發展黏性耦合三維小板法於風力發電機氣動力分析（英文版）.pdf

AcknowledgementsKouh,forFirst of al l,I am immensel y grat eful t o my PhD supervisor,Profes!t aking me in and giving me t he chance t o push mysel f harder t han J,have ever done before.To my oral defence commit t ee members,t hank you for your pat ronisat ion,and for al l of your suggest ions and comment s,wit hout which,t his t hesis woul d be hal f of what it is t oday(l it eral l y and figurat ivel y).I am very proud of t he work t hat I have done.I have t hat it al so meet s your high st andards.To al l my friends who have support ed me over t he years,t hank you.To everyone from CadLab:Pork,Tony,瓜哥，A-bing,Wayne,才白文，Rex.(Wow,t hereve been a l ot of you over t he years.)And especial l y t o my part ners-in-pain,大隔 小辣，and Daniel.小辣，I owe you big t ime,brot her.Thanks t o Mom and Dad.It was incredibl e having you here wit h us in Wul ai.I know t hat neit her Teresa nor I woul d have made it t o t he end or our degrees wit hout al l t he l ove and support(and del icious food)t hat you gave.Final l y,and most most most import ant l y,t o my gorgeous wife,Teresa-Banat my Amis warrior-princess.I l ove you wit h my whol e heart.Thank you for al l your pat ience when I was grouchy.And for your kindness when I was sad.And for your l ove al ways.The l ast few mont hs have been especial l y t rying on bot h of us,but Im al l yours now.doi:10.6342/NTU201800594AbstractIn addit ion t o t he many t ypical fail ure mechanisms t hat affl ict wind unit s iTaiwan are al so suscept ibl e t o cat ast rophic fail ure from t yphoon-indued ext remel oads.A key component of t he st rat egy t o prevent such fail ures a fast,accurat e aerodynamic design and anal ysis t ool t hrough which a ful l er underst anding of t he aerodynamic l oads act ing on t he unit s may be derived.Present model l ing approaches range from l ow fidel it y,such as t he Bl ade El ement Moment um BEM t heory,t o high fidel it y,such as Navier-St okes(NS)sol vers.The former is fast and comput at ional l y inexpensive,but l imit ed in t erms of fl ow condit ions which may be model l ed,whil e t he l at t er are very comput at ional l y expensive,and t herefore impract ical for design work.To t his end,a viscous-coupl ed 3D panel met hod is herewit h proposed,which int roduces a novel approach t o simul at ing t he severe fl ow separat ion so preval ent around wind t urbine rot ors.The Hess-Smit h panel met hod was adopt ed for t he inviscid cal cul at ions,and an empirical l y based boundary l ayer anal ysis is t hen performed t o det ermine t he separat ion point.The separat ed t hick wake is t hen model l ed as an ext ension of t he surface geomet ry al ong which a const ant pressure dist ribut ion is assumed.The wake geomet ry is det ermined it erat ivel y,and an out er it erat ive l oop is run t o updat e t he l ocat ion of t he separat ion point.As proof of concept,t he proposed met hod was first val idat ed against experiment al and numerical resul t s for several high t hickness wind t urbine airfoil s.At l ow angl es of at t ack,pressure dat a predict ed by t he current met hod showed excel l ent agreement wit h t he experiment al dat a,as wel l as wit h t he referenced numerical dat a,comput ed by an NS sol ver.At higher angl es of at t ack,t he current met hod showed reasonabl e agreement wit h t he experiment al dat a,whil e t he referenced numerical dat a significant l y overest imat ed t he pressure dist ribut ion al ong t he suct ion surface.The abil it y of t he current met hod t o simul at e t he more compl icat ed case of a rot at ing 3D wind t urbine rot or was t hen assessed by code-t o-code comparison wit h RANS dat a for a commercial 2 MW wind t urbine.Al ong t he out board and inboard regions of t he rot or,pressure dist ribut ions predict ed by t he current met hod showed very good agreement wit h t he RANS dat a,whil e pressure dat a al ong t he midspan region were sl ight l y more conservat ive.The power curve predict ed by t he current met hod was correl at ed very wel l wit h t hat provided by t he t urbine manufact urer.Taking int o account t he high degree of comparabil it y wit h t he more sophist icat ed RANS sol ver,t he excel l ent agreement wit h t he official dat a,and t he considerabl y reduced comput at ional expense,t he aut hor bel ieves t he proposed met hod coul d be a powerful st andal one t ool for t he design and anal ysis of wind t urbine bl ades.iiidoi:10.6342/NTU201800594Contents口就委具含寄定善.Acknowl edgement s.Abst ract.Cont ent s.Nomencl at ure.viChapt er 1 Int roduct ion.11.1 Mot ivat ion.11.2 The evol ut ion of wind power.41.3 St at ement of t he probl em.71.4 Lit erat ure review.91.5 Thesis object ive.15Chapt er 2 The inviscid sol ver.192.1 El ement ary pot ent ial fl ow.202.2 The invisicd fl ow over a circul ar cyl inder.212.3 Zhukovskii aerofoil s.262.4 Panel met hods.32Chapt er 3 Val idat ion of t he invisicd sol ver.453.1 Fl ow over a 2D cyl inder/3D sphere.463.2 Symmet rical Zhukovskii aerofoil s.503.3 NACA 4415 aerofoil.543.4 Val idat ion of t he hel ical wake.60Chapt er 4 Fl ow separat ion.714.1 The boundary l ayer.724.2 The separat ed wake.814.3 Viscous-inviscid coupl ing.90Chapt er 5 Case st udy 1:Wind t urbine airfoil s(2D).945.1 Target airfoil s.945.2 Wind t unnel descript ion.975.3 RANS model descript ion.975.4 Resul t s.995.3 Discussion.108Chapt er 6 Case st udy 2:Rot at ing wind t urbine rot or(3D).1106.1 The t arget wind t urbine.Il l6.2 RANS model descript ion.1136.3 Resul t s.1166.4 Discussion.125Chapt er 7 Concl usions&fut ure work.129References.132Appendic A Derivat ion of t he inviscid sol ver.137ivdoi:10.6342/NTU201800594NomenclatureVNabl a operat orV2Lapl ace operat orufl ow vel ocit yDot operat or(Pdoubl et st rengt hwake doubl et st rengt ht rail ing edge doubl et st rengt hs0vel ocit y pot ent ialinner pot ent ialEVpanel l ocal coordinat esosource st rengt hVdist ance from singul arit ynnormal direct ionUg free-st ream vel ocit yo)Rot at ional speed of rot orW0TE t rail ing edge offset angl e6W wake panel azimut h angl ep Pressure8*boundary l ayer displ acement t hickness5*p fl uid densit ytw shear st ress at t he body surface kinemat ic viscosit yI Thwait es XXXA dimensionl ess pressure gradient paramet ers上 l inear dist ance al ong surfacea infl uence coefficient saw wake infl uence coefficient scN user-defined fract ionCf skin-frict ion coefficientCp)i pressure coefficient at panel iCpsep pressure coefficient at separat ion pointE ent rainment vel ocit yH shape fact ort urbul ent shape fact or k wake decay const antdoi:10.6342/NTU201800594N Np Nw Nb Nsw Pi Re Q s SbSw SgUi*Ue&el V%sep XTE Zi Zi zfnumber of vert ex chordwise vert icesnumber of body panel snumber of wake panel s number of bl adesnumber of panel s on t he separat ed wake fiel d point(col l ocat ion point jof panel i Reynol ds number vol umet ric fl ow rat esurface boundarybody surfacesmal l spherewake surfacefarfiel d surfaceFl ow speed at rot or farfiel d fl ow speed inviscid fl ow speed rel at ive fl ow vel ocit y cont rol vol ume x l ocat ion of t he separat ion pointx l ocat ion of t rail ing edge init ial val uessmal l(0.5%)pert urbat ion of z2 vert ical l ocat ion of wake panel ividoi:10.6342/NTU2018005941Int roduct ion1.1 Mot ivat ionWarming of t he cl imat e syst em is unequivocal(Figure 1.1.The at mosphere and ocean have warmed,t he amount s of snow and ice have diminished,and sea l evel has risen/Cont inued emission of greenhouse gases wil l cause furt her warming and l ong-l ast ing changes in al l component s of t he cl imat e syst em,increasing t he l ikel ihood of severe,pervasive and irreversibl e impact s for peopl e and ecosyst ems.Limit ing cl imat e change woul d require subst ant ial and sust ained reduct ions in greenhouse gas emissions whicht oget her wit h adapt at ion,can l imit cl imat e change risks/These are t he key concl usions of t he Fift h Assessment Report(AR51 publ ished in November,2014,by t he Int ergovernment al Panel on Cl imat e Change(IPCC工 a scient ific body est abl ished t o provide an object ive source of informat ion about cl imat e change 86.1doi:10.6342/NTU2018005940)m e d E B J.Figure 1.1 Gl obal mean est imat es based on l and and ocean dat a 88However,despit e t he irrefut abl e proof t hat our pl anet is warming(sixt een of t he sevent een warmest years on record have occurred since 2000 88,and t he overwhel ming scient ific consensus t hat t his warming is t he direct resul t of human act ivit ies(in a review of 11 944 peer-reviewed scient ific papers,Cook,et al.13 found a 97%consensus t hat humans are causing gl obal warming,a significant port ion of t he human popul at ion st il l do not bel ieve in ant hropogenic gl obal warming,preferring t o t rust in propaganda provided by t he mul t i-t ril l ion dol l ar fossil fuel indust ry,which,over t he l ast hal f cent ury,has invest ed heavil y in an el aborat e disinformat ion campaign t o promot e publ ic skept icism of environment al science and environment al regul at ions,whil e spending,over t he past t wo decades,more t han US$2 bil l ion 90 l obbying t o affect US energy pol icy and weaken regul at ory agencies.But put t ing aside t he pol it ical l y cont roversial t opic of man-made gl obal warming,t here are st il l many ot her probl ems associat ed wit h our dependence on fossil fuel s,such as:Air pollution In addit ion t o CO2,ot her by-product s of t he coal/pet rol eum burning process incl udes:part icul at e mat t er,bot h a heal t h hazard(l inked t o heart disease,l ung cancer,chronic bronchit is,aggravat ed ast hma,and so on and a t hreat t o t he environment;sul phur dioxide and nit rogen oxides,which are responsibl e for,2doi:10.6342/NTU201800594amongst ot her t hings,t he phenomenon of acid rain”;and mercury,hich is，承匕especial l y harmful t o aquat ic eco-syst ems.Furt hermore,t housandsradioact ive t race el ement s,such as uranium and t horium,are rel eased gl obal y eachyear by t he burning of coal 91.Habitat destruction-The ext ract ion of coal is usual l y achieved by ext remel y dest ruct ive mining t echniques,such as st rip-mining and mount ain t op removal,whil e t he dril l ing and shipping of crude oil are highl y prone t o l eaks and spil l s,which can devast at e ecosyst ems.On t op of t hat,rising sea-wat er t emperat ures are causing mass coral die-offs around t he worl d,a probl em which is furt her exacerbat ed by t he acidificat ion of our oceans by rising CO2 l evel s.(And I wont even ment ion t he ot her rrcont roversialH inst ances of gl obal warming-rel at ed habit at dest ruct ion.War-Most of t he recent t urmoil in t he Middl e East can be t raced back t o of Brit ish and American int ervent ion during t he l ast cent ury,such as t he 1953 Iranian coup d*et atwhere t he Brit ain,t oget her wit h t he American Cent ral Int el l igence Agency,overt hrew t he democrat ical l y el ect ed prime minist er of Iran aft er he nat ional ised t he Brit ish Angl o-Iranian Oil Company,now known as BP.Simil ar power pl ays have t riggered viol ent confl ict s al l over t he worl d.A recent art icl e by Kl are 34 focuses on four areas,namel y Iraq/Syria,Nigeria/Sout h Sudan,t he Crimea/Ukraine,and t he Sout h China Sea,t o discuss how t he desire t o cont rol val uabl e oil and gas asset s is fuel l ing l ong-st anding hist oric t ensions.It shoul d go wit hout saying t hat t he human race urgent l y needs t o kick it s dirt y fossilfuel habit.This,in my opinion,is best achieved by t he promot ion,devel opment,and impl ement at ion of renewabl e energy sources.This is my personal mot ivat ion.3doi:10.6342/NTU2018005941.2 The evol ut ion of wind power髀Humans have been harnessing t he power of t he wind for t housandsjyears,by at t aching sail s t o wat ercraft,whil e t he first cl ear descript ion of a so-cal t ea t urbine,i.e.a rot at ing machine t o ext ract energy from a fl uid fl ow,is from t he first-cent ury BCRoman archit ect Vit ruvius,who t el l s of a combinat ion of t he separat e Greek invent ions of t he t oot hed gear and t he wat er wheel int o one effect ive mechanical syst em for harnessing wat er power for t he purposes of grinding grain and pumping wat er 92.This cross-fl ow design was l at er adopt ed by t he Persians for t heir panemones(essent ial l y,vert ical axis wind t urbines driven by drag forces.In t he 12t h cent ury AD,axial-fl ow windmil l s(Figure 1.2,RHS),t he fore-runners of t he modern horizont al-axis wind t urbine(HAWT工 began t o appear across Europe.Like t heir predecessors,t hese earl y windmil l s were al so drag based,del ivering high t orque at l ow speeds(bl ade t ip vel ocit ies of drag based t urbines are l imit ed t o t hat of t he ambient air fl ow,which was perfect l y suit ed t o t he t asks of wat er pumping and grain grinding,but is far l ess suit abl e for el ect ricit y generat ion,for which high rot at ional speeds and l ow t orque are more effect ive.Figure 1.2 Typical wat ermil l(LHS and windmil l(RHS)4doi:10.6342/NTU201800594The next major breakt hrough in“windmil l rot or design didnt come about unt il.t he t urn of t he 20t h cent ury.Inspired by Rankines moment um t heory 1863 andrpoude bl ade el ement t heory(1878),t he Wright brot hers real ised t hat an aeronaut ical propel l er is essent ial l y a wing rot at ing about a horizont al axis,where t he rel at ive angl e is 燮.唠二讲 different for al l point s al ong t he l engt h of t he rot or(Figure 1.3.They consequent l y int roduced a l inear t wist dist ribut ion t o t heir propel l er design.Impressivel y,t heir original propel l ers,pioneered more t han a cent ury ago,were onl y about 5%l ess efficient t han t heir modern equival ent 2.Figure 1.3 Il l ust rat ion of how t he rel at ive infl ow /rei varies in direct ion and magnit udeal ong t he l engt h of a wind t urbine bl adeThe design of t wist ed bl ades was furt her improved by Gl auert 23,whose bl ade el ement moment um(BEM)t heory broke t he bl ade down int o a number of independent el ement s,and is,wit h a few modificat ions,st il l used for t he design and anal ysis of modern wind t urbine rot ors.5doi:10.6342/NTU201800594For most of t he 20t h cent ury,however,t here was l it t l e int erest in t he use of winjd power for generat ing el ect ricit y.It was onl y during t he oil crises of t he 1970s t hat t he idea as real l y t aken seriousl y,wit h int ensive R&D project s being impl ement ed by It jimerican Depart ment of Energy(DOE)and t he Nat ional Science Foundat ion(Denmark).z/)7 t 1Today,wind energy is a mul t i-bil l ion dol l ar indust ry,and is present in more t han 80 count ries,of which 28 count ries have more t han 1 GW inst al l ed,incl uding 17 in Europe;four in Asia-Pacific(China,India,Japan&Aust ral ia;t hree in Nort h America(Canada,Mexico,US);t hree in Lat in America(Brazil,Chil e and Uruguay;and one in Africa(Sout h Africa 84,whil e Taiwan,l ist ed by marine energy consul t ancy 4C Offshore as one