Univ lor ABS resea view large achi effec num Dyn they the f unst mod and, a H turbi wind mea orien for V uniq betw high macr the n inter espe impr takin
A CO
STRUC
Aless Dept. of In Università bianchi F Lorenzo Fe DESTEC versità degl Pisa renzo.ferrari@ Pisa, Ita STRACT The Darrieus archers and m w of small-siz e-scale offsho eve a more ctive design o merical techniq namics are con y provide a deflow, which i eady aerodyna
In this study, del was applie especially, th H-shaped Dar ine was studie d tunnel, in surements, in nted to the co VAWT simul que set of ex ween simulati hlighting the ro-structures a numerical mod On these ba rest for the sc ecially as a f roving the un ng place past D
OMBINED
CTURE AN
sandro Bian ndustrial En degli Studi ni@vega.de Firenze, Italy errari C li Studi di @unipi.it aly s concept is i manufacturers ze installation ore rotors mou efficient, s of Darrieus V ques making nsidered as th etailed and coms instrumenta amics of the r a two-dimens ed to analyze he wake chara rrieus VAWT ed experimen ncluding bot n the frame onstruction o ations. Thank xperimental d ions and exp main aerody and discussing del.
ases, the pre cientific com first support nderstanding o Darrieus wind
EXPERIM
ND PERFO
chini gineering di Firenze .unifi.it y Giaco Dipartime Politecn giacomo.p Mil increasingly w of wind turbin ns in the built unted on floati structurally s Vertical Axis use of Com he most prom mprehensive al for the unde rotor.sional U-RAN both the ove acteristics on T for micro ntally in a larg th performan of a compre f an experim ks to the avail data, systema periments we ynamic pheno g the predictio esent study i mmunity worki to future an of the unstead d turbine blade
MENTAL A
ORMANCE
Fran Dept. of In Università balduz omo Persico ento di Ener nico di Milan persico@poli ano, Italy welcome by b nes, especially t environment ing platforms. ound and co Wind Turbin mputational Fl mising tool, si representation erstanding of NS computatio erall performa the mid plane ogeneration. T ge-scale, open nce and w ehensive activ mental benchm lability of suc atic comparis ere carried o omena and fl on capabilities s considered ing in this fi nalyses aimed dy aerodynam es. ProceeAND NUME
E OF A H-S
ncesco Bald ndustrial En à degli Studi zzi@vega.de Firenze, Italy o rgia no imi.it V Dip P vince both y in t or . To ost-nes, luid ince n of f the onal ance e of The n-jet wake vity mark ch a sons out, flow s of d of ield, d at mics INT the sui ins see (e. are eas po (Pa han em Un fro im of beh sta (Ra and for com enh the to p flo how exp piv edings of theERICAL AN
SHAPED D
duzzi ngineering i di Firenze e.unifi.it y Vincenzo Do partimento d olitecnico d enzo.dossen Milano, I TRODUCTIO Vertical Ax e Darrieus con itable for new stallations in c e Balduzzi et g. Borg et al e mainly conn se of install sitioning of araschivoiu [3 ndling of tur missions and nfortunately, t om that of con mprove efficien the physical havior is need all (Simao-Fe ainbird et al d so limited c r them when d Within th mputational re hancing the ese rotating bl provide a deta ow. Due to th wever, the p perimental da votal to tune e 1st Global P Jan 16-1NALYSIS
DARRIEUS
G Dept. of I Università giova ossena di Energia di Milano na@polimi.it taly ONxis Wind Tur nfiguration, h w types of ap complex terrai al [1]) or of [2]). The reas nected to the lation and m generation 3]). Moreover, bulent and co high reliab the efficiency nventional ho ncies further, l phenomena ded. Some of rreira et al [6]), are inde orrections, or designing the r is context, esources, CFD comprehensio lades, since it ailed and comp
he complexity possibility of ata to be us the numerical Power and P 18, 2014, Zu
GPPF
OF THE W
S WIND TU
iovanni Fer Industrial E à degli Stud anni.ferrara@ Firenze, Ita Lore Dept. of and Mec Univer lorenzo. Tr rbines (VAW have been rec pplications su ins (e.g. the u ffshore large asons for this superior mec maintenance, equipment , Darrieus VA complex flow bility (A. Bh y of Darrieus orizontal axis a more in de that govern these phenom [5]) or flow eed not comp r no correction rotors. thanks to D can play a fu on of the flu t represents th mprehensive re ty of this typ f having relia sed as valida l tools. Withi Propulsion Fo GPPF urich, Switzer www.pps.gF-2017-WAKE
URBINE
rara ngineering di di Firenze @unifi.it ly enzo Battisti f Civil, Envir ch. Enginee rsità di Tren .battisti@uni rento, Italy WTs), especial cently identifie uch as small-urban environm floating platf increasing su hanical simpl afforded by at ground AWTs present s, with low n hutta et al turbines still wind turbine epth understan n Darrieus tu mena, e.g. dyn curvature ef pletely unders n at all, are ap the incre fundamental ro uid structures he only model epresentation o pe of simulat able and det ation test casin this framew orum 2017 rland lobal
-51
e i ron. ring nto tn.it lly in ed as -scale ment, forms ccess licity, y the level good noise [4]). lacks s. To nding urbine namic ffects stood pplied asing ole in past l able of the tions, tailed es is work,an e carri the f parti [8]. tropo flow diffe mak futur than data the plan dime som [9]). the p RAN appr deem pron on D simu are resou show estim term here struc VAW blad to th in Fi Figure 1. D H-shaped xtended expe ied out in coll frame of a N ially documen Throughout t oskien-blade w conditions, erent aerodyn ke results ava re projects on nks to the avail , a systematic experimental ne and the ensional U-R e of the autho The main ae prediction cap One may no NS simulation roach to this med to produc ne to breakdow Darrieus VAW ulations towar however still urces needed. wed that the ap mation of the ms of instantan e thought to b ctures attended WT TEST MO The turbine p de, H-shaped D he support stee ig. 1. (a) Design (a) a d VAWT und rimental resea laboration wit National found nted in Dosse this project, tw and open-H-b with the ai namics of the ailable as an n similar ma lability of suc c cross-compa results of th numerical o RANS comput ors in the last erodynamic ph pability of the
otice that the ns is probab kind of analy ce vortices th wns than real WTs is inde rds fully three l hampered b . On the other pproach here aerodynamic neous torque be also exploi d in the wake ODEL prototype cons Darrieus VAW el pole by two (b) nd picture ( der conside arch program th the Univers ded project, w ena et al [7] a wo rotor archi blade) were t m to give i e two config experimental achines. In th ch a unique se arison was car he H-shaped r
ones coming tational mode
few years (e. henomena are numerical mo e use of two bly not the m
yses since 2D hat are more c 3D ones. The ed trying to e-dimensional by the enorm r hand, recent proposed can cs of these m profile. On t itable to asses of a H-Darrie sidered in this WT. The blad o aluminium s ) b) of the eration. on VAWTs w sity of Trento whose results and Persico e itectures (clos ested in seve nsights into gurations and l benchmark he present stu t of experimen rried out betw
rotor at the m g from a tw el developed .g. Balduzzi e e highlighted odel is discuss -dimensional most appropr D simulations coherent and l e present resea move the C l analyses, wh mous calculat studies (e.g. n give an accur machines, even these bases, i ss the main fl eus rotor. s study is a thr des are connec spokes, as sho was o, in are et a. sed-eral the d to for udy, ntal ween mid wo-by et al and sed. U-riate are less arch CFD hich tion [6]) rate n in it is flow ree-cted own 1.0 by det EX aga aut the as the exp wh Wi per Mi sec the for 55 scr tur the ob con blo “op up rot cla blo [7] rep con Ins com me bet thr Ins fre for am wa the the vel sin the abo wir The rotor h 028 m), and it un-staggered tails on the lab XPERIMENT In this pape ainst wind tu thors in a pre e frame of thi performance e turbine wak perimental la hile full details
ind Tunnel The measu rformed in the ilano (Italy). I ction (4.00 m e aerodynamic r micro-gener m/s. A comb reens provides rbulence inten Since the so e wind tunne served in co nfined (or “cl ockage, furthe pen chamber” stream of the tor directly f assical correla ockage is neg ]; therefore, fr presentative o nsidered for c strumentatio The experim mbining an easurements tween elastic rough the shaf Data were struments) da equency of 2 r each working mount of data f as computed b eory; the read e uncertainty locity magnitu ngle-sensor ho e rotors. Unce out 2% after re probe wa has a swept ar t is equipped d NACA0021 boratory mode TAL CAMPAI er, CFD simul unnel measure
vious, but rec s experimenta measurements ke and perfor ayout is repo s on the test ca urements rep e large-scale w t consists of a wide, 3.84 m c characteriza ration, up to bination of ho s high flow qu sity at the inle olid blockage l chamber, a nventional w losed chambe er tests were c ”) configuratio diffuser leadin facing the up ations for free gligible in the ree-jet wind tu
f the open fie omparison wi on and Data mental turbine absolute en and a preci joints, to m ft. registered us ata acquisitio kHz. Recorde g condition, in for time-avera by resorting to der is referred as a function ude and the tu ot wire probes rtainty in the calibration in as mounted a rea of 1.5 m2 with dihedral 1 profiles (c el are reported IGN
lations are tho ements perfor cent, experime al activity, ae s were carried rmance. A br orted in the ampaign can b ported in t wind tunnel of a closed-loop m high and 6 ation of real-s o a maximum honeycombs a quality, resulti et lower than of the rotor a a relevant blo wind tunnel te er”) configura carried out us ion, by remov ing to the fans pstream tunne e-jet blockage e present case tunnel results ield configura ith CFD simul a Processing e performance encoder for ision torque measure the t sing a Comp on board, se ed time series n order to gua aging. Measur o the classical d to [7] for th of the TSR. urbulence fiel s were travers velocity mea n a low-speed aligned with 2 (H x D, 1.5 l blades comp = 0.086 m). d in [7]. oroughly comp rmed some o ental campaig rodynamic as d out to invest rief review o following sec be found in [7 this paper f the Politecni facility, whos .00m long) al cale wind turb m wind spee and anti-turbu ing in a freest 1%. amounts to 10 ockage effect ests performe ation. To limi ing a ''free jet ving the test s, and installin
el. Applicatio e suggests tha e (below a few
are believed ation, and are
lations. g e was obtaine rotational s meter, mou torque transm act-RIO (Nat etting a sam s lasted 3 min arantee a suffi rement uncert l error propag e quantificatio To investigat d in the wake sed downstrea surements res d jet. The firs the wind s 5 m x posed Full pared of the gn. In s well tigate of the ction, 7]. were ico di e test llows rbines ed of lence tream 0% of t was ed in it the t'' (or room ng the on of at the w %) to be here ed by speed unted mitted tional mpling nutes icient tainty gation on of te the e, two am of sulted st hot speed
direction, thus exposing the wire normal to the wind to measure directly the velocity magnitude. The second probe was mounted in the vertical direction and was rotated during operation around its own axis to measure the transversal velocity components, by exploiting the angular sensitivity of wire. The so-called “triple decomposition” was performed on hot wire data, by extracting the time-averaged, phase-resolved, and turbulent components of the velocity, as discussed by Persico et al [8]. The phase-resolved component was obtained by ensemble averaging the velocity signals, using the encoder as a key-phasor. The streamwise turbulence intensity was determined by extracting the unresolved unsteady fluctuations of the velocity signal. A pneumatic five-hole probe was also applied throughout the tests to measure the pressure level and the 3D flow direction. Measured pressure values confirmed the absence of blockage-induced over-speeds outside the wake region; flow angle measurements with the vertical hot wire and the pneumatic probe also indicated negligible transversal and vertical velocity components in the wake at midspan.
NUMERICAL SIMULATIONS
The two-dimensional simulations presented in this paper were carried out using the commercial code ANSYS®
FLUENT® [10]. The time-dependent unsteady
Reynolds-averaged Navier–Stokes (U-RANS) approach was adopted in a pressure-based formulation. The fluid was air, which was modeled as an ideal compressible fluid following the suggestion of Balduzzi et al [9]. Inlet air conditions were the same monitored during the experimental tests, i.e. a pressure of 1.01×105 Pa, a temperature of 293 K and an inlet
turbulence level of 1%. Some of the authors have recently presented (Balduzzi et al [9]) the assessment and validation of the main settings for a proper CFD simulation of Darrieus wind turbines using the FLUENT® code, whose accuracy has
been also successfully verified by means of experimental data (Balduzzi et al [11]). Moreover, in recent works (Balduzzi et al [12] and Bianchini et al [13]) the accuracy of the proposed numerical approach has been verified exactly on the same case study considered here. For the sake of completeness, the main numerical settings are briefly summarized below. The turbulence closure was achieved by means of the k-ω shear stress transport (SST) model (Menter [14]), coupled with the Enhanced Wall Treatment embedded in the FLUENT® code for the computation of the boundary
layer in the near-wall regions; since the y+ was constantly kept < 1 in the boundary layer, the linear law of the wall is actually used.
The Coupled algorithm was employed to handle the pressure-velocity coupling. The second order upwind scheme was used for the spatial discretization of the whole set of RANS and turbulence equations, as well as the bounded second order for time differencing to obtain a good resolution (Amet et al [15]). The global convergence of each simulation was monitored by considering the difference between the mean values of the torque coefficient over two subsequent revolutions: according to Balduzzi et al [11], the periodicity error threshold was set to 0.1%.
To simulate the rotation of the turbine, the sliding mesh technique was employed, i.e. the computed domain was split into a circular zone containing the turbine, rotating with the same angular velocity of the rotor, and an outer rectangular fixed zone, determining the overall domain extent. The final dimensions of both domains were defined, according to the sensitivity analysis reported by Balduzzi et al [9], so to allow a full and unconstrained development of the turbine wake. In this way, CFD results can be properly compared to experiments obtained in unconfined configuration. A velocity-inlet boundary condition was imposed at the inlet section (40D upwind from the revolution axis). The ambient pressure condition was imposed at the outlet boundary (100D downwind), while a symmetry condition was defined on lateral boundaries (30D).
In the present study, the finest mesh defined during the sensitivity analyses reported by Balduzzi et al [12] was used for all the three considered tip-speed ratios. Even if redundant at TSR=3.3, this refinement level indeed ensured a very accurate discretization of the whole area around the rotor, greatly limiting any bias error due to the numerical discretization. The mesh is of unstructured and hybrid type, with triangular elements in the core flow region, and an O-grid made of quads in the boundary layer. As discussed, the first element height was chosen to guarantee that the dimensionless wall distance at the grid nodes of the first layer above the blade wall does not exceed y+ ~ 1. The expansion ratio for the growth of elements starting from the surface was kept below 1.1 to achieve good mesh quality. The airfoil surface was discretized with 1400 nodes, resulting in 1.2×106 elements for the rotating region, while the
stationary region was discretized with 2.0×105 elements. The
mesh characteristics fully accomplish the criteria based on dimensionless thresholds proposed by Balduzzi et al [16]. Based on the same criteria, in order to limit the Courant Number in proximity of the blades, an angular timestep of 0.21° was used, i.e. the finest one identified by Bianchini et al in [13]. With the described settings, the calculation time to achieve a complete revolution of the rotor is around 24 hours in a 16 CPUs (2.8 MHz each) calculation center. The required number of revolutions to achieve a periodic solution is dependent on the TSR, varying from 10 to 20 revolutions.
Figure 2 finally shows some visual details of the mesh in the rotating region (a) and near the leading edge (b).
RESULTS AND DISCUSSION
To assess the prediction capability of the CFD approach, Fig. 3 compares the curve of experimental power coefficient cP, defined according to Eq. 1 and measured by Dossena et al [7], with the one obtained with present simulations.
3 0 P 2 1 c AV P ρ = . (1)
It is worth remarking that numerical results were corrected to account for the parasitic torque of the supporting struts of the rotor, which were of course not simulated in the two-dimensional approach.
Figu Fig [7] origi [17] matt coef aero proje calcu azim deta the s mod wind betw parti the p both discr ure 2 - Mesh and gure 3 - Com ] and presen s The correcti inal one-dime , assuming a r ter of fact, the fficient for th dynamic drag ection of the r ulated with muthal positio il on the mod same authors del was assess
d tunnel. Upon exami ween experim icular, numer power curve h in terms of repancy, quit h refinemen near the lea
mparison be nt CFD simu struts’ paras ion was intr ensional mode rectangular sh e model consid he section an g during the relative speed a multitube on) on the tan
del, the reader in [17] and [1 sed by means ination of th ments and s ical results ar over the enti f qualitative t te common t in the rota ading edge etween expe ulations corr sitic torque. roduced by r el developed b hape of the str ders a constan nd assumes t e revolution d (variable wi s BEM app ngential direc r is referred to 18], where the s of experime he figure, so simulations i re able to rep ire range of T trend and cP in all CFD ating region (b). eriments fro rected for th resorting to by Bianchini e ruts section. A nt equivalent d
that the indu is given by th the radius proach for e tion. For furt o the analyses e reliability of ental tests in ound agreem is apparent. produce prope TSRs of inter values. A li simulations (a) om he the et al As a drag uced the and each ther s of f the the ment In erly rest, ittle of Da thi con sev det the rou mo cha piv the for we me [7] for 14 a c pro tha op (ar du vor par sta tha the in t the tow kin sta tor dow F arrieus VAWT s region of nditions for a vere flow sep taching even ese condition ughness, the t odel, etc. can
aracteristics votal for the f ese results, th r the purposes To analyse ere selected easurements w ], the tip-spee r comparison .2 m/s, 9.0 m constant angul Figure 4 c ofiles, cT (Eq. By looking at the TSR=1 erating point, round approxi e to the onse rtices detachin rticular, as wi all in the secon
at are convect e downstream
the torque pro The TSR=2 e operating c wards ϑ=90°, nd of airfoil d all occurs, the
rque region be wnwind. Figure 4 - CF Ts, can be no the curve, considerable paration pheno
from the lea ns, additiona trailing-edge n play a fun of such sep final torque pr he CFD appro of the presen the wake char
for analysi were available ed ratios of 1. n, correspond /s and 6.5 m/ ar speed of th ompares the . 2), for the thr T c A ρ = at the torque 1.5 condition with an abrup imately ϑ=60
t and the sub ng from the bl ll be discusse nd quadrant o ed downstream blades, result ofile. 2.4 condition urve, but the
i.e. where it during conven e trend is aga etween ϑ=120 FD torque p three invest oted at low t the airfoils fraction of th omena and la ading edges al elements refinement in ndamental ro parations, wh roduction. Ov oach can be c nt study. racteristics, th is, for whic e. Based on .5, 2.4 and 3. ding to a w /s, respectively he turbine equ CFD-based hree investigat 2 0 4V D T . (2) profiles, one n represents pt stall onset i °), followed bsequent prop lades (see Ra ed later on in t originates stro am by the flow
ting in the inte still lies in th e upwind pea t is commonl ntional turbin ain quite steep 0° and ϑ=180° profiles com tigated TSR ip-speed ratio work in st he revolution, arge-scale vor of the airfoil like the su n the experim ole in setting hich are how verall, by virtu considered rel hree relevant T ch detailed the indication 3 were consid wind speed V y, having assi al to 400 rpm. torque coeffi ed conditions can readily n a highly uns in the upwind by intense rip pagation of m inbird et al [6 the study, the ong macro-vor w and interact ense ripples sh he unstable pa ak is now sh ly located for ne operation. p, with a neg ° and large rip
mparison at t Rs. os; in talled with rtices ls. In urface mental g the wever ue of liable TSRs flow ns of dered V0 of igned . icient . notice stable zone pples macro-6]). In deep rtices t with hown art of hifted r this After gative pples the
cond extra as ty poss field func num dime CFD rake of th and equa for prof trave turbu prof agre expe term From that posit In th pres look repo Fig Finally, TSR dition, with no action is stron ypically found To understan sibility of hav d becomes c ctioning point merical result ensionless vel D data have b es positioned e he experiment “far” 1.5D d ally spaced po at least 3 re files were also Figure 5 rep erse, in the fo ulence profile files, the uns eement is ge eriments, in te ms of flow an m an in-depth CFD data pr tions, i.e. arou hese positions
ent wide ripp king at the com orted in Fig. 6 ure 5 - Near and turbu R=3.3 repre o evidence of ngly unbalanc d for three-bla nd the physics ing experimen crucial In p ts compariso ts and the locity (V/V0) been acquired exactly at the tal traverses, downstream th oints were acq
volutions. Ba o calculated an ports the resu orm of measur es. Super-impo steady disper enerally foun erms of wake d turbulence analysis of th esent a high und Y/D=-0.4 , both the vel les, whose na mputed veloci for a given in r traverse: d ulence inten esents a sta abrupt stall, e ced towards t ade rotors of s s underlying ntal measurem particular, for ons were ma experiments and turbulenc d in correspon e same distanc i.e. “near” 0.7 he shaft. At e quired at a freq ased on these nd added to th ults at TSR= red and compu osed to the com
sion is also nd between
width and def morphology he trends, it is
dispersion, es 4, Y/D=-0.2 an ocity and the ature can be c ity field, who nstant.
dimensionle nsity (b) at T
able function even if the tor he upwind zo similar solidity these trends, ments of the fl r all the th ade between s in terms ce intensity (I ndence of virt ce from the sh 75D downstre each section, quency of 10 k e data, avera e comparisons 1.5 for the n uted velocity mputed avera reported. Go simulations ficit as well a inside the wa worth remark specially in th nd 0.1 <Y/D<0 turbulence fie comprehended ose distribution ss velocity TSR=1.5. ning rque one, y. the flow hree the of ITU). rtual haft eam 100 kHz aged s. near and aged ood and as in ake. king hree 0.4. elds d by n is (a) firs of exp tow ori inc allo ph of far not flu Fig F F Upon exam st two ripples strong mac perienced by t The third wer) is instea iginate from cidence angle ow the flow enomena have
the rotor. Ful r traverse (Fi ticeable, even uctuations inte Considering g. 8 reports the Figure 6 - Ins Figure 7 - Fa and turb mination of the correspond to cro vortices the airfoils, al ripple (locat ad connected the airfoil t is decreasin to reattach t e a major imp ly comparable ig. 7), with n though wit nsity. g the conditio e results again stantaneous r traverse: d bulence inte e figure, one c to the onset an connected to lready discuss ted approxim to another se trailing edge ng again towa to the blades pact on the ov e trends were
the main flo th a general on correspon n for the near
s velocity fie
dimensionle ensity (b) at
can notice tha nd the propag o the deep sed in Fig. 4. mately behind et of vortices e as soon as ards values w s. Both these verall perform also found fo ow structures reduction o ding to TSR traverse. eld at TSR=1 ess velocity t TSR=1.5. at the gation stall d the s that s the which e two mance or the still f the R=2.4, 1.5. y (a)
disp snap 9, a Even term regio velo devi inter in ag pow notic turbu and clear posit rema mea the c of th Fig. inten Fig In this case, t ersion is low pshot of the in more unifor n if the overa ms of wake wid on behind th city deficit hi iation might raction betwee greement with wer coefficient A general de ced in Fig.8( ulence intensi of low magn rly marked, p tion and rela arkable agre surements, a center of the w he tower. To better und 10 reports th nsity in the sam
ure 8 - Near and turbu the velocity pr wer. In furthe nstantaneous rm wake struc all accuracy o dth, a discrepa he rotor, wh igher than tha be connected en the flow an h the slightly previously di ecrease of the (b). With res ity profile pre nitude in the c eaks at the tw ative intensity eement with further peak wake, again s derstand the n he instantaneo me instant in r traverse: d ulence inten rofile is smoo er detail, as velocity field cture was pre of simulations ancy can be he here experim at predicted nu d to a highe nd the tower. y lower value iscussed in Fig e turbulence l spect to the edicted by CF center of the wo sides of the y of these pe h experimen is however d uggesting a st nature of the tw ous distributio time of Fig. 9 dimensionle nsity (b) at T
other and the d testified by d reported in F edicted by CF s is still good ere noticed in ments revealed umerically. S er impact of This evidenc of the predic g. 3. evel can be a lower TSR, FD is almost wake, with tw e wake itself; eaks is again nts. In tun distinguishable tronger influe wo lateral pea on of turbule 9. ss velocity TSR=2.4. data the Fig. FD. d in n the d a uch the ce is cted also the flat two, the n in nnel e at ence aks, ence (a) F Fi tur dif eve int mo dee reg det con me rel dow bel aga wit def TS air ten rev tho tur Figure 9 - Ins gure 10 - In Figure 10 c rbulence inten fferent vortice en more evide tense one, loc ore precisely r ep stall condi gion of the up tached from nvected down echanism disc ated to the wnwind regio low the stall o The main p ain clearly di th an even m ficit in the wa Finally, Fig SR=3.3 for the As discusse rfoils do not e nds to remain volution, with ose described rbulence level stantaneous stantaneous at TS learly suggest nsity are relate es, already no ent thanks to t cated approxi related to the m itions experie pwind part of the airfoil s nstream by t cussed in Fig flow reattach on as soon as one. henomena des istinguishable ore pronounc ke in correspo . 12 and Figur e near and far t ed, in this hi exhibit the on attached to th no formation for lower TSR s are accordin s velocity fie s turbulenc SR=2.4.
ts that the two ed to the prop oticed at TSR= the vorticity a imately at Y/ macro vortice enced by the f the revoluti surface, the the main flow g. 6. The othe hment in the s the incidenc scribed for th e also in the ced discrepanc ondence to the re 13 report th traverse, resp igh-loaded TS nset of deep he airfoil wall n of vortical st Rs. The data ngly lower. eld at TSR=2 e intensity f o local increas pagation of the =1.5, that are analysis. The /D=-0.5, is in es generated b airfoil in the ion; once they
vortices are w, with the er peak is in e last part of ce angle decr he near travers far one (Fig. cy for the vel e central towe he results for t
ectively. SR condition
stall and the s during the w tructures simi dispersion an 2.4. field ses of e two e now more ndeed by the e last y are then same nstead f the reases se are . 11), locity er. the at ns the flow whole lar to nd the
Fig
Fig
gure 11 - Far and turbu
gure 12 - Ne (a) and tur
r traverse: d ulence inten ear traverse rbulence inte dimensionles nsity (b) at T : dimension ensity (b) at ss velocity ( TSR=2.4. nless velocit t TSR=3.3. (a) ty Fi vel bet pre sub tra CO me win CF me app ana pre som com wa pas the dow bla rip ran ma sta and igure 13 - Fa and turb Experiment locity deficit tween data evious TSR bstantially co verse, with ev ONCLUSION In the pre easurements i nd turbine h FD simulation Overall, an easurements a proach is usua alyses. In par edict the perfo me main flo mpatible with In particula ake is charact ssage of macr e upwind regi wnwind regio ade. For an int pples are still
nge of inciden ay lead to stal able part of the
d smoother wi ar traverse: bulence inte s again show in the wake, is satisfying. conditions, oherent with ven smoother t NS sent study, a inside the wa have been co s of the same n interesting a and simulatio ally thought to rticular, the C ormance of th ow structure experimental ar, it has been
terized by ve o vortices det ion after the s on as soon as t
termediate TS clearly disting nce angles exp l onset. Finall e functioning ith less intens
dimensionl ensity (b) at owed a sligh even if the o . As mentio the velocity those meas trends and low
a comprehen ake of a Dar ompared with rotor in the m agreement w ons, even if to be not fully CFD data were he turbine and es whose ef l trends. n shown that ery intense ri taching from t stall angle is the flow tends SR, intense tur guishable in r perienced by ly, for high T curve, the wa se ripples and ess velocity t TSR=3.3. htly more in overall compa ned for the y profiles re ured at the wer dispersion nsive set of rrieus vertical h two-dimens mid plane. as found bet f a 2D U-R y suitable for e able to corr d also to high ffects were t at low TSR ipples, due to the blades eith reached, or i s to re-attach t rbulence and reason of the the airfoils, w SRs, typical o ake is more re turbulence, ev y (a) ntense arison two emain near n. flow l-axis sional tween RANS these rectly hlight fully Rs the o the her in n the to the some wide which of the egular ven if
the velocity deficit is now increased due to the higher load experienced by the turbine.
NOMENCLATURE
A turbine swept area [m2]
cP power coefficient [-]
cT torque coefficient [-]
CFD Computational Fluid Dynamics
D, R turbine diameter, radius [m]
ITU turbulence intensity [m]
k turbulent kinetic energy [m2s-2]
P power [W]
RANS Reynolds-Averaged Navier-Stokes SST Shear Stress Transport
T torque [Nm]
TSR Tip-Speed Ratio
V wind speed [m/s]
VAWT Vertical Axis Wind Turbines
y+ dimensionless wall distance [-] Greek letters
ϑ azimuthal angle [deg]
ρ air density [kg/m3]
ω Specific turbulence dissipation rate [s-1]
Ω revolution speed [rpm]
Subscripts
0 value at inlet ACKNOWLEDGMENTS
Thanks are due to Prof. Ennio Antonio Carnevale of the University of Florence for supporting this study, and to Dr. Andrea Tanganelli for his support in the CFD simulations.
The authors would like also to acknowledge the Company Tozzi-Nord Wind Turbines for the rotor model and the technicians and collaborators of Universitá di Trento and Politecnico di Milano for their precious support before and during the tests.
REFERENCES
[1] Balduzzi, F, Bianchini, A, Carnevale, EA, Ferrari, L and Magnani, S, “Feasibility analysis of a Darrieus vertical-axis wind turbine installation in the rooftop of a building,” Applied Energy, 97, 2012, pp. 921–929
[2] Borg, M, Shires, A, and Collu, M, “Offshore floating vertical axis wind turbines, dynamics modelling state of the art. Part I: Aerodynamics”, Renewable and Sustainable Energy Reviews, 39, 2014, pp. 1214-1225
[3] Paraschivoiu, I, Wind turbine design with emphasis on Darrieus concept, Polytechnic International Press, Montreal, Canada, 2002.
[4] Aslam Bhutta, MM, Hayat, N, Farooq, AU, Ali, Z, Jamil, ShR, and Hussain, Z, “Vertical axis wind turbine – A review of various configurations and design techniques,” Renewable and Sustainable Energy Reviews, 16(4), 2012,1926-1939
[5] Simao-Ferreira, C, van Zuijlen, A, Bijl, H, van Bussel, G, and van Kuik, G, “Simulating dynamic stall on a two-dimensional vertical-axis wind turbine: verification and
validation with particle image velocimetry data”, Wind Energy, 13, 2010, pp. 1-17
[6] Rainbird. J, Bianchini, A, Balduzzi, F, Peiro, J, Graham, JMR, Ferrara, G, and Ferrari, L, “On the Influence of Virtual Camber Effect on Airfoil Polars for Use in Simulations of Darrieus Wind Turbines,” Energy Conversion and Management, 106(December 2015), 2015, pp. 373-384 [7] Dossena, V, Persico, G, Paradiso, B, Battisti, L, Dell’Anna, S, Brighenti, A and Benini, E, “An Experimental Study of the Aerodynamics and Performance of a Vertical Axis Wind Turbine in a Confined and Unconfined Environment”, 2015, ASME Journal of Energy Resources Technology, 137(5) 2015, paper code 051207 (12 pages) [8] Persico, G, Dossena, V, Paradiso, B, Battisti, L, Brighenti, A and Benini, E, “Three-Dimensional Character of VAWT Wakes: an Experimental Investigation for H-Shaped and Troposkien Architectures”, Proc. of the ASME Turbo Expo 2016, Seoul, South Korea, June 13-17, 2016.
[9] Balduzzi, F, Bianchini, A, Maleci, R, Ferrara, G and Ferrari, L, “Critical issues in the CFD simulation of Darrieus wind turbines”, Renewable Energy, 85(01), 2016, pp. 419-435
[10] ANSYS®, Inc., 2015, “FLUENT® Theory Guide”,
release 16.0
[11] Balduzzi, F, Bianchini, A, Maleci, R, Ferrara, G, and Ferrari, L, “Blade Design Criteria to Compensate the Flow Curvature Effects in H-Darrieus Wind Turbines”, J. of Turbomachinery, 137(1), 201, pp. 011006-011006-10 [12] Balduzzi, F, Bianchini, A, Gigante, FA, Ferrara, G, Campobasso, MS, and Ferrari, L, “Parametric and Comparative Assessment of Navier-Stokes CFD Methodologies for Darrieus Wind Turbine Performance Analysis”, Proc. of the ASME Turbo Expo 2015, Montreal, Canada, June 15-19, 2015
[13] Bianchini, A, Balduzzi, F, Ferrara, G, and Ferrari, L, “Aerodynamics of Darrieus wind turbines airfoils: the impact of pitching moment”, J. of Engineering for Gas Turbines and Power, 2016, paper in press
[14] Menter, F, “Two-equation Turbulence-models for Engineering Applications,” AIAA Journal, 32(8), 1994, pp. 1598–1605
[15] Amet, E, Maître, T, Pellone, C, and Achard, JL, “2D numerical simulations of blade-vortex interaction in a Darrieus turbine”, J. of Fluids Engineering, 131(11), 2009, pp. 1-15
[16] Balduzzi, F, Bianchini, A, Ferrara, G, and Ferrari, L, “Dimensionless numbers for the assessment of mesh and timestep requirements in CFD simulations of Darrieus wind turbines,” Energy, 97(15 February 2016), 2016, pp. 246-261 [17] Bianchini, A, Ferrari, L and Magnani, S, “Start-up behavior of a three-bladed H-Darrieus VAWT: experimental and numerical analysis”, Proc. of the ASME Turbo Expo 2011, Vancouver, Canada, June 6-10, 2011.
[18] Bianchini, A, Balduzzi, F, Ferrara, G and Ferrari, L, “Influence of the blade-spoke connection point on the aerodynamic performance of Darrieus wind turbines”, Proc. of the Turbo Expo 2016, Seoul, South Korea, June 13-17, 2016.
