Comparison of Flight Measured Helicopter Rotor Blade Chordwise Pressure Distributions and Two-Dimensional Airfoil Characteristics

COf.ff..PARISON OF FLIGHT MEASURED HE:LICOPTER ROTOR BLADE CHORDWISE PRESS{JRE DIS1rRIBUTIONS AND TWO-DIMENSIONAL AIRFOIL CRft.RACTERISTICS By James Scheiman and Henry L. Kelley NASA Langley Research Center For Presentation at the CAL-TRECOM Symposi~~ on Dynamic Loads Problems Associated With Heli copters and V/STOL Aircraft, Buffalo, New York, ?'

• I

q ~

[

(

" '1 COMPARISON OF FLIGHT MEASURED HELICOPTER ROTOR BLADE CHOB~WISE PRESSURE DISTRIBUTIONS

I~

AND TWO~D~lSIONAL AIR]UIL CHARACTERISTICS

(

By J sme s Scheiman and Henry L. Kelley NASA Langley Research Center

I'

~ SuMMARY

•

a

A comparison is made between air~011 chord1r.ise pressure distribu~ tiona from helicopter rotor flight t~;3ts and static two-dimensional

I

wind .. tunnel tests. Differences in actual and tvro-dimensional ... airfoil pressure distributions are shown to exist. These differences in air- - '

I

foil characteristics are expected to amplify the blade flapw-:i.se and.

I

torsional vIbratory forces determined from two .. dimensional ... airf'oil data. Possible reasons for thr£!se airfoil differences are briefly dis-

~

cussed.. The point is made that in endeavoring to coni'j.rm current refined theories of calcu.lati.ng section angle of attack" it is essential,

I

in making data. comparisons" that care be used to prevent these differences

I

..

-- between actual and static two-dimensional-section (tata from obscuring

!W' the effectiveness of the anglr:;-of··e,ttack ca.lculations.

I

If I, Experience has shown that the ability to perform an adequate

I

....

structural dynamic analysis of the ;r'otor blade is marginal. This lack of ability has generally been viewed as attributable to unknown air

I

loads and in particular to l.lnknown inflow velocities rather than to t ' the applicability of two ... dimensional-airfoil characteristics. This

I

it view has tended to be confirmed, for example, by the results of rotor test inflow veloai ty measurements and by the adequacy of pred.:J.cting

I

I '

(

!

~,,,,,,~,-

- 2 ...

helicopter performance by use of two-dimensional airfoil data. In any Case, both :i.!'i~low -velocities a-nd actual airfoil characteristics must be knmID in order to perform a reasonably accura:te dynamic blade analysis.

In regard to the inflow velocities, it iN believed that the capability of theo:t"j' to pred.ict these velocities for trim level flight has significantly improved recently. With these new theories the danger exists for blaming the remaining inadequacy of the inflow theory for any lack of correlatlon between test and. theory, when the differences may actually be caused by airf?il characteri.stic discrepancies.

The validity of two-dimensional de.ta has been gi ~ven little detailed attention because of a Jack of actual. operatin£,' test d.ata. Partly to help fill this gap, the NASA Langley Research Center has recently completed, a helicopter flight-test program which has utilized extensi va blade pressure instnwentation. These data.. provide a comparison of the actual and wo-dimensional-airfoil chordwise pressure distributions to the extent needed to illus'crate that importarlt airfoil characteristic discrepanCies do occur in the flight conditions sampled~ Portions of the flight measured chordT,.;rJ,.se pressure dist:;:-;tbutions for two f'light conditions are discussed. Samples of these distri'bu ...

tions are direc'bly compro.'ed 'With two-dimensional full- scale data (see ref. 1) by equ.ating the two normal force coefficients. The chordwise pressure distribution for other flight condi'ciona and. the movement of the blade center of pressure are discussed.

,. 3 - .... ,'.

All reference to two-dimension@..l-airfo',ll characteristics in this paperref'ers ,to static two·~dimensioni1.l charaoterist:i.cs in distinction "~:' ' to (,ls<!illating 'U..."lsteady two··dimension~u charai.~teristics.

I

SYMBOLS

I

c airfoil chord eN norrnal;o.force coefficient

I

~p differential pressure measured on the air:i\')iJ ' .

" q dynamic pressure

I

:r radia,ldistance to blade elemen'~' measured from center of ;0'~

I

rotation I' ..

R blade radius measured from center of rotation

I

J;,'~:' VF forward speed x chordwise distance measured from blade leading edge

I

" X center of pressure of, airfoil section measul'ed from leading ~,-- ...

I

edge nondimensional tip-speed ratio, ·7v

I

blade nominal azi!i1uthangle"measured from downwind position i-': I,:.'

in the direction of rota'bion and disregarding blade lag

I

.', motion ,:;'

I rotor angular velocity

I

~~

I

<, II ,.::P;

I

~:"~l

I ~-, ~~ ... , ~ , ,,~ : 1

.. -4 ..

DISCUSSION Normal-Force Coefficient for Flight With Blade Section Stall A plot of the local normal-force coefficients CN along the blade radius for different azimuth positions is shown in figure 1. The flight condition is for a trim, level-f'light.;;cruise forward speed and a reduced rotor rotational speed and is thus for a flight condition expected to produce local blade-section stalling. Further details of' this flight condition. are available i~ table. IV of reference 2.

Notice the high values of normal""force coefficient in the area of lJ.1J:J.ese coefficients correspond to dynamic pressures of approximately

50 pounds per square -foot at r/R = 0.55 and 100 pounds per square

foot at r/R ~ 0.75. The normal-force coefficient values; in these areas of the rotor, Correspond to values above the maximum st.atic two-dimensional values of CN- Just prior to these high normal ...

force coefficients a rs:pj.d rate of change in the normal-force coefficient is noted. This change in C1\1 can be directly related.

to a two-dimensional .. airfoiJ. angle-ot-attack change and the corresponding . .

high rates of angle-of-att~lck cha.nge can be explained, for example, by· , , ' > the rapid changes in local inflowvelocitiea throUgh the rotor. In ; ."

this instance ,an estimate based on su.ccessi ve ilormaJ. ... force coeffiCients, in the previously mentioned high angle-or-attack area of' the rotor, indicates a rate of' roughly 100 per second or 1 per 2-1/2 blade- chord lengths. This rapid angle-of~attack increase will.

I

~:

I

- 5- .. :,:~{r' .

I

provide, a parti,al explanation fOr the lack of chord'Wise-pressure- ~--~- 4istrihutioh correlation explored :f'Llrther in this paper.

I

The circles on this figure are points where chordwise pressure -

,I

: , distributions are discussed in figures 2 to 4; the solid' circJ"e~ are ~ points where two-dimensional and flight data 'are compared.

~ :~.:~

I

) 1"',"" Chordwise Pressure Distributions for Flight . .:-

I

Wi th Blade Se,ction Stall '.', ' " Figures 2to 4, are for the same flight condition as figure 1, ' '

I

which was selected with the expectation of producing local blade- ;'-'" ,", section stalling. A plot of the chordwise pressure coeff:lcient dist:ri- <-

I

" :.q.

0 0 bution for r/R ~O.55 and'ii :=. 165°, 195°, 210°,225 , and 255 is ~ shown in figure 2. The blade ... azimuth position, i,htegrated norma.l-force

I

" coeffiCient, 'aild the centers of pressure are as indicated. At 'it = 165°, ',-

I

the flight-test distri'bution agrees with the two-dimensional data; the center of pressure is close to (slightly aft of) the gMarter chord.

I

The normal-force coeffioient is below the two~dimensional stall pOint.

At 1jr = 195 the norml:ll-force coefficilent of 1.3 is above the two-

., I

---"\1 dimensiona.l stall value but the pressu::c'e distribution appears unstalled.

I

At the remaining azimuth locations the normal-force coefficient is above the airfoil section two-dimensional stall point and no two-dimensional

I

da'ca are ~vailable for comparison. The pressure distribution is such I;l.S to correspond to some separation and, theref'ore,the section can be

I

irl.ewed as exhibiting stall characteristics although the details of the -,i,

I

distribution ha'V'e no counterpart in tivl'Q ... dimensionru. data. Based on examination of the contours of figure. 1, this increased maximum ON

I

(

,;t,~

.. 6 - is expected to increase the vibratory amplitude of the actual airfoil blade loads as. compared with the predicted two-dimensional air loads.

It may be of interest that some variations in pressure occurred between rotor revolutions for these last three plots (average values . are shown); however, the distribution shapes are belie·vBd .. representati ve ..

These variations in themselves suggest a stalled type of flow.

Sample pressure distributions for TIR = 0.75 are. shown in

figure 3; these distributions are for the same flight condition shown

j.nfigures 1 and 2. For '" = 165 and 195 the distribution shows

good agreement with two-dimensional data" As the azimuth angle increases 0 0 from 1Jr = 195 to 240 , the normal-force coefficients again increase to values above the two-dimensl,onal stall point, although the actual air- foil reta.ins the unstalledtwo-dimensional pressure distri'.jution.

The pressUre distr:i .. bution for r/R = 0.95 is shown in figure 4.

The normal-force coefficients are all be,low the static two-dimensj.onal stall point rElnd ·therefore good pressure ... ,d1str1bution correlation would 0 0

be expected. For 0/ = 45 and 75 the agreement between the flight and

two-dimensiona.l data is indeed reasonable, but at 1jr = 90 and 120 the

correlation is not so good. Thus, while a large part of' the rotor does behave in accordance with two-dimensional data, figures 2 to ~. show that poor correlation can occur to a degree which would be eXpected to have a major. effect on ~eriodic blade loads.

,

"

J

;;1

.,

,I"

- 7 ...

:1 'j

Chordwise Pressure Distributions for Flight '\::' With High Blade~Tip Mach Numbers .II ",--.-'

J

The next trim. level flight condition, discussed for flight at a high tip Mach number, for which th~ maximum blade-tip Mach number " .

I

was 0.76. The chordwise pressure distribution for the 0.95 blade

;1

station is shown in figure 5.' The norma1. ... force .coefficients are all below the two-d'; i;~xl~ional-airfoil stall point; and good correlation

t

would be expected. The centers of pressure, however, are all farther '.

forward than would be expected, even for high Ma,ch number operation of'

I

a'two-dimensional. airfoj.l. F~r W, = 30 there is reasonable agreemen.t ,,' :1 ·,-.,L-,.: between the flight andtwo~di~~nsion.al distributions, althOUGh for ",

* = 75°, 90°, and 105° the flight data depart from the two-dimensional

J

data.

The 0.75 ... radius station shown in figure 6 is for the previou.Sly

I

described- high tip Mach number flight. TIle fl:Lght-tneasured chordwise ,.1: J ' pressure a.istributions, the centers of pressure, and tlQrmal-force

I

coefficients are typical of unstalled two-dimensional data. The -'

I

correlat;i.on shown is good.

') The high tip Mach number test for the O.55-blade-span statlon is

I

shown in figure 7. Again the normal-force coefficients, centers of

pressure, and the distribution. are typicaJ. of two-dimensioneJ. dc{..ta.'

I

'F' The correlation v;ri th two-dimensional d.ata is again good.

I

In sUJ'rmlary, figures 5, 6, and 7 for the high tip Mach number flight indicate that a large percentage of the actual chordwise pressure distri ...

I

butions are in agreement wi t,h two .. dj.mensional airfoil data. Only a

I

I

.. -. ......... "

" ..

::-.; '," small (though important) percentage of the pressure distributions are not in agreement. Since the disagreement in this case is primarily in the high Metch num.ber regions, it appears that with care:ful selection of the flight test conditions, it will be possible to find cases that· warrant comparison with theories using the two-dimensional data to study the adequacy of new angle ... of-attack prediction theories.

Other Flight Conditions A small portion of the chordwise pressure distributions for a number of other trim level-flight conditions. have be.en revievled ar'ld the results were similar to theresul ts of the prevj.ou8 two flight conditions discussed in detail in this paper; namely, that portions of the actual operating helicopter blade do not behave in accordance with two=d.imensional airfoil data. Because there are these cases where important differences do arise" an exact knowledge of the rotor inflow velocities is not necessarily SUfficient to describe the exact rotor blade loading. Caution should therefore be exercised in interpreting the correlation of flight measured and theoretical rot~);;, .... blade span- wise loadings.

Measured Center-of~Pressure Movement .'t· In an atte:n:q>t to generalize the /3,ctual airfoil center;"of-pressure mo'rehlent, a plot was made of the center of pressure as a function of the blade azil)IUth angle :foX' three different flight conditions, and this plot :Ls shown in figure 8 I' Note t,he forward shift in center of pressure on the advancing side of the rotor for all three night

[

(

.. 9-

[

conditions and the rearward center of pressure on the retreating side of the rotor for the first flight condition (flight with blade section

[

stall dl.scussed Tlreviously). The forward shift in center of preasuxe could not be explained by Mach number effects.

(

; ~. ' The flight-test blade was a modified NACA 0012 airfbil(ref. 1) . ,

I

and hence had no camber. The interesting possibility thus arises that if a. small amount' of camber, which would tend to add a constant moment

(

coefficient with varying angle of attack below stall were added, the variB.tions in dynamic pressure with azimuth would then modifY the ' i

I

m.easurea centers of pressure in such a we:y ,as to result in reduced

I one-per-revolutiol'l aerodYnamic control forces. In other w()rds, the

added' source or moment variation 'With azimuth would be expected to have

I

a phase angle such as to offset partially the measured ~rlations.

I

Discussion of Actual and Two-Dimensional-Airfoil Pressure-Distribution Differences

I

The reasons for the differences'found between actual and two- dimensional airfoil data aJ."e not com;ple.tely ·understood. As is well

I

known, the flow conditions on a rotor are highly complex and many

I

potential contributing explanations have long been at hand should such problema arise. Since the problem haa now been verified in

I

tangible form" an effort is being made to sort ou'b su;m.e of these pOssibilities.

I

As one exc:urrple, the f~!Lct that a high rate of increeLse in angle

I

of attack can give higher than static CN values is '\;Tell l"JlOwr.l max ; (for example, ref. 3), and this effect has long been looked for in

I

N ,

I

~ "

'.

" - 10 - rotor measurements. In some early in'\"estigations this effect was apparently negligible, that is, stall was evidenced roug:bJ.y "ihere expected. Apparently the other complexities of rotor inflow, and the specific design details, prevented any sii!snificant occurrence of bigher than static CWmax vall1es. In severa]. more recent irrvestigat1ons, including the present one, the 0r:;;08i te has been true. Dynamic!9lly, this effect would be expected to increase the actu.al amplitude of the oscillating" air loads as compared to the calculated loads based on twX' -dimensional data.

It should be noted that the most drastic source for high ra.tes of change of angle of attack is likely to be the striking of the tip 'VOrtex from the previous blade. Consequently, the high rates Of change ar.ld the C values in excess of static tivo ... dimensional values may N occur in specific cases in basically mild flight conditions as well as in the low rotor speed or high forward velocity conditions normally associated 'With blade-sect:l.on stalling.

Time-varying blade yaw angles, spanwise flow on the blade, and nonuniform velocity gradients in front of the airfoil are other possible factors that may ca.use disagreement between actual and two-

.,'

dimensional airfoil characteristics.

CONCLUDiCNG REMARKS It has been shown that the actual helicopter rotor blade does not always belLC'J,'\"e 111- accordance wi'uh two ... dimensional airfoil data. These air:foil"'characteristic differences are expected to amplify both the

, ff~

L

.=-~

lr

It

- 11 - l~ ...

[

fr: flapwiseand torsional blade oscillating loads. Fossible reasons for

tL

these differences were briefly explored .. The point is made that when ..

(f" -~!; the loads are predicted from :refined inflDw theo:,ries as compared m'th

rL

experimental loadings, caution ~hould be exerci~ea in interpreting ~.~ ~l .tt.

.~ dif'ferencesin blade loading,since these may arise because of the ~ "lack of applicability of two-dimensional data rather than inadequacies

II

in the inflow theory. Thus, before comparisons of actual and predicted f'~, , Ii .--"j air loads are used to" determine validity of angle-ot-attack calcula~ ~, tions, each experimental case used must be revieweq, for evidence of

[~

the presence or abs n~1e of discrepancies betw"een the actual section aerodynamic characteristics as reflected by chordwise pressure distri- ~.

,,1- bution; and the section charac'teristics being assumed in 'the analysis.

ft-c =:~~' .

fj ~ - .~

[

l

I

I"

i

•

I

,.

"'

I

.,~~

- 12 - REFERENCES 1. Lizak,. Alfred A.; Two-Dimensional Wind Tunnel Tests of an H-34 Main Rotor Airfoil Section# TREe Tech. Rep. 60-53 (SER-58304), U.S. kn:(Jy Transportation Res. Commana (Ft. Eustis} Va.}, '1"0"- Sept. 1960.

2. Scheiman, James, and Ludi, LeRoy H.: QuaJ..i tati ve EvaluEi:tion of' Ef'f'ect of' Helicopter Rotor-Blade Tip Vortex on Blade Airloads.

NASA TN D-1637, 1963.

3. Silverstein, Abe,,. and Joyner J Upshur T.: Expe:tlimentaJ.. Verif'ica- tion of the Theory of' Oscillating Airfoils. NACA Rep. 673, 1939·

-:-'""" 1,,1, ' , :-_~. t,_ ,

~! I

: ;

" I""

.-- -,; .

". II

,I

• FLIGHT AND 2-'DIM. TUNNEL~ t CASES FOR CHORDWISE PLOTS

I

o FLIGHT ONLY J

I

.. I I 'I~ "

j

~

I

,I

NASA

~I

Figure l. - u:lcal normal-force coeffj.clent B from flight data" JJ.:::: 0.2:3"

J

';I

r

fL ~ 0.23, R l:O.55

o FLIGHT DATA -- .... 2-DtM. DATA

V' :; 210

6~ '1'=1650 4~ t eN ;: 0.9

eN = 1.4

4 _ 00

-

2 x-074 0

c-' .V

~ ~=O.28 2 Oct . ~--o- __ _ 0 ." :: 225 o~·-----~ 6p eN :: I. 6

--

-

q 0 0 x

SsC .y = 195

'0=46

r CN=I.3

- . q ~ x 02 .1, :: '255° .;....= 5 "t' o C ,.

00 eN:: 1.5 2 00 .-

·'l' 0 0 00 0 '%=0.37

_-w ...... 1 ~I --",-' ,*-J .... 1 ...... 1'--'-' ........... o.a.-JI -.. ....... 1 _._LL " J_! Q.j

o .5 1.0 0 .5 1.0

x C NASA Figure 2 .... Chordwise pressure distr:l.butions.

0' "I~ j

,~ I

(

I

o FLIGHT DATA

--

2-DIM. DATA

I

W =225 ,.

8--- eN:: 1.4

I

~ =0.23 °0

I

0/= 240 !

"I; eN= I. 5 ~ =0.23

I °0

. 0 I b.J 0 .5 1.5

I

X.

c

I

I

NASA

I

Figure 3~- Chordwise pressure distributions.

J

I

I

I

.... lfi·~·¢"~

}k=O.23, ~::: 0.95 o FLIGHT DATA --2-DI~I1. DATA ~=900

'" = 45°

eN =0.2 CN=O.5 {- =0.21 ~ ~ 0.21

o

~p

--

q NASA Figure 4.- Chordwise pressure distributions.

.' ?"

.' , . I' .. , ~ , .

, .

, .

i """::.:; -.

I

i

'I

.$ o . . .... r

. JL = 0.24" R :: 0.95

::1, .

~ 1 o FLIGHT DATA --~- 2-0IM. DATA

, '" = 90

. C =O.21

N

OO .... l -' .

\/01 -%=0.18

\

o .... -

'9:. ,) •.

/ -' .. -_ c.

I - '" = 105

.' " o C =O.19

N

o -?A . .

\/0 \ !=O.17

\'\"'.' " cl'

, c·· ......

------~o= ~,.xo_

x"

.•• ~.,:~.I: " 0 .~ . . .

'" = 135

.15 000 C}J=O.16

c~ "I~ 1,' f=O.l8 - ..

'.'

:';"- ,'I

o .5 10

...... \,' x .. '

.-

:-'

c

r; :1_ .'.

NASA I').'

Figure 5.- Chordwise pressure di.stributions.

:tr:: , I, :; '.'

,;;'.t .. , .. ~ ~·I.

>", - , IL = 0.24, if =0.75.

o FLIGHT DATA . --- 2-DIM. DATA 0 0 " ";.

+= 90 2 ~ . lJI= 180 .

eN = 0.22 o~. . eN = O.5f).

'-x 2 I Q X 00 Oa 0 - = O. I 'Q. o' C = 0.27 -0..." C . "0. .... .... ...; .. r\- .'

0---·'-- ...... ---.....-"..,;;;..=-:..:-.-0- 0 I.o...- ___ .~ v_-..;. ... _..-~ 2", = 120 2 \ t/I= 195 eN = 0.27 q" . eN = O.6~ . l\p .... ~

-x o' 'Qx, - 0 25

q h. ..... n -co = 0.22 ~ - - ~11<. ........ C • "C>-_ - oof'\- 0 .... -o--:.-"Cb

o

. "'~15; 4 Q '" ;: 225

2-

eN = 0.38 \ eN = 0.84

QQ _x _ . 2 ~ X

-Q",o.... c'" 0.28 ~o..._ C = 0.24

.......... -'--.-...........a. .... ~b.Ir...bJ I C r=t -1....6-J

o .5 1.0 0 .5 1.0

x

-

c NASA.

Figure 6.- Chordwise pressure distributions.

I,

~: ·1

I'

. " '.".' ,

; I" . "

. 'I

,~: ", , . r'··, '.

, , }L= 0.24" R = (155 .

.. ,,~I

o FLIGHT DATA , --- 2'-DIM. DATA

" I"

. ','-, .'

..

'tjI = 165 C =O.64

I

N X ' 2 ~ ,c=O.27,' " ( 0-0...

A P O· -: ...o_!l!rt..a 0

,J' I

-- ' q 6 ",=195 '" = 240 -( C~=a94 C =a73 N \ )( - 0 27 - ,'( .

b c- ~ ~. j~=O.26

?~ .. "'0'"

. III ~ji'?:I-t=o.J ... I I ?"-Q-Pi-b.ol

[

o .5 1.0 0 .5 1.0

x - C ( NASA [.

Figure 7-- Chordwise pr.essure distributions • .... ~ ( '.' ('J, ,e I' i , , " ['.'

·!/ ..

r

-

R 0.95 0.75 o .30" }L =0.23 0 . 0 o 8 8 t:I . D 8 fJ ij C C 8 ~ a 8 0 0 ~ ~ 0 ij 0.°

o ~6e;to .. . 0°

.20 .10 .30 jJ; =0.18 8epoooB8ijooOOcoc8ao~88do

x

.20 ~ooo . 00

-

c J-L=O.29 I __ I I I _~.,.J 160 240 320 l/I,DEG NASA Figure 8.- Centers of pressure.