd Control of Micro Hydro Synchronous Gnrator Using Fuzzy PID Controllr C. S. Chin K. T. K. To P. Nlakantan Elctrical and Elctronics Enginring Program, School of Enginring and Information Tchnology Univrsiti Malaysia Sabah, Lockd Bag No. 2073, 88999 Kota Kinabalu, Sabah, Malaysia Tl: +60-8-832-0000 x 353, E-mail: knto@ums.du.my Abstract - Hydro nrgy is on of th common rnwabl nrgy utilizd for powr gnration systm in Malaysia. Th hydro nrgy rotats a turbin in which th turbin is usd to driv th gnrator. In this study, stand alon micro hydro powr gnration systm is chosn to bring in lctrical nrgy to rural ara sinc it is costly to transmit nrgy to rural ara. Th gnratd lctrical nrgy by stand alon micro hydro gnrator is distributd to th usrs by transmitting th lctrical nrgy to th load. Howvr, th load is not consistnt as thr is diffrnt powr dmand at diffrnt tim intrval. Th rspondd rotor spd will b dtrmind and corrctiv action will b applid to gnrat powr to match th load. Proportional-intgral-drivativ (PID) controllr will b implmntd into stand alon micro hydro powr gnration systm to control th gat of th watr flow. PID controllr is subsquntly rplacd by Fuzzy PID for auto tuning purpos..0 INTRODUCTION Enrgy is ssntial to maintain quality of lif of human bings. Elctrical nrgy is usually gnratd by gnrators which ar drivn by turbins. In gnral, th drivn turbins ar opratd by using convntional sourcs such as disl and natural gas. Howvr, th dpltion of th convntional sourcs has raisd th world s concrn. In addition, convntional sourcs for powr gnration produc a lot of wasts and pollutants that contribut to acid rain, smog and climat chang. As a rsult, rnwabl sourcs for powr gnration hav bn xplord. Rnwabl nrgis such as hydro, solar, wind, gothrmal and so on ar clan, abundant and flxibl to match th load growth []. Rnwabl nrgy systm also dos not rquir nvironmntal clanup or wast disposal systm. Malaysia has th tnth highst rainfall ovr its land ara in th world. Th annual rainfall in Malaysia is 289 cubic cntimtrs of watr volum pr squar cntimtr of land ara [2]. This statistic shows that hydro nrgy has a grat potntial to b dvlopd in Malaysia. Th watr is rsrvd in th damp at highr ground. Th potntial nrgy of th running watr is usd to rotat th mchanical turbin. Th mchanical turbin howvr will driv th gnrator and consquntly, th lctricity will b gnratd [3]. To upgrad th quality of lif of th rural ara, nrgy nds to b supplid to th rural ara. Howvr, th utility company rfuss to do so bcaus of high installation cost. In addition, popl from th rural ara consum only small amount of nrgy, which th utility company will not arn any profit from transmitting powr to rural aras. In addition, th utility company nds to spnd for maintnanc of long distanc powr supply. Thus, stand alon powr gnration systm is ndd to supply nrgy to rural aras. Th gnratd powr by micro hydro powr gnration systm is distributd to th usrs. Load powr howvr changs with tim as thr is diffrnt load dmand. It will affct th rotor spd of th gnrator. Corrctiv action nds to b takn to maintain th rotor spd at normal spd to rtain th quality of lctricity. Good quality of lctricity should hav constant voltag and frquncy. PID controllr is applid to th systm to control th gat of th running watr and thus control th rotor spd of th turbin. Th PID controllr is vntually rplacd by Fuzzy PID controllr sinc Fuzzy PID controllr is abl to tun th paramtr of th PID automatically at diffrnt conditions. Fuzzy PID controllr is xpctd to prform bttr than PID controllr vn thr is disturbanc in th systm. 2.0 METHODOLOGY 2. Micro Hydro Synchronous Gnrator A micro hydro powr gnration systm consists of a hydro turbin and synchronous gnrator. Th block diagram of micro hydro powr gnration systm is shown in Figur. Error PID Controllr Hydro Turbin Comparison of powr and rotor spd Rotor Spd Synchronous Gnrator Load Output Activ Powr Figur : Block Diagram of Micro Hydro Powr Gnration Systm Hydro turbin is usd to driv th synchronous gnrator which th rotation spd of hydro turbin is vry much dpndnt on th spd of watr flow. Howvr, th spd of watr flow can b controlld by controlling th valv or gat of th hydro turbin. By comparing th turbin rotor spd and th dmandd lctrical powr, rror is dtrmind and th corrctiv action will b takn by PID controllr. Th stability of th gnratd powr is takn into considration sinc th load powr is changing with tim. As th load powr is changing, th rotor spd of th synchronous gnrator will b changd. Th rotor motion is dtrmind by Nwton s scond law [4], givn that 253
Procdings of th 2nd Sminar on Enginring and Information Tchnology Jα m (t) = T m (t) T (t) = T a (t) () whr J = total momnt of inrtia of th rotating masss, kgm 2 α m = rotor angular acclration, rad/s 2 T m = mchanical torqu supplid by prim movr minus mchanical losss, Nm T = lctrical torqu that accounts for th total thr phas lctrical powr output of th gnrator plus lctrical losss, Nm T a = nt acclrating torqu, Nm Th mchanical losss and lctrical losss ar nglctd. T m and T ar positiv for gnrator opration. In stady stat opration, T m and T ar qual and thrfor, th nt acclrating torqu is zro. Howvr, if th load powr dcrass, T will b rducd. T m will b gratr than T. Th nt acclrating torqu will b positiv rsulting in positiv rotor angular acclration. Th rotor spd of synchronous gnrator will b incrasd. This situation could caus th lctrical frquncy of gnratd powr bing varid and affcts th quality of gnratd powr. Thus, a PID controllr is introducd to control watr flow basd on th fdback output rotor spd and output activ powr of th synchronous gnrator. 2.2 PID Controllr Th lctrical load powr and th mchanical turbin rotor spd hav bn compard as th modl is simulatd in pr unit (pu) form. Figur 2 shows th structur of rror acquisition systm to obtain th rror signal which is th input of PID controllr. Rfrnc rotor spd Synchronous gnrator rotor spd 2 Synchronous gnrator output activ powr Rfrnc Mcahnical powr Diffrnc of Rotor spd Diffrnc of Powr Input rror Figur 2: Structur of Error Acquisition Systm Th diffrnc of rotor spd can b obtaind by comparing th rfrnc rotor spd and actual synchronous gnrator rotor spd. Th comparison of rfrnc mchanical powr and synchronous gnrator output activ powr provids diffrnc of powr. Th rror from th diffrnc of rotor spd and th diffrnc of powr contributs to rotor angular acclration. In stabl situations, th rror should b zro. Howvr, th changing of load dmand causs th output activ powr to b changd and th systm bcom unstabl. Thus, th divrgnc of rotor spd and powr is fd into PID controllr to tak corrctiv action maintaining th stabilization of th powr gnration systm. Th rror signal is fd into a PID controllr. Th structur of PID controllr is shown in Figur 3. Input rror In -C- -C- Kp Ki Kd Intgral /s du/dt Drivativ Figur 3: PID Controllr Output controlld Out Th PID controllr taks corrctiv action to nsur th rror signal is always zro. Th corrctiv action can b don by controlling th gat of watr flow, which th spd of running watr dtrmins th rotor spd of hydro turbin and vntually dtrmins rotor spd of synchronous gnrator. PID paramtrs nd to b st corrctly for th wll functioning systm. Th quation of PID controllr can b shown as blow. u = K p + K i + K d (2) whr, K p = proportional gain K i = intgral paramtr K d = drivativ paramtr = input rror signal u = output controlld signal Th paramtrs K p, K i and K d ar st using Ziglr- Nichols tuning mthod [5]. Ziglr-Nichols tuning mthod provids som ruls and formulas to obtain th paramtrs valu. By only using proportional controllr, th gain is incrasd from zro until ultimat gain at which th output first xhibits sustain oscillation. Th paramtr valu of th proportional controllr is th ultimat paramtr and th corrsponding priod is ultimat tim. Gnrally, proportional controllr racts fast with th rror signal and rducs th rising tim of corrctiv action. Howvr, th systm xprincs ovrshoot and stady-stat rror. Drivativ controllr corrcts th ovrshooting and th intgral controllr corrcts stady-stat rror. 2.3 Fuzzy PID Controllr Th Fuzzy PID controllr has a bttr prformanc compard to PID controllr. Fuzzy PID controllr is an ffctiv tool to dal with disturbancs and uncrtaintis in trm of vagunss, ignoranc and imprcision [6]. Th combination of th fuzzy tchnology and convntional PID forms th Fuzzy PID controllr [7]. Th Structur of Fuzzy PID controllr is shown in Figur 4. Fuzzy rsponss to th rror signal and chang of rror signal and dcid th bst magnitud for proportional gain, intgral paramtr and drivativ paramtr basd on th mmbrship function and rul basd st. Th chang of rror signal is th diffrnc of xisting rror and past rror through a dlay. Th Fuzzy PID controllr has bn dvlopd to hav an individual fuzzy for ach of P controllr, I controllr and D controllr. Th fuzzy 254
Procdings of th 2nd Sminar on Enginring and Information Tchnology controllr works individually to dtrmin ach of th proportional gain, intgral paramtr and drivativ paramtr. In Dlay Dlay Dlay 2 c c c FUZP FUZI FUZD Product Product 2 Product 3.32 Proportional 0.848 Intgral 2.357 Drivativ Figur 4: Fuzzy PID Controllr /s du /dt Drivativ Sinc P controllr and I controllr rquir mor wightag on th rror signal, mor mmbrship functions ar built in input rror in fuzzy infrnc systm (FIS). Manwhil, Fuzzy D controllr has bn dsignd to hav mor mmbrship function in input chang of rror bcaus D controllr nds mor wightag on chang of rror. Mor mmbrship functions ar built in corrsponding input rror and chang of rror in Fuzzy P, Fuzzy I and Fuzzy D to nsur th FIS acts snsitivly to dtrmin th suitabl paramtr. Figur 5 shows mmbrship functions st to input rror, input chang of rror and output gain of Fuzzy P. Input rror Out Figur 5 shows that mmbrship functions of output gain is st mor dviat to th right hand sid. This is bcaus high proportional gain nd to b chosn to corrct high rror. Th implmntd Fuzzy P aims to b mor snsitiv on slcting high gain. Th rul ditor is ssntial to instruct th FIS to dtrmin P gain, I paramtr and D paramtr. Figur 5 shows som ruls for Fuzzy P, Fuzzy I and Fuzzy D. Th input rror of Fuzzy P has bn dividd into 5 fuzzy rgions whras th input chang of rror has bn dividd into 3 fuzzy rgions. Thus, Fuzzy P has 5 ruls to b cratd. Fuzzy I howvr hav 7 fuzzy rgions of input rror and 3 fuzzy rgions of input chang of rror. Thrfor, thr ar 2 ruls to b formd. Fuzzy D hav 3 fuzzy rgions of input rror and 5 fuzzy rgions of input chang of rror. Hnc, Fuzzy D has 5 ruls to b considrd. Th abbrviation of nh, nm, nl, z, pl, pm, ph, n and po ar dfind as ngativ high, ngativ mdium, ngativ low, zro, positiv low, positiv mdium, positiv high, ngativ and positiv. Positiv and ngativ valu of rror and chang of rror hav bn takn into considration and dividd into high, mdium and low dpnding on th concrnd district. Th ascnding ordr of output control action of Fuzzy P, Fuzzy I and Fuzzy D howvr ar L4, L3, L2, L, H, H2, H3 and H4 with L4 is th lowst gain and H4 is th highst gain. Rfrring to Figur 6(a) first rul, if th rror is ngativly high and th chang of rror is ngativ, it is prdictd that th rror is bcoming largr. Thus, Fuzzy P is xpctd to slct th highst gain. Hnc, th output Fuzzy P has bn st to th highst in th rul ditor. Figur 6(b) is th rul ditor for Fuzzy I. Fuzzy I will only corrct th rror whn th chang of rror is approaching zro. Positiv and ngativ chang of rror should hav th minimum of intgral paramtr. Thrfor, 8 th and 9 st rul in Figur 6(b) which consists of ngativ chang of rror is st to hav th lowst intgral paramtr. Considr th 2 nd rul, if th rror is positivly high and th chang of rror is zro, it mans that thr is a high positiv valu of stady stat rror. High I paramtr nd to b slctd by Fuzzy I and thrfor, th output Fuzzy I has bn st to th highst in th rul ditor. Input chang of rror (a) (b) Output proportional gain Figur 5: Mmbrship Functions of Fuzzy P (c) Figur 6: Partially Rul Editor of (a) Fuzzy P, (b) Fuzzy I and (c) Fuzzy D 255
Procdings of th 2nd Sminar on Enginring and Information Tchnology Fuzzy D has mor mmbrship functions in chang of rror sinc chang of rror is th main concrn for Fuzzy D. Considr th 8 th rul in Figur 6(c), if th rror is zro and th chang of rror is zro, thr is no chang of rror. Thus, th output Fuzzy D is st to th lowst. Th dfuzzification mthod applid in FIS is cntroid which th calculatd output control gain is th cntr of ara undr curvd [8]. Th ovrlapping of mmbrship function in output proportional gain of Fuzzy promiss a smooth output gain dspit th nonlinar input rror and chang of rror. Rul viwr of Fuzzy illustrats th way of dtrmining output gain via dfuzzification mthod. Figur 7 shows th rul viwr of Fuzzy P. Figur 7: Rul Viwr of Fuzzy P Th rul viwr displays th fuzzy infrnc procss of input rror, input chang of rror and output gain. Each row of plot rprsnts on rul. Sinc 5 ruls hav bn st in Fuzzy P, thr ar 5 rows shown in rul viwr as Figur 6. By adjusting th indx lin, th output gain can b chckd. If th rror is -0.32, and th chang of rror is -0.36, thr will b ovrlapping of rror and chang of rror in st rul, 2 nd rul, 4 th rul and 5 th rul. Through cntroid dfuzzification mthod, th cntr of total ara undr curv is calculatd. Th output Fuzzy P gain is 0.77 in th particular cas mntiond. 3.0 SIMULATION RESULTS & DISCUSSIONS Th changs of load dmand caus th unbalanc btwn lctrical powr and mchanical powr. It will lad to acclration or dclration of th rotor spd. Fuzzy PID has bn introducd to th systm to maintain th lctrical powr sam as th mchanical powr. Th systm modl is simulatd in pu form. Figur 8 shows th systm modl of Fuzzy PID micro hydro powr gnration systm which has bn dvlopd in MATLAB SIMULINK. Th synchronous gnrator is connctd to th load, whr th load powr is dsignd to b varid to tst th modl. Th rotor spd and th output activ powr of synchronous gnrator hav bn fd back to Fuzzy PID controllr. Th Fuzzy PID will dcid th proportional gain, intgral paramtr and drivativ paramtr to control th gat of th running watr and subsquntly, control th rotor spd of hydro turbin. Thus, th rotor spd of th synchronous gnrator which is drivn by hydro turbin can b controlld to maintain at pu. Figur 9 shows th dsignd signal of dsird load powr at diffrnt tim intrval. Th load powr which is rprsnting th usr load dmand will affct th synchronous gnrator rotor spd. Th problm howvr can b solvd by controlling th mchanical rotor spd of hydro turbin to match to th synchronous gnrator rotor spd. <Rotor spd wm (pu)> <Output activ powr Po (pu)> Fuzzy PID <Stator voltag vq <Stator voltag vd (pu)> (pu)>.0 vrf Dlay c c FUZP FUZI Product Product 2.32 Proportional 0.848 Intgral /s vd Vf vq vstab Excitation Systm m Pm A B Vf _ C Synchronous Machin 200 MVA 3.8 kv Conn Conn2 Conn3 Load Dlay 2 FD 2.357 du /dt c FUZD Product 3 Drivativ Drivativ Dlay spd limit srvomotor position Hydro Turbin Plant 0/3 0.07 s+ s turbin Mux (u(2)/u())^2 prmannt gain 0 /s droop 0.756 Figur 8: Fuzzy Basd Micro Hydro Powr Gnration Systm 256
Procdings of th 2nd Sminar on Enginring and Information Tchnology Figur 9: Load Powr Signal Figur 0 shows th simulatd rsult of micro hydro powr gnration systm. It can b obsrvd that th output activ powr gnratd by synchronous gnrator is similar to th dsird load powr. Th simulatd rsult shows that thr is som ovrshooting whn th changs of dmandd load powr taking plac. This is du to th changing of dmandd load powr altring th rotor spd of synchronous gnrator and subsquntly changing th gnratd output activ powr. It is howvr can b corrctd by Fuzzy PID controllr. Fuzzy PID controllr will tak corrctiv action to maintain th rotor spd at pu. In fact, th rotor spd varis btwn 0.95-.05 pu during th corrctiv action for somtims. Th systm nds tim to sttl down at xact pu. Figur shows signal of input rror, input chang of rror, Fuzzy P gain, Fuzzy I paramtr and Fuzzy D paramtr. Figur 0: Output Activ Powr and Rotor Spd of Synchronous Gnrator Figur : Error Signal, Chang of Error Signal, Fuzzy P Gain, Fuzzy I Paramtr and Fuzzy D Paramtr 257
Procdings of th 2nd Sminar on Enginring and Information Tchnology Th input rror signal is obtaind by comparing th diffrnc of rotor spd and diffrnc of powr (Rfr Figur 2). It can b obsrvd that th proportional gain, intgral paramtr and drivativ paramtr changs with tim, dpnding to th condition of rror and chang of rror signal. Basd on th stting of mmbrship function, and th rul st, th Fuzzy PID is abl to function corrctly to slct th most suitabl gain for rror corrcting action. Initially, th rror signal is vry larg. Th Fuzzy P gain is xpctd to slct larg gain to corrct th rror. Th rror will b rducd and th proportional gain should b rducd with th dcrass of th rror. Rsults show that Fuzzy P gain is slctd as 0.85 initially as th larg rror, followd by proportional gain of 0.60 to corrct th intrmdiat rror and lastly proportional gain of 0.25 to corrct th small rror at th first 00s tim intrval. Ovrall, it was found out that th Fuzzy P slcts larg proportional gain whn th rror is larg and small proportional gain whn th rror is small. Thus, Fuzzy P mt th objctiv of slcting suitabl proportional gain to corrct th rror. Drivativ control rducs th ovrshooting and improvs th transint rspons of th systm [9]. Thus, Fuzzy D should slct larg drivativ paramtr if th systm xprincs larg ovrshooting. Th rror signal in th systm should b controlld to maintain at zro. Sinc initially th rror is larg, ovrshooting has takn plac as shown in Figur. Fuzzy D is abl slct larg drivativ paramtr trying to rduc th larg ovrshooting of th systm. As th dcrasing of th ovrshooting, Fuzzy D is abl to tun th suitabl drivativ paramtr to corrct th ovrshooting. Gnrally, Fuzzy D is functioning wll, rsponding to th rror and chang of rror signal and dciding on suitabl drivativ paramtr for th systm. Intgral control has a function of corrcting th stady stat rror. Rsult from Figur shows that th Fuzzy I is changing th intgral paramtr. Th systm has lss stady stat rror. It can b obsrvd from Figur 0 that although th gnratd output activ powr is matchd with th dmandd load powr, th rotor spd of synchronous gnrator hardly sttling down at xact.0 pu. Intgral control has th ffct of liminating stady stat rror, but it could also caus th transint rspons bcom wors [9]. Th mmbrship function and rul bas of Fuzzy PID nd to b rvisd for bttr prformanc. 4.0 CONCLUSION Micro hydro stand alon powr gnration systm is usd to supply powr to th rural ara which is far away from th city. Th gnratd powr is supplid to th usrs, but th consumabl powr is not constant. In this study, th micro hydro powr gnration systm has bn modlld in MATLAB SIMULINK. Th load which is rprsnting usr s powr dmand is connctd to th synchronous gnrator. Th modl has bn invstigatd with diffrnt dmandd load powr. Fuzzy PID controllr has bn implmntd to th systm to corrct th rror. It can b concludd that Fuzzy PID rsponss wll in ordr to gnrat sam amount of dmandd powr whil maintaining th rotor spd at th ratd.0 pu. Th prformanc of micro hydro powr gnration systm could b improvd. In futur, fin tuning of mmbrship function of Fuzzy PID will b carrid out. Through th tuning, it is xpctd Fuzzy PID can b usd to assist th powr gnration systm achiv fast rspons, lss ovrshooting, lss oscillation and at th sam tim liminat stady stat rror. In addition, dlay will b takn into considration sinc practically th rspons of th gat of watr flow acquir som tim. 5.0 REFERENCES [] Stanly, R.Bull. 200. Rnwabl Enrgy Today and Tomorrow. Procdings of Th IEEE, Vol 89, No. 8: 26-226. [2] Th Lvrhulm Trust. 2006. Rainfall Volum. SASI Group (Univrsity of Shfild) and Mark Nwman (Univrsity of Michigan). Map 0. [3] Mohibullah, Mohd. Amran Mohd. Radzi and Mohd. Iqbal Abdul Hakim. 2004. Basic Dsign Aspcts of Micro Hydro Powr Plant and Its Potntial Dvlopmnt in Malaysia. National Powr & Enrgy Confrnc (PECon) Procdings, Kuala Lumpur, Malaysia: 220-223. [4] J.Duncan Glovr, and Mulukutla S.Sarma. Third Edition 2002. Powr Systm Analysis and Dsign. Brooks/Col [5] Guillrmo J.Silva. Aniruddha Datta. and S.P. Bhattacharyya. 2005. PID Controllrs for Tim Dlay Systms. Birkhäusr [6] Baozhu Jia, Guang Rn, and Gang Long. 2006. Dsign and Stability Analysis of Fuzzy Switching PID Controllr. Procding of th 6 th World Congrss on Intllignt Control and Automation, Dalian, China: 3934-3938 [7] Kyoung Kwan AHN, Dinh Quang Truong, and Yoon Hong Soo. 2007. Slf Tuning PID Control for Hydraulic Load Simulator. Intrnational Confrnc on Control, Automation and Systms COEX, Soul, Kora: 345-349. [8] Th Mathworks. 2009. Fuzzy Logic Toolbox Usr s Guid. [9] T.T.K.Knnth. G.Sainarayanan. W.H.Khoo. 2006. Control of Exothrmal Raction Using Fuzzy PID Controllrs. Procding of th Third Confrnc on Artificial Intllignc in Enginring & Tchnology, Sabah, Malaysia:78-83 258