Smart Electrical Energy Storage System for Small Power Wind Turbines

Transcription

1 1, 1h Inernaional Conference on Opimizaion of Elecrical and Elecronic Equipmen, OPTIM 1 Smar Elecrical Energy Sorage Sysem for Small Power Wind Turbines M. Georgescu, L. Baroe, C. Marinescu, L. Cloea, Members, IEEE * * Transilvania Universiy/Elecrical Engineering and Compuer Science, Brasov, Romania mgeorg@easyne.ro, luminia.baroe@unibv.ro, corneliu.marinescu@unibv.ro, luminia.cloea@unibv.ro Absrac Energy sorage devices and soluions are required for power qualiy and balance wihin wind sysems. In he conex of rapidly expanding of disribued energy sources, he wind energy converers are in he cener of ineres. In his case, he direc dependence of he power generaion capabiliy for a given wind speed represens a major problem of wind energy conversion wih regard o large-scale nework inegraion. This paper proposes an overall soluion which consiss of a wind plan wih a Smar Sorage Modular Sysem (SSMS) where he wind source -as sochasic one- is coupled over a dc bus wih hree sorage modules and a smar grid inerface. The invesigaed sysem consiss of fours modules: a) he sochasic module (variable wind urbine wih PMSG and ac/dc + bus converers), b) he shor erm sorage module (flywheel wih power elecronic converer), c) he medium/long erm module (Vanadium Redox Flow Baery wih power elecronic converer) and d) he medium erm module (Lead Acid Baery). All of modules are inerfaced wih he nework and insulaed loads by a grid inerface module (power elecronic converer, filer and ransformer). To he relaed SSMS, are accomplished compuer simulaions. In order o validae hem, laboraory ess are also presened. I. INTRODUCTION During he recen years, renewable energy sources for elecric power supply have received a considerable aenion due o he global concerns associaed wih he convenional generaion and poenial worldwide energy shorages. In he conex of a marke rapid expansion, for disribued energy, a major ineres is represened by he wind energy sysems. A major problem of he wind energy conversion regarding he large-scale nework inegraion is he direc dependence of he power generaion capabiliy on he given wind speed. Also, an imporan problem is he sysem conrollabiliy, aking ino accoun ha wind energy is wih inermien oupu. Generally, he majoriy of he remoe communiies in he world are supplied wih elecrical energy produced by diesel generaors which are no quie advanageous because of he fuel consumpion and price. In order o reduce energy coss, invesigaion of renewable energy sources represens an ineresing alernaive. In his case he necessiy of energy sorage is becoming more imporan regarding specially he high energy coss during maximum load period and he consanly raising base load in he neworks. Ones conneced o he nework, he energy sorage devices are providing he main following services: frequency sabiliy, balances of he maximal energy need, load balancing and ready-o-use sored energy during he blackous. All hese feaures bring advanages for boh o he energy producers and users providing possibiliies o reduce he energy ransmission coss by energy losses. Based on he Kai Srunz concep of Sochasic Energy Source Access Managemen (SESAM) inroduced in [1], and in order o solve he conflic beween he sochasic naure of he energy source and he need o schedule he power oupu, he auhors proposed a Smar Sorage Modular Sysem (SSMS) able o work for a small wind urbine in neworking condiions and for insulaed loads. II. SYSTEM DESCRIPTION The general block diagram of he SSMS for a small wind farm is presened in Fig. 1. As resuls from he figure, he SSMS consiss in he following main four modules: Sochasic Source Module (SSM) which comprises he renewable energy source (wind) wih sochasic oupu; Shor Term Sorage Module (STSM) based on a flywheel wih Inducion Moor (IM); Medium/Long Term Sorage Module (MLTSM) based on a Vanadium Redox flow Baery (VRB); Medium Term Sorage Module (MTSM) based on a Lead Acid Baery (LAB). An auxiliary module (Converer 5 + Filer + Transformer) is represened by he Grid Inerface Module (GIM) and provides connecions wih he main nework and he insulaed loads. I allows in addiion boh he acive power ransfer and he reacive power generaion. The whole SSMS is managed by a Global Sysem Conrol (GSC) which auomaically conrols all he modules hrough local conrol unis (LCU) and is based on fuzzy logic algorihms. All he modules are inerconneced hrough a dc bus. The acive power and reacive power oupus o he grid are adjused by he GIM. The GSC is hierarchically srucured being associaed wih he LCU conrollers. The GSC and LCU mainain by conrol he desired value of he dc bus volage across he capacior in order o obain he desired oupu power value. The designed SSMS includes he following desirable feaures: based on renewable energy, acive & reacive power deerminisic generaion, clean energy, good conrollabiliy and efficien mainenance coss /1/$6. '1 IEEE 119

2 Fig. 1. Block diagram of he SSMS for wind sysems. The cos and mainenance efficiency of SSMS is promoed by he design ino modules wih well defined inerfaces and conrol unis. A. Sochasic Source Module The SSM comprises he following pars: a) he wind energy source wih sochasic oupu; b) he Permanen Magne Synchronous Generaor (PMSG) and c) he power Converer 1. To describe he wind energy source wih sochasic oupu is considered an aerodynamic wind mahemaical analysis based on he following main parameers: - aerodynamic wind power: - aerodynamic orque: - ip speed raio: 1 3 P W = ρπr vwindc p ( λ, β ). (1) 1 3 T W = ρπ R vwindc p ( λ, β ) / λ. () WTR λ. (3) = ω v wind R where: ρ is he air densiy, R he roor blades lengh, v wind is he equivalen wind speed, λ is he ip speed raio, Cp(λ,β) is he power coefficien, β is iling angle of wind urbine roor and ω WTR is he angular velociy of he roor blades. The wind generaor roaes wih a variable speed and generaes a variable power which sricly depends on he wind speed. In Fig. is depiced he speed/ime characerisic of he wind generaor, for a shor ime and in he absence of sorage sysem. Taking ino accoun his characerisic, and he wind mahemaical model Kaimal [], [3] i was depiced he aerodynamic model in Malab/Simulink, as shown in [3]. I was used he Kaimal mahemaical model because i is in accordance wih he Danish Sandards. Anoher imporan parameer for he wind urbine is he power curve as relaionship of C p and λ which is represened in he Fig. 3. Fig.. Wind generaor speed. For a given wind speed, he power capured by he wind urbine highly depends on C p. Usually he power curve characerisics are included in he warrany assessmen procedures as par of he wind plan commissioning. Turbine oupu power [p.u.] maximum power curve 8 m/s 7 m/s 6 m/s 5 m/s 9 m/s 1 m/s 11 m/s 1 m/s Turbine speed [p.u.] Fig. 3. Wind urbine power characerisic. 1193

3 The PMSG is equipped wih permanen magnes and has no damper windings. Fig. 4 shows he equivalen circuis in he d-q coordinaes ha synchronously roae wih angular speed ω. where: J is he flywheel ineria which roaes wih he angular speed ω. P Gwind P Gne i d R a i cd i d ωl i q q P ref Energy sorage v d R c v d Energy delivery (a) d axis equivalen circui Fig. 5. Waveforms o conrol he sysem flywheel-wind generaor. vq i q R a i cq R c i q v q ωldi (b) q axis equivalen circui Fig. 4. Equivalen circui of he PMSG. d ωψ a As indicaed in Fig. 1, he Inducion Moor (IM) is suiable o conver he elecromechanical energy in accordance wih he following waveforms (Fig. 5), where P Gwind is he power provided by he wind generaor (as known one), P Gne is he power which is provided by he flywheel ino he nework and he P ref is he reference power, calculaed as follows: P ref = P P. (7) Gne Gwind The PMSG orque is expressed as [4]: T G 3 Pn = [ ψ aiq + ( Ld Lq ) idi q]. (4) where: ψ a is he magne flux-linkage, P n is he number of pole pairs, L d, L q are he d- and q-axis inducances and he currens are indicaed in he figure. For his applicaion is used a salien pole PMSG wih he model in roor oriened axis and he currens as sae variables. The elecromagneic orque conrol rule for he PMSG is he following one: T e = k Ω W = K f. (5) where: k, K are consans, Ω W is he wind urbine roaions number and f is he PMSG saor frequency. The power Converer 1 consiss in a diode recifier and a boos converer. To maximize he wind urbine oupu power and adjus he PMSG speed, a Maximum Power Poin Tracking (MPPT) conrol is used [5]. Depending on he wind speed, he MPPT conrol adjus he power ransferred o bring he urbine operaing poins ono he maximum power curve, as shown in he Fig. 3. This leads in changing he angular velociies ω WTR in order o generae ac volages a differen frequencies. The PMSG ac waveforms are recified in he diode converer. The boos converer (buck chopper) helps o obain he dc bus desired volage. Here, he capacior is an energy buffer for he generaor. B. Shor Term Sorage Module The STSM consiss in a flywheel ha sores kineic energy, based on he following equaion: Jω E k =. (6) If P ref >, means ha exiss energy in excess which can be sored. If P ref <, a lack in energy exiss and i will be replaced by he sored energy. To conrol he IM, a DTC conrol scheme has been considered and implemened in he laboraory, as presened in [6], [7], [8], [9], [1]. The DTC is used o conrol he IM, because he auhors esimaed a meaning of 5% decrease in calculus ime of DSPs comparable wih he vecor conrol mehod. To mainain he flywheel in a secured operaing mode a General Monioring Sysem has been designed by he auhors. I is a smar one because i works based on fuzzy logic conrol. The inerface o he dc bus is achieved by he power Converer which is a PWM volage source inverer (VSI). The dc bus curren is supplied by he Converer and is necessary o have a correc esimaion of i because deermines he volage V dc value which mus kep consan. C. Medium/Long Term Sorage Module The MLTSM is based on he Vanadium Redox Baeries (VRB) as sorage source and he Converer 3 which provides a consan dc volage o charge/discharge he baery. The auhors consider ha VRB is a viable opion for boh volage and frequency conrol [3], [5], because of he main following advanages: more flexibiliy, higher efficiency, accepably for larger deph of discharge, lower mainenance coss, longer life span, more environmenally friendly and lower cos per kwh of energy sorage. The block diagram of VRB sysem configuraion, he conrol mode of he bidirecional charge conroller and Simulink modeling are presened in [9]. The bidirecional charge conroller provides suiable charging condiions conrolling he curren flow o avoid overcharge for he baery proecion [5]. Because he charge conroller, included wihin he Converer 3, has been modeled as a lossless device, 1194

4 he VRB curren is calculaed as follows: I V I DC DC VRB =. (8) VVRB where: V DC is he dc volage a he oupu of he buck-bus converer (included in he Converer 1), I DC is he dc bus curren and V VRB is he VRB volage. In his paper, he VRB is used as sorage device only for he insulaed loads. D. Medium Term Sorage Module The MTSM is based on he he Lead Acid Baeries (LAB) as sorage source and he Converer 4 which is a VSI-PWM inverer buil wih MOSFETs in order o operae wih high currens a low volages. On he dc side (LAB side) of he inverer, an elecrolyic capacior is se in order o provide an inverer sable bus volage. A simplified equivalen model of he VSI-PWM inverer is presened in he Fig. 6, as follows: E K Q b = E b Q i d b. (11) where: E b is he no load volage a he LAB raed charge, K is he polarizing volage, Q is he LAB capaciy and I b is he LAB discharging curren. As inpu of he charging/discharging conroller is used as parameer he SOC of LAB, defined as follows: Q n SOC [%] = SOC [%] Q ibd (1) where: Q n is he raed capaciy of LAB. To have an opimal conrol of he charging/discharging conroller is recommended he minimum of % value of SOC. Fig. 6. The simplified equivalen model of he VSI-PWM inverer. As seen from he Fig. 6, he dc side of he inverer is a conrolled curren source and he ac side is a volage source based on he following equaions: i dc _ inv v v inv i = η v inv v b inv b. (9) = D. (1) where: D is he modulaion facor, η is he inverer efficiency and L σ is he sray inducance of he ransformer. The equivalen scheme of he LAB consiss in a conrolled volage source (E b ) in series wih he inernal resisance (R in ) and he LAB volage (v b ), as shown in he Fig. 7. Fig. 7. Equivalen model of he LAB. I is known ha he E b volage depends on he charging sae, baery ype, emperaure and is expressed by he following relaionship: III. SIMULATION AND EXPERIMENTAL RESULTS The proposed sysem has been mahemaical modeled and compuer simulaed using he Malab/Simulink sofware package. To validae modeling and compuer simulaions, by pracical resuls, a es laboraory bench was buil in he laboraory. I consiss in: IM moor of 5,5 kw, 95 ro/min conrolled by a dspace sysem DS113; PMSG of kw, 4 ro/min, lead by he IM moor and conrolled by a dspace sysem DS113; flywheel which consiss in an IM of 3 kw a 15 ro/min conrolled by a PWM inverer and using DTC for a maximum dc bus of 4-4V. The flywheel ineria is of,15-,65 kgm. The flywheel is conrolled by a dspace sysem DS 114. VRB sysem has been replaced in he laboraory of Transilvania Universiy of Brasov by a lead acid baery bank of 56 kv/11a, 6 kw. LAB sysem which consiss in a series of 1 baeries (each of 1V/6Ah). Laboraory load sysem: elecrical sove of 1kW wih conrol seps of,5kw and a radiaor of kw wih conrol seps of 1kW. The simulaions and laboraory experimens consis in: sysem operaion in he case of sorage sysem absence; sysem operaion wih he flywheel conneced only wih he nework for a shor ime; sysem operaion wih VRB conneced only wih he insulaed loads; sysem operaion wih LAB conneced only wih he laboraory load sysem. A. Operaion in he case of sorage sysem absence The wind generaor roaes wih variable speed generaing variable power which depends on he wind speed (see Fig. 1). In his case, aking ino accoun he whole sorage sysem 1195

5 absence, he delivered power in he nework is depiced in he Fig. 8, as follows: Fig. 8. Delivered power in he nework (case of he sorage sysem absence). B. Operaion wih he flywheel conneced The firs simulaed siuaion for he IM of he flywheel i was he following: a idle saring, he magneic flux has raed value and he angular speed is zero, as seen in Fig. 9. compuer simulaion resuls, as depiced in Figs As example, a consan wind speed of 1 m/s, he urbine power of 3 kw can supply he oal power of 4 kw requesed by loads. Figure 1 shows ha he difference of power is supplied by he charging/discharging operaing modes of VRB, which is able o mainain he power balance of he sysem. The VRB parameers for he maximum Sae of Charge (SOC) are 56 V/ 11 A/ 6 kw. In he Figs. 13 and 14 are depiced he VRB volage and curren waveforms for a baery iniial SOC of 8%. Power [W] Loads VRB Wind urbine Time [s] Fig. 1. Power balance of he sysem provided by he VRB. Fig. 9. Angular speed during he idle saring of he flywheel IM. Afer one second, a sudden orque is applied o he flywheel, as reference orque for he IM (see Fig. 1). VRB Volage [V] Fig. 1. Elecromagneic orque of flywheel inducion moor, a saring. The flywheel IM sars a raed flux, bu afer he speed of 153 ro/min, he moor field weakening begins (see Fig. 11). VRB Curren [A] Time [s].81 Fig. 13. VRB volage and curren waveforms VRB SOC [p.u.] Fig. 11. Flywheel IM curren, a saring. Since of he field weakening, he IM power is limied o is raed value for a speed of n=3 ro/min. In his case he flywheel ineria is considered of J=,5 kgm. Based on he Malab/Simulink implemened model, he characerisics during he operaing mode of he sysem wind generaor-flywheel are made in he absence of he reacive power and are presened in [1]. C. Operaion wih he VRB Conneced In his case, he SSM works in parallel conneced wih he MLTSM and boh of hem coupled wih he insulaed loads hrough he GIM. Based on he VRB mahemaical and Malab/Simulink implemened models [9], are presened Time [s] Fig. 14. VRB sae of charge. All hese compuer simulaions are made for he seady sae of PMSG and for a dc bus volage of 5 V. D. Operaion wih he LAB Conneced In his case he SSM works in parallel conneced wih he MTSM based on he LAB and is coupled wih he laboraory load sysem. To have good resuls wihin he pracical laboraory simulaion of he wind sysem working wih he LAB sorage coupled wih he load sysem, he auhors have 1196

6 buil a simulaor for he wind urbine conrolled in real ime by a dspace DS113 sysem and depiced in Fig. 15. Fig. 15. Block diagram of he wind sysem simulaor. To have an opimum conrol of he sysem, near he simulaor are joined monioring elemens for mechanical speed, orque and wind speed. Before of simulaor saring i mus o sar he Danfoss inverer a zero speed reference value. As pracical laboraory resuls are presened he waveforms of he LAB volage, curren and sae of charge (SOC). In his case, wind speed increase from 5 m/s (a =1s) o 9 m/s (a =6s). The enire energy load demand is considered.5 kw. During his process, he LAB volage increases by abou V (Fig. 16). As can be seen in Fig. 17, a wind speed of 5 m/s he power produced by he wind urbine (approximaely 3 W) canno supply he enire energy load demand (.5 kw), herefore he baery will supply he difference. The iniial LAB SOC is considered 8 %. In he ransien regime, he baery SOC decreases in order o ensure he sable supply for he loads (Fig. 18). Fig. 16. The LAB volage variaion. Fig. 17. The LAB curren variaion. IV. CONCLUSION This paper presens a smar sorage sysem designed and used for a small power wind farm, which has a modular and flexible srucure. I is able o deliver power in sandard neworks by using as sorage module a flywheel. For insulaed loads he sysem uses as sorage modules a VRB and LAB ones. The power ransfer beween he individual modules is performed over a dc bus. Through his muli-level sysem deerminisic wind power oupu o he grid is made possible in differen ime frames. All he modular seup is conrolled by a smar general sysem based on fuzzy logic algorihms. This one provides efficien coordinaion and reduces he coss. ACKNOWLEDGMENT This work was suppored in par by he Romanian Minisry of Educaion, Research and Innovaion hrough conrac CNCSIS-IDEI 134/7. V. REFERENCES [1] K. Srunz and E. K. Brock, "Hybrid plan of renewable sochasic source and muli-level sorage for emission-free deerminisic power generaion," Proceedings of he Cigre-IEEE Inernaional Symposium of he Elecric Power Delivery Sysems, Monreal, Canada, Ocober 3. [] P. Rosas, "Dynamic influence of wind power on he power sysem," Ph.D. Thesis, Secion of Elecric Power Engineering, DTU, Denmark, March 3. [3] L. Baroe, L. Cloea, "MPTT conrol of a variable - speed wind urbine," Bullein of he Transilvania Universiy of Brasov, vol. 13, series A1, ISSN , Brasov, Romania, pp , 6. [4] T. Nakamura, S. Morimoo, M. Sanada and Y. Takeda, "Opimum conrol of IPMSG for wind generaion sysem," Power Conversion Conference, IEEI, pp , vol.3,, Osaka, Japan. [5] L. Baroe, R. Weissbach, R. Teodorescu, C. Marinescu and M. Carsea, "Sand alone wind sysem wih VRB energy sorage," Proceedings of Inernaional Conference OPTIM 8, vol. II-B, pp , Brasov, Romania, May 8. [6] C. Lascu, I. Boldea and F. Blaabjerg, "Direc orque conrol of sensorless inducion moor drives: a sliding mode approach," IEEE Trans. Ind. Appl., vol. 4, no., pp , Mar/Apr. 4. [7] G. Cimuca, M. M. Radulescu, C. Saudemon and B. Robyns, "Performance Analysis of an Inducion Machine based Flywheel Energy Sorage Sysem associaed o a Variable Speed Wind Generaor," Proceedings of he Inernaional Conference OPTIM 4, Brasov, Romania, May 4. [8] L. Leclercq, C. Saudemon, B. Robyns, G. Cimuca, and M. M. Radulescu, "Flywheel energy sorage sysem o improve he inegraion of wind generaors ino nework," Proceedings of he inernaional Conference ELECTROMOTION 3, vol., Marrakesh, Morocco, nov. 6-8, 3. [9] L. Baroe, C. Marinescu and M. Georgescu, "VRB modeling for sorage in sand-alone wind energy sysems," Proc. of he Power Tech 9 IEEE Conference, ISBN: , Buchares, Romania, June/July 9. [1] L. Baroe, M. Georgescu, C.Marinescu "Smar sorage soluion for wind sysems" Proc. of he Power Tech 9 IEEE Conference, ISBN: , Buchares, Romania, June/July 9. [11] M. Georgescu, "Elecrical energy sorage sysems," Romanian Research Naional Cener, Tech. Rep., Projecs IDEI 134/7 and e-farm 134/8, Nov. 8. Fig. 18. The LAB SOC variaion. 1197