`
`SEEDS FOR SUCCESS IN CIVIL
`AIRCRAFT DESIGN IN TH£ NEXT
`TWO DECADES
`
`19 - 20 May 1976
`
`PROCEEDINGS
`
`. .
`THE ROY AL AERONAUTICAL SOCIETY
`4 HAMIL TON PLACE, LONDON W.1, ENGLAND
`
`GE-1005.001
`
`
`
`A VARIABLE PITCH FAN FOR AN ULTRA QUIET DEMONSTRATOR ENGINE
`
`by D G M Davies. Dowty Rotol Ltd
`D C Miller. Rolls-Royce (1971) Ltd
`
`Paper Presented at the Royal Aeronautical Society's Spring
`Convention on "Seeds for Success in Civil Aircraft Design
`in the Next Two Decades" held on 19 and 20 May 1976 .
`
`GE-1005.002
`
`
`
`A VARIABLE PITCH FAN FOR AN ULTRA QUIET DEMONSTRATOR
`ENGINE
`
`D. G.M . Davies
`D. C . Miller
`
`Dowty Rotol Ltd .
`
`Rolls-Royce (1971) Ltd.
`
`1.
`
`INTRODUCTION
`
`A little over five years ago aircraft designers on
`both sides of the Atlantic were engaged on a new breed of
`civil aircraft for operation close to city centres .
`For a
`variety of reasons, environmental pressures being far from
`the least, Downtown STOL, per se, did not prove to be just
`round the corner, despite some encouraging experiences for
`example in Canada with the twin Otter between Ottawa and
`Montreal. However the arguments for such a system still
`exist, (albeit so do some of the objections) but furth c r(cid:173)
`mo're there is also a need for a new quiet aircraft to ·
`replace the ageing types engaged on intercity routes, which
`are mainly turboprop powered. Such an aircraft must have
`a field performance equal to that of the turboprop, and be
`able to satisfy the environmentalists, and so before STOL
`in the 1990's we can reasonably expect Q . R.T . O.L . in the
`1980's .
`
`The engine requirement for a short field (short or
`reduced is only a question of degree) is a higher ratjo of
`take-off to cruise thrust than that associated with
`low
`present day fan engines . A quiet engine requires a
`hot jet velocity for low rear arc noisP, Mnrl a Jow fan Tip
`speed for forward arc noise . Both noise and thrust require(cid:173)
`ments may be achieved with cycles of high by-pass ratio and
`low fan pressure ratio . These cycles also have a
`low s . f . c.
`and so they are therefore highly suitable for engines for
`RTOL and STOL . For low pressure ratios below about 1 . 4:1,
`efficient operation over all of the speed range of the
`aircraft requires some variability, which in the case of a
`fixed pitch fan could be effected by a variable area
`exhaust nozzle.
`If however a variable pitch fan is empioyed
`to take advantage of the flexibility of control and other
`features, that it confers on aircraft operation, it can
`itself provide this adjustment for efficient operation and
`eliminates the need for a variable nozzle .
`
`2 .
`
`JUSTIFICATION OF VARIABLE PITCH
`
`The variable pitch fan about which this paper is
`concerned has many claims of merit. The principal one oi
`these is that it provides propulsion characteristics, that
`allow aircraft to operate safely and quietly within an ATC
`system, involving steep approach and climb out.
`In fact it
`is difficult to envisage an alternative system that fulfills
`the requirements so completely.
`
`GE-1005.003
`
`
`
`2
`
`The a c· ievement of near zero thrust levels is a
`requi remC'nt fo r 6 -7° appr oach0s and this must bC' ac;sociatNi
`wi th an abilit y
`t u mudula le t Ile thrus l rap.idly j 11 unkr to
`fly accurately down a prescri bed path . Further, to ensure
`saf e ty and t o enabl e landings in the worst visibility
`c ategori es , a v ery fas t r espons e to full take-off thrust is
`essential.
`
`Under these approach conditions, the V.P. fan , when
`driv en by a two spool engi ne having an I.P . compressor, can
`r es ult i n the achievement of near zero thrust with almost
`instantaneous modulation r esponse and restoration of full
`take - 0££ thrust in a n emergency in about 1 second . The low
`thrust is achieved by adopting fine pitch of the fan at
`high fan RPM, when the I.P. compressor acts as a
`11brake 11
`enable the use of stable H. P . spool speeds . Further the
`high speed of the Fan/I .P . compressor heavily supercharges
`the 11 . P. spool, l eading to the fast accel e ration capability
`for r estoration of fu ll thrust. Variable I.P . compressor
`bleed is used in conjunction with fan blade pitch angle to
`preserve the best I .P. compressor running conditions .
`
`t o
`
`There are very many other advantages of V.P., which
`collectively ar e very attractive, particularly when
`associated with the above main feature .
`
`These ar e :-
`a)
`Elimination of the n eed for variable fa n duct
`nozzles to compensate for the effect of forward
`speed on fan operating line .
`Fine t uning of engine running for optimum power
`and SFC a t different atmospheric temperature s .
`
`b)
`
`c)
`
`d)
`
`e )
`
`:f)
`
`g)
`
`Reverse thr ust without special reversers ;
`thoug h the a bility with V.P . to achieve reve rs e
`thrust down to zero forward speed should not be
`overlooked, or the value of that under-estimated .
`
`Rapid achievemen t of reve r se thrust for normal
`landing and aborted take-off .
`
`Acoustic advantages by selection of particular
`fan speed for any part thrust setting .
`
`••Fr ee" air for wing blowing under all part
`thrust con ditions .
`
`Others involve topics such as water injection,
`engine adjustment to increase TBO, ~abin con(cid:173)
`ditioning, anti-icing cross wind immunity etc.
`
`On single spool High Bypass e ngin es V. P . might be
`essential for engine starting ( t.~ much load on starter
`unless fan in fine pitch) and l o w speed running , and
`because idle thrusts are otherwise too high for aircraft
`ground handling . The Turbomeca Astafan is an example .
`
`Many of the above features .. are also very attractive
`for militar y a i r craft .
`
`GE-1005.004
`
`
`
`3
`
`High bypass r a tio cycles normally req uir e h igh l y l oaded
`L . P . turbines , but the use of a reduction gear enabl es t he
`low speed variabl e pitch fan to be driven wi th a hi g h speed
`s haft ; thus allowing the number of turbine stages t o be k ep t
`l ow . This together with the elimin ation of a separate
`r ever ser system has significant weight advan t a ges .
`
`Having briefly s ur veyed t h e applicabili t y and chara(cid:173)
`cteristics of the V. P . £an engine , we will now mo v e o n t o
`describe the V.P . £an of a demonstrator engine , c alled
`M45SD- 02 , its design philosophy both aerodynamic a nd
`mechanical, its construction, its performance ,
`i t s f unction(cid:173)
`ing, a nd some early ~ est results from the engine runnin g .
`The basic parameters of the demonstrator engine and
`i ts
`proposed fully rated production successor the M45 S- 11 ar e
`shown in the following table .
`
`3 . ENGINE DEFINITION
`
`The basis of the demonstrator engine is t he M45H- Ol,
`which is currently in service on the VFW . 614 feede r air(cid:173)
`l iner . Changes are minimal to incorporate a variable
`pitch geared fan, but includes the addition of a 4th stage
`to the L . P . turbine, to transfer more shaft powe r to drive
`the fan with the consequent ial lowering of ho t jet velocity,
`which is a requirement for the achirvcment of low noise .
`
`M45SD-02
`
`M45S-ll
`
`M45H- 0 1
`
`10027(44 •6) 14370(63 • 9 ) 7600 (33 • 8)
`2 · 85
`8 • 73
`9•8
`1 • 60
`1 • 31
`1 · 27
`1 • 50
`1 • 18
`1 • 20
`1377 (420 )
`900 (274)
`900 ( 274 )
`
`1 133 (345 )
`
`1495 ( 456)
`
`96
`
`2 • 38 :1
`0 • 3 1 (8 · 8)
`
`1 05 · 5
`(Un si l e nced )
`
`0 • 46 (13 •1)
`
`Thrust lb (KN)
`By-Pass Ratio
`Fan Outer P . R.
`Fa n Inner P . R.
`llot Jet Velocity
`f/s (m/s)
`Fan Tip Speed
`f / s
`(m/ s)
`Noise Predicted PNdB
`at 500 ft .
`(170 m)
`Reduction Gear Ratio 2 •38:1
`SFC SLS lb/hr/ lb
`0•32 ( 9 •1)
`(mg/N/sec)
`
`1 027 (313)
`
`95
`
`The precise choice of pressure ratio for techn ology
`demonstra tion fan design is not particul arly impor t ant .
`It
`is not proposed to justify the choice hcrP.in beca us0 it i s
`firmly believed that the overall 0xperiment is not sen s i t i ve
`to the design variables and that there will be an adequa t e
`read across from the M45SD- 02 to a comparatively large
`range of fan pressure ratios and therefore to a n y particula r
`requirement dictated by a Q . R. T . O . L . project. For this
`reason the aerodynamic considerations discussed below tend
`to be of a more qualitative nature .
`It is interestin g t o
`note that American Designers have aJso chosen a simi l ar
`design point fan pres~ure for their Q . C . S . E . E . p roject .
`The general arrangement of the M45SD-02 0nc::d nr is :. hown j 11
`Fig . 1.
`
`GE-1005.005
`
`
`
`4
`
`4 . OUTLINE nr. FAN BLADE DESIGN PHILOSOPHY
`
`4 . 1 Fan Design
`
`The V. P . fans proposed and constructed to date by
`Rolls-Royce and Dowty Rotol have all been based on an
`aerodynamic design philosophy which has followe d t h e
`state-of-the-art of fixed pitch fans in general . Any
`departures from this have been as a r es ult of the influences
`introduced by variability .
`
`Conventional £an design , as practised by most
`authorities, has featured usually :-
`
`Supersonic tip speeds enabling a direct drive to be
`adopted with a minimum number of L . P . turbine stages .
`
`The resulting fan design r e quir ed the adoption of
`s harp edge airfoil sections s u ch as biconvex together with
`high solidity tip sections to reduce s hock losses .
`
`A high flow per ann ulus a r e a
`t ogethe r with a low hub
`tip ratio allows a minimum f rontal ar ea to b e achieve d with
`low weight .
`
`A further objQctive of high stage pr es sure ratio
`requires a highly loaded hub design, achieved by very high
`solidity, together with a reduction in area across the sta ge
`to limit diffusion .
`
`These features result in an i nverse tap e red blade
`with a ratio of tip to hub chor d of approximately 1 • 4 and
`this leads to low flex ural and torsional fre que ncies .
`
`Any problems arising from thus such as flutter can be
`handled by the adoption of low aspect ratio blades , or
`alternatively designs incorporating snubber s may ha ve
`significant weight advantages .
`
`4 . 2 Effect of Variable Pitch
`
`By comparison with the conventional fan design, t he
`V. P . fan ha s several furthe r geometric constraints .
`
`In order to change pitc h a constant o u ter casing
`diameter has to be adopted and this cannot b e f ully compen(cid:173)
`sated with hub flare , resulting in an area ratio near
`unity . Hence there is a
`l arger drop in axial velocity
`across the stage than normal, with an inhere nt highe r
`diffusion, which limits the work capacity of the design .
`
`There are two possible modes of r e v erse thrust
`operation; via fine pitch o r via coars e pitch .
`Each mode
`of revers e operation has advantages a nd disadvantages and
`because of this, it was decided that the demonstrator design
`and test programme would enable the invest i gation of both
`modes . Operation in reverse thrust via :fine p itch follows
`
`GE-1005.006
`
`
`
`5
`
`the 1 .tural sequence of control during landing when the
`blades are in fine pitch , as is the c a s~ wi Lh prope ll<'r s ,
`but in r everse f l ow the blad es operate wi Lh neua tj v c e<lmuer
`at a high incidence and conseque nt risk of flutter . Via
`coarse pitch , the angular movement r e quir e d to c hange to
`reverse flow is greater, and during this mov ement the blades
`pass through stalled conditions with hi gh risk of flutt e r .
`However in this £low regime the blades operate with positive
`camber and the potential l e vel of reverse thrust is greater
`than reverse thrust via fine pitch . However it shou ld be
`noted that the blade twist is wro n g in both cases .
`
`The requirement of reve r se thrust via fine pitch
`implies that the blade solidity must be l ess than unity
`since the blades must pass each ot h er and this inevitably
`limits the level of diffusion attainable at the hub .
`In
`order to achieve a design pres sure ratio of say 1• 27 :1
`with a blade mean so lidity less than unity and an area
`ratio near unity , a transonic blade design was required with
`a tip solidi ty of •8 to maintain the s h ock syste m within the
`blade passage .
`
`Now a blade , with a hub solidity near 1 and a tip
`solidity of 0 •8 , has a high inverse taper ratio . If~ is
`hub- tip ratio, the n taper rati o can be exp r essed as: -
`
`= c/ s tip
`c/ s hub x /""-
`
`Tip Blad e Chord
`I Iub Blade Chord
`
`= • 8
`t-'
`Since the hub solidity is approximatel y half t ha t of
`a conventional fan, the inve r se taper is much hi g h er r es ult (cid:173)
`ing in very low flexural and torsio nal f r equencies , and
`introducing the risk of f lutt e r . Furthe r
`t h e r e quir ements
`of varjable pitch prohibit the adoption of snubbers .
`
`'
`Eval uation of the aerodynamic d esign of th e blade , and
`consideration of the s tructural r e quirements l ead to a f inal
`choice of hub tip ratio of 0 •49, a nd a blade tap e r ratio of
`1. 7.
`
`Despite the high hub tip rat io the adoption of a high
`design axial vel ocity enabled the flow p er frontal area of
`32 lb/sec/ sq . ft . (1 • 35 kg/sec/ sq . m) to be achi eved . With
`a tip Nach number of 1•18 , a biconvex blade profil e was
`adopted . This al so g ives a symmetrica l sect ion fo r rever s in <J
`via coar se pitch , when the blade trail ing e dge J ead s .
`
`The following l ists the main features of th e c h os0n
`design which emerged from the above : -
`
`GE-1005.007
`
`
`
`Pressure Ratio
`
`Tip SpC'C'CI
`
`flub Tip Ratio
`
`Pressure Rati o Tip
`/ \lf/U2 Tip
`
`Pressure Ratio Root
`2
`1\11/U Root
`
`Solidity Tip
`
`Solidity Root
`
`Inlet Axial Mach No .
`
`Mass Flow/ Unit Frontal Area
`
`6
`
`1 · 27 : 1
`
`1027 ft/ sec .
`
`( 3 1 3m/scc )
`
`0 • 4 9
`
`1 • 36 :1
`
`0 • 28
`
`1 • 1 8 :1
`
`0 • 57
`
`0 · 83
`
`0 • 98
`
`0 • 6 1
`
`32 lb/s ec/ sq . f t .
`(1• 35 kg/sec/ sq. m)
`
`The test results of a 0 •35 scal e mo d e l ar e discuss e d late r
`in the paper .
`
`4 , 3
`
`Special Structural and Mechanic al Conside ratio n s of
`V .P . Fan
`
`The main st ructural probl e m of any f an blade is the
`achievement of a satisfactvry level of blade f r e que ncy .
`Fan blades are excited by virtue of dis turbances , fix ed o r
`periodic , upst ream or downstream of the fan, or by v ibration s
`t ra nsmitted through the sha ft . The e xc itation fr e qu e ncy is
`u s ually a who le number product of fan r . p . m. and the r e quire (cid:173)
`ment is to ensure that , in the n o rmal operating range , the
`blade resonant frequency and the e x c itat i on frequenc y do not
`coincide .
`It is a l so important to ens ure that the natural
`tor sional frequency and the 2nd mo d e f lexural fr e que ncy a r e
`not coincident . Even if these r e quir eme n t s are satisfied ,
`the aerodynamic excitation can s t i ll gi ve rise to f lutte r,
`a nd so it is also necessary to e nsur e that the fr e que ncy
`parameter ( 2 TT x Frequency x Chord• incidence v e l ocity )
`exceeds an empirical limiting value .
`
`Blade frequency under centrif u gal c ondi t i o ns is
`larcly a function of the geometrical s hap e and the ma t eria l
`properties . Thus the s ame basic c onside rations had to be
`made for the V. P . fan , a s would normall y be made f or a F. P.
`fan .
`
`The structural design proble m however had to be s o lved
`wit hout resorting to mid s pan snubbe r s . This problem was
`accentuated by the h igher than usua l
`tape r ratio, dictated
`by the aerodynamic design conside rations, and in order to
`get a satisfactory solution , theo r e t ical c alcula tions
`indic~ t rd that a thickness c hord rati o a t 1hC' tip or o nl y
`
`GE-1005.008
`
`
`
`7
`
`1-!% '';as required . However there was concern that other
`problems might arise with such a thin blade , and so a single
`blade was manufactured for frequency tests, (a) to Vl'r.i.Iy
`the theoretical calculations, and (b} to determine whether
`any undesirable modes of vibration existed .
`
`The test results were extremely interesting , because
`they highlighted a sensitivity of torsional frequency to
`r.p.m. and this presented an easement with regard to IT and
`2F coincidence . As a result of this , it was possible to use
`a 2% tic ratio at the blade tip.
`
`5. STATOR AND OUTLET DUCT DESIGN
`
`The stator was designed conventionally but with two
`additional considerations .
`
`1) Noise
`
`2) Reverse Flow
`
`To reduce fan tone noise by use of cut off, the number
`of stators was chosen to be just over double that of the
`rotors, while the gap between the rotor and stator was chosen
`at twice the rotor chord .
`
`The requirement of reverse flow involved the choice of
`a biconvex profile , to reduce the trailing edge thickness in
`reverse flow and enable the stator, to act as an inlet guirle
`vane with minimum loss.
`
`An undivided stator is adopted with the swan neck
`inlet duct to the engine placed behind . This , togethe r with
`a relatively blunt splitter nose, enables the reverse flow
`to enter the core with minimum loss .
`
`In forward flow the fan duct exit nozzle has approx(cid:173)
`ima tely 70% of the fan £low area .
`In addition it h as the
`normal sharp outer lip . Thus in reverse flow, the effective
`noz zle area is inadequate, thereby limiting the attainable
`reverse thrust.
`In order to avoid this, auxiliary revers e
`flow intakes are positioned aft of the fan stator , to augment
`the reverse flow .
`
`6 .
`
`FAN BLADE RETENTION
`
`Having established the blade geometry and its speed of
`rotation, in the manner described above, the blade weight
`and the centrifugal force , it will generate , can be compu ted .
`The blade must now be restrained in the hub against this
`centrifugal force, and yet be free to rotate about its
`pitc h axis. To illustrate the magnitude of the problem , the
`M4 5 SD-02 V . P. fan diameter is 5 ft .
`(1 · 5 m) and rotates at
`4,000 r.p.m . The blade chord for 14 aluminium blades ranges
`from 6 in . {15 cm) at the hub to 10 in , {25 cm) at the tip .
`The weight, of each aerofoil part only , approaches 3 · 3 lb.
`(1 • 5kg) and, at the design speed , cx0rts 15 tons cc11tr.iiu1_Jal
`
`GE-1005.009
`
`
`
`8
`
`force . The shank of the b lade and other parts of the root re(cid:173)
`tention system r sults in the n eed to provide a total restraint
`of about 40tons . In a ddition, there is the aerodynamic force
`on the b) ade, which acts a lmost normally to the plan e of the
`blade, but which, due to the twist on the blade , it is more
`conveniently resolved into a torque forc e and thrust force .
`These ar0 relatively small compared with the centrifugal
`fOJ cc generated . Typical values are 350-400 lb .
`(160-180 kg).
`
`A great deal of effort was put into solving the blade
`retention problem on propelle rs . The solution developed ha s
`been applied to the v ari a b l e pitch fan, and a cross-section
`is shown in fig . 22 . The sche me uses a rolling element bear(cid:173)
`ing (lower bearing in sketch } to transmit the blade centri(cid:173)
`fugal force loads to the hub, while leaving the blade free t o
`pivot . The thread of the b lade bolt, which secures the
`bearing to the blade, must be prelo aded to minimise fatigue
`damage , and this is done by ti g htening it against a 2nd
`bearing (upper bearing in sketc h} located on a shoulder of
`the blade . This 2nd upper bearing also serves to oppose the
`bending moments, induced a t th e roo t of the blade, by the
`a erodynamic forces acting on the blade a e rofoil.
`
`In order to establish t he integrity o1 the blade root
`retention system, fatigue st r e n g th tes ts were carried out,
`before the fan was run on t he e ngine. Those tests were
`carried out by mounting the blade root fatigue specimen in
`a " pot", the top end of which was f abric ated to represent
`the hub .
`The lower end or t he s p e cime n was identical to
`the b l ade root base, and the upper e nd was formed into a
`thread shank . The specime n i s shown in fig . 3 .
`
`Fatigue testing was c arried out by applying the
`centrifu gal load , genera ted b y the aerofo il alone , to the
`upper portion of the specime n, with a
`t e nsion member .
`Centrifugal load which is n ormally gene rat e d by the blade
`hearing components, was app l ied, by me ans of a hydraulic
`jack in compression , under t h e b lade root . Superimpo s e d on
`to this centrifugal l oad wa s th e design bending load. The
`bending stress induced was the n varied by a repr e s e ntative
`osci llatory stress . Typical numbers for this test we re
`4 7,000 lbs (21 , 400 kg) tension at the upper e nd of the
`spec imen , 23 , 000 lbs (10 , 400 kg) compres sion under the blade,
`4 450 lbs-in . (5120 kg - cm) steady bending moment and +3 750
`l bs - in .
`(4310 kg - cm) vibratory bending mome nt . The first
`assembly was tested to 30,000 ,000 reve r s als a nd then the
`vibratory bending moment i ncreased to !4200 lb-in .
`(4830
`kg - cm ) and a further 30 , 000 ,000 reversals we re applie d .
`
`Wh e n further e n g ine running has been carri e d out and
`a f ull appr a isal of t h e vibratory stres s es made, a 2nd
`specimen wi l l be tested with these measured vibratory loads
`applied.
`
`GE-1005.010
`
`
`
`9
`
`7 . Hf'q STRUCTURE
`
`The design problem associated with a V.P. fan hub, which
`has to restrain 14 blades , each exerting about 40 tons
`radial l oad, is not an C'asy one . Tht' solution was found
`in a hub, comprising 2 compJl'l e rings of hi<Jh hoop
`strength , carrying on the out er periphery a numbc•r of
`"pots 11 , one for each blade. The transition , between each
`blade 11pot" and the main structural rings , proved to be the
`most difficult structural design problem.
`In order to
`verify our calculations a full photo-elastic test was
`conducted. The work was carried out by Stress Engineering
`Ser~ices Ltd. who constructed a full size model of the hub
`in Aralaite CT200 material . The complete model was rotated
`in an oven, and the temperatu re increased to above the
`critical value, for the material. The temperature was then
`slowly reduced, whilst the hub was spinning , to freeze in
`the stresses . Blade centrifugal and bending loads we r e
`simulated by applying dummy weights, which screwed into
`each blade port . On removal from the oven , the Araldite
`hub was then sectioned and examined under polarised l i9ht .
`A typical segment of the model is shown in fig . 4 , and some
`typical photo-elastic fringe patterns are shown on fig . s .
`The results were used to modify the final design in order to
`reduce the stresses in the critical areas .
`8 . PITCH CHANGE MECHANISM
`
`Just as the centrifugal blade load is the main direct
`force to be opposed, so the main pitch change moment to
`overcome is generated by the centrifugal fie ld i n which the
`blade operates . This result s from the blade mass being
`displaced on each side of the pitch change axis , and the
`blade lying obliquely across the plane of rotation . Each
`element of mass , away from the pitch change axis , tends to
`take the blade towards fine pitch (fig . 6) . This i s
`termed
`centrifugal twisting mome n t
`(C . T . M.) and increases rapidly
`with chord . The forces , which generate this , also have
`components along the blade, which tend to straighten o ut the
`twist , which is built in for aerodynamic reasons . Howe ver ,
`this lat~er moment manifests itself as a stress within the
`blade , whereas the fo rmer has an overall resultant, which
`must be catered for by the pitch change mechanism .
`
`In addition to the C . T . M. there is a resultant moment
`due to the aerodynamic force acting through the ce n tr~ of
`pressure, which seldom coincides with the section centre of
`gravity, about which the blade is "stacked" . The magnitude
`of the aerodynamic twisting moment (A.T . M. ) varies wi t h the
`design and blade loading. For most forward thrust
`conditions the A.T . M. acts to oppose C . T . M. but in the event
`of loss of power , the fan starts to windmill, the sense of
`the A.T.M. changes and it becomes additive to C . T.M . These
`then , are the moments , the algebraic sum of which mu st
`always be opposed in order to hold the blade in any one
`position. When a blade pitch chang0. is attempted , friction
`at the blade root must also be overcome .
`
`GE-1005.011
`
`
`
`10
`
`On the M~ssD- 02 fan C . T . M. per blade is of the order of
`3000 lb-jn (34'>0 kg-cm) . The A. T . M. is ahout 700 lh-jn
`(800 kg-cm) at altitude, in the opposite sen~c to C . T . M.,
`and the friction moment , which of course alway~ opposes
`movement, is about 800 lb-in (920 kg-cm). To coarsen the
`blade at cruise conditions it is therefore necrssary to apply
`a moment of 3100 lb- in (3750 kg- cm ) to each blade root .
`This gives a total torque requir ement, taking into account
`gear ratios between blades and mechanisation, o:f approximately
`5000 , 000 lb- in (575 , 000 kg-cm) .
`
`The motive powc>r for applying th<' pi 1 ch chnn~.J<.' torque
`to each blade is hydraulic . Trad itionally Dowty Rotol
`propellers have employed a piston and cylinder arrangement,
`with connecting rods to offset pins in the base of each
`blade . This arrangement was r ejected for V.P . fans , because
`of the large number of blades and the high "G" :fore es . The
`chosen sol ution was a small bevel pinnion splined o~ to the
`end o:f each blade and meshing with two large bevel ring
`gears . Rotation of these ring gears is by means of a
`rotary vane type actuator . The ring gears are mounted , one
`to th e actuator inner rotor , the other to the casing , so
`thC\t ignoring friction a pure couple is applied to each
`blade . This removes a ll gear side loads from the main
`l>l."\tk• bl•aring . The concep t was first tri e d on the V. P . fan
`fpasibi Ji Ly demonstrator , fitted l o C\n /\.s ta~uu c•nginc ,
`details of which have been published elsewh0rc . The mode of
`operation is shown in fig . 7 . The basic design problem ,
`associated with the actuator , was one of sealing , to
`minimise leakage between fine and coarse pitch chambers .
`This was achieved by using a combination of synthetic rubber
`and P . T . F . E . around the periphery of each of the vanes .
`In
`order to demonstrate the satisfactory functioning of the
`pitch change mechanism prior to engine testing , a vane
`motor was mounted inside a hub, :fitted with dummy weights to
`represent the blade loads and subjected to a series of tests
`over a range of r otational speeds and operating pressures .
`This proved to be a relatively routine development job and
`it is not proposed to amplify the details further .
`
`9 .
`
`P ITCI I CONTROL
`
`Traditionally the control of a V. P . propeller has been
`by means of a C . S . U . mounted on the cng inr wheel case> , and
`supplying oi l to the propeller via a rotating transfer muff .
`This ensures that th e power developed by the engine is
`absorbed at a pre-determined propeller r . p . m. a1 all times .
`Whil st the propeller control system functions satisfac t ori ly
`with respect to per:formance, the need to open up the oil
`system each time a propeller is removed from the engine, has
`given rise to the proposal of a self-contained unit . Thus
`the ground rules for the control system for the M45SD-02 fan
`were that the unit should be capable of being removed from
`the cn<Jine without disturbing the oil system , and tha t any
`structural failure in the region o:f the rotating gear should
`n ot result in a control system failure.
`
`GE-1005.012
`
`
`
`11
`
`7 he self-conta ined req uirement was met by making the
`hub driving centre into the main oil tank, and mountino ther e (cid:173)
`in the high pressure oil supply required by the w1.ne motor
`(fig . 8) . This comprise d
`two pumps driv<.>n from ;1. sta tj on;1 ry
`member , within the fan , and which is s pigotted an<l dow( • ll1 ~d
`to the engine gearbox casing . Modulation of the pressure oil
`to the vane motor is by means of a conventional fo llow- up
`servo valve . This is disposed radially within the fan hub,
`and is a directly operated control . The transfer of the
`valve input signal, from the static to the rotating m0mbe r,
`is via a plain thrust bearing . This is the essence o.f th"
`control system , but also mounted in the hub driving cen t re/
`oil tank are relief valves to regulat e press ur e , a change (cid:173)
`over valve to bring the 2nd pump into the circuit for hi g h
`rate pitch change filters, and drillings to provide a source
`of lubrication to the pump drive gears. This hydraulic
`circuit was created by mounting the components into an
`aluminium rin g , which was drilled to connect the various
`compone~ts in the prescribed manner .
`It has not bee n poss (cid:173)
`ible to test this system separately from· the other components ,
`but a complete fan test has been carrie d out to check the
`correct functioning of all the hydraulics .
`
`The engine/fan control system is inte nded to b0 of th~
`single lever type, but for the ini tia 1 d e mons tr a tor r unninu
`independ e nt control of fuel, fan pitch and L . P. pressu r e
`bleed is being used . The engine performance will b e mapp e d
`using this system, and later a black box system will be u sed
`to control the engine along the desired working line .
`The
`complete V.P . fan assembly is shown in fig . 9 .
`
`10 . RIG TESTING OF MODEL FAN
`
`Tes ting of a 0•35 scale model fan has b een c onducted
`using a rig driven by a Gnome gas turbine . Thi s
`i s s h own jn
`fig . 10 . The free power turbine drives the mod e l fan from
`the rear, via a gearbox and torque meter . The mass flow may
`be altered by varying the width of the flared outlet duc t .
`Intake flow is measured by means of a standard b e llmouth,
`with static pressure tappings . Other measurements we r e exit
`total and static pressures, exit total t e mperatur e , shaft
`horsepower and speed.
`
`the>
`The blade pitch was adjusted betwee n test runs t o
`required value . Test runs were carried oul , with a 1ixcd
`pitch, at a con s tant outlet throttle setting . Me a s ur e me nt s
`were taken at varying s p eed s , and with a succession of pitch
`s ettings in forward flow .
`For a given pitch setting, a
`compressor characteristic was obtained by cross plotting t he
`measurement s from the constant throttle runs . The sur ge
`behaviour was very gentle , and so a surge point was clr•fined
`whe n the outlet static pressure attaine d a maximum value .
`The design pitch characteristic at varying speeds is shown
`in fig . 11 and the effect of varying pitc h is s hown in
`fig . 12 .
`
`GE-1005.013
`
`
`
`12
`
`It will b< seen from fig . 12 that the pressure ratio
`cfpclinC's som0what i'\S the b lades arP set finrr, but at zrro
`tip pi tell tllt'H' is s t i 11 a ~ i uni1ica11l p!<' ss un' ratio
`th'veJopPd, d0spi te the sma ll f l o w. Fig . 11 shows that an
`op0ratin9 pressure ratio of 1•3 can be s ustained with the
`blading, in excess of the design pressure ratio 0£ 1 · 27 .
`order to determine the reason fo r this, traverses were
`carried out with probes to meas ur e the radial variation of
`total and static pressure and total tempera tu re , behind the
`rotor and stator .
`It was concluded that there was a signifi(cid:173)
`cant untwist of the carbon fibre rotor blades at desi gn
`spe0d , thereby further increasing the tip wo rk. A further
`mod0l was made of the engine blade , in which a pretwist was
`incorpora t0d, and the traverse o:f the f low .from this blade
`was much nearer to the design value . As a res ul t of this
`t(·~ t, i l was decid0d to i ncorporate a pretwist into the full
`si2P engine aluminium blade .
`
`In
`
`Reverse flow testing has also b een carried out by
`setting the outlet ducting open to its gr eates t e xtent . The
`intake bellmouth was removed, and a cowl substituted to
`~imulatc the pod intake of the full siz e engine . Revers e
`facin0 rakes were fitted in front of the rotor to mea s ure
`total pressure . Tests were carried out at various blade
`srttings in reverse flow , via fi n e pitch and via coarse
`pitch . The tests demons trated that t h e tip section generat ed
`most of the reverse press ure ratio .
`In addition the reverse
`pitch via coarse pitch gen e rated a high e r rev e rs e pr es sure
`ratio than via fine pitch .
`
`Noise measurements were al so taken statically at
`various pitch settings with bellmouth and pod intakes .
`These indicated that a minimum noise occurred at an inter(cid:173)
`mediate pitch setting , consistent with a moderate aerodynamic
`loading . Further measurements were take n behind a
`low speed
`fan driven by a Proteus engine to simulate forward air speed .
`These tests indi cated that the pod intake had a 5 dB
`reduction in noise at forward speed compared with the s tatic
`tests and differed in its noi se behaviour from the bellmouth
`intake at the design tip speed .
`
`11. DEMONSTRATOR ENGINE
`
`The core of the M45-H engine has been used as the basis
`of the demonstrator e n gine fig . 13 . The r e quir e ments of the
`demonstra tor engin