Document
,
NASA Contractor Report 178248
e
The E f f e c t of
H e a v y R a i n on an
’ , . . I .
c
Coleman duF. Ijonaldson and Roger D. Sul” ilvail
ARAP Group, Titan Systems, I n c .
Princeton, N e w Jersey
Contract N A S l - 18088 March 1987
NASA
ha1 m a l Aeroriaurlcs and S m c e A ~ T n s t w o r l Langley Research Center
*
Ha rfi p t o n V i r g I n i a 2 3665 - 52 2 5
NASA Contractor Report 178248
The Effect of
Heavy Rain on an
c
Airfoil at High Lift
Coleman duP. Donaldson and Roger D. Sullivan
ARAP Group, T i t a n S y s t e m s , I n c .
P r i n c e t o n , N e w J e r s e y
Contract NAS 1 - 18088
March 1987
National Aeronautics and Space Administration Langley Research Center Hampton,Virginia 23665-5225 1. Introduction Although the effect of heavy rain on aircraft performance was discussed t as early as 1941 by Rhode (l), no serious studies of the relationship of heavy rain to aircraft safety were made until Luers suggested in 1981 that the tor- rential rain that often occurs at the time of severe wind shear might substantially increase the danger to aircraft operating at slow speeds and high lift in the vicinity of airports. While Luer's ideas were not published until early 1983 ( 2 ) , appropriate measures were taken by NASA to study the effect of heavy rain on the lift of wings typical of commercial aircraft.
These tests, reported by Dunham, Bezos, Gentry, and Melson (3), were the sub- ject of a number of discussions between the senior author of this report and Mr. Earl Dunham of NASA during the fall of 1984. One of the aspects of these tests that seemed confirmed by the data was the existence of a "velocity effect" on the lift data. The data seemed to indicate that when all the nor- mal non-dimensional aerodynamic parameters were used to sort out the data, the effect of velocity was not accounted for, as it usually is, by the effect of dynamic pressure. Indeed, the measured lift coefficients at high lift indi- cated a drop'-off in lift coefficient for the same free-stream water content as d velocity was increased.
This behavior was explained at the time by the authors as being due to the variations in momentum deposition of the droplets that are splashed back into the air stream after shattering upon striking the airfoil surface. Since the higher the speed the smaller are the splashed back droplets, it follows that the higher the speed the closer to the airfoil will be the layer where these droplets are reaccelerated by the air stream. It was suggested that if this splashed-back momentum effect was large enough and took place close enough to the airfoil, the airfoil could stall. This would be a very serious condition if, indeed, it could occur. These ideas were put forward at a meet- ing of contractors working the heavy rain problem for NASA in April of 1985
( 4 ) -
Subsequent to this meeting, funds were obtained from NASA to modify the A . R . A . P . ARB code to allow the computation of the effect of splashed back droplets on the location of separation on airfoils at high angle of attack to see if stall could be induced and under what conditions of speed and scale Q such a phenomenon might be hazardous to aircraft operation. A secondary ob- jective of this calculation was to obtain such information as could be gleaned to aid NASA in conducting properly scaled tests of airfoils in heavy rain.
Bilanin ( 5 ) has given a more detailed review than is attempted here of the mechanisms that may be involved when aircraft operate in heavy rain.
c
Q
2. Ejecta Scalinv To derive an expression for the extent of the momentum-defect layer that results from droplet splash-back, we first determine the order of magnitude of the diameter of the droplets that splash back from a surface after drop im- pingement.
We consider a drop of radius r impinging normally with velocity V on a frictionless surface unwetted by the drop. As the drop hits the surface it is deformed into a pancake whose instantaneous radius is a and whose thickness is t as shown in Figure 2.1. From continuity, we have the result that
- 4 nr3 = na2t
3 0 At any given time, the surface energy of the distorted pancake is uw(2na2 + 2nat)
a
where uw is the surface tension of water in contact with air.
In view of (1). this may be written
- 4 nr; (7 2 2 + --) uw
(3) Generally when the drop breaks up into droplets, t<<a, so we write the surface energy approximately as 4 2
- nr3 - u
( 4 ) 3 0 t w
d
t ,
a
+., C Q, E a , M C d a E d a
*
L a cu 3 , L c , 0 : E
a
c
Now this surface energy cannot increase without bound as t becomes smaller, for the only energy available to supply this surface energy is the kinetic energy of the original drop. In an actual impingement process, energy is lost to both internal and external friction, but if we neglect these and other losses and say that droplets of diameter d = t must form where a portion of P f of the original kinetic energy has been converted to surface energy, we may write
*
or #
a
The above expression gives the order of magnitude of droplet size that can be formed. From this expression we learn that the faster one flies through the air the smaller are the drops that are splashed back from the surface of an iB airfoil. Also, we see that the droplet splash-back size is expected to be relatively independent of the initial drop size ro.
Having obtained an expression for the size drop one expects to see splashed back, we may now derive an expression for a length that is typical of the distance over which the splashed drops are accelerated. Consider the deceleration of a particle by its drag: PaV2 dV A m - - p d t - ' D T p
e
where m and A are the mass and cross-sectional area of the particle, CD is P P V is the magnitude of the the coefficient of drag, pa is the air density, velocity, and t is time. We may write this, since d s = Vdt, as
a
dV - ‘D* ’ a - - - - o - d s V m 2 P
*
There is clearly a characteristic length associated with the acceleration process, and if we assume C 1 it is D U (9) tells us is that droplets splashed back from the surface of an What Eq.
airfoil will have the momentum defect that they bring into the air flow about of Q from the surface.
the airfoil adjusted within a distance of the order C We may look at this length then as a momentum-defect deposition or adjustment I We may express this length in terms of the splash droplet diameter, length.
d by noting that for a sphere PI Putting Eq. (10) into Eq. ( 9 ) we obtain where pw is the density of water.
a
Using the diameter d given in (6) we find finally that the momentum-defect P adjustment length is What will be important in considering the effect of the momentum defect splashed back by the impinging droplets will be the ratio of this length to a typical dimension of the wing or airfoil in question. We therefore expect that an appropriate parameter f o r scaling the effect of heavy rain will be c 3f pavzc - = - - 16 Ow kc where c is the chord of the airfoil. We can get rid of the constant 3f/16
a
and take as a parameter simply This parameter (the deposition length parameter) is a mixed Weber number con- taining the density of air and the surface tension of water.
e
It is useful in order to understand the relative importance of rainfall rate and N on C , to consider the sketches in Figure 2.2a and 2.2b. In D Lmax Figure 2.2a we show for a fixed ND (thus a fixed momentum-adjustment length) an extremely simplified picture of what increasing rainfall rates will do to the velocity distribution on the upper surface of an airfoil at angle of at- tack. Clearly as the rainfall rate is increased, a larger momentum-defect will be realized, and with increasing defect will come less and less ability of the boundary layer on the upper surface to recover pressure. Thus, we
a
a
a
Y \ ' \ Figure 2.2a Effect of increasing rainfall rate on near-surface velocity distribution at a fixed value of ND, i.e. fixed momentum- defect adjustment length relat'ive to chord.
might expect the CL of an airfoil to fall off (perhaps linearly) as rain- max fall rate is increased at a fixed value of N D' On the other hand, if we consider what happens for a fixed rainfall rate as N is increased, we must see something like what is depicted in Figure D 2.2b. In this case, as N is increased, the momentum adjustment length be- D comes smaller and smaller; thus a given momentum deficit that is associated with the rainfall rate must be carried in a thinner and thinner layer. The result, at least as far as the flow at the surface is concerned, is equivalent to increasing the rainfall rate. We would expect t o find then that the effect of increasing the chord of an airfoil at a fixed rainfall rate might be
a
similar to the effect of increasing rainfall rate for a fixed chord.
a
a
0 4
a
Effect of increasing ND on near-surface Figure 2.2b velocity distribution at fixed rainfall rate.
3. Initial Examination of Data as a function of rainfall rate and N was A small amount of data on C D Lmax given to the authors by Mr. Earl Dunham of NASA/Langley. These data are shown
*
in Figure 3.1. A fair amount of liberty has been taken in fairing curves through the data on the basis of our feeling that the effect of rainfall rate on C might be linear. The fairing was also biased by our notion that the Lmax
*
effects of ND and rainfall rate might be similar.
With great trepidation the results of the initial fairing of these data are shown in Figure 3.2, as a function of N with rainfall rate as a D parameter. The CL data are plotted in terms of 1-CL /(CL ) o , where max max max ) o is the value of C for zero rainfall rate.
(CL max Lmax On this figure we have also indicated the range of N that will be found D for large jet transport aircraft during landing operations and the range of N D for which water impingement tests have been carried out at MIT.
First of all, it should be noted that Reynolds number a parameter that must be considered, perhaps by a consideration of the ratio of to the C boundary layer thickness. However, it is now believed best to use N as given D here as the essential parameter and consider Reynolds number effects c separately.
In view of the possibility that N effects are similar to rainfall rate D effects as pointed out above, an initial conjectural extrapolation of the data to the N values that might be associated with large scale transports was made D and is given in Figure 3.2 for a rainfall rate of 75 mm/hr (3 in/hr). This extrapolation indicated that such a rate might cause a 15 to 20 percent reduc- tion of CL
. Clearly the implication of this hypothetical scaling poses a
max
a
e X x o x m = Y N .. 9 c . ( cy -0 rD CI
- s
M ' c y d -0 (0 * X x X c3 z y m c ( ? m N cy (9 '9 s Y 9 m cy cy N N N X W I '3 P z Y rn cu 0 -4 c 'rl
\
c , c cu
\
m v) Q X
\
m m -0 E rn C
\
m c, m U n
\
L
-
m
c -
m U A z cu M c .4 c ,
\ 0
\ L
\
\ 4
a L
'Tt S
r L z m m r L
B
v)
L 3
c, (z s FY Q m FY (z .r)
P
-c( E
-
u 5 :
-
(u Q L M 'rl P , (u - r n I -
I " I I I " ' I I " '
& sufficiently severe operational problem so that a large scale wing should be tested in a rain environment.
In addition it was clear than an attempt t o calculate the effect of I droplet splash-back on the maximum lift of an airfoil would be a most useful exercise. In what follows we will describe the program that was used to make such calculations and the results of these calculations.
a
Q
4 . Proerram Development and Use The calculation of the maximum lift of a two-dimensional airfoil even without rain is not yet a science. Therefore, rather than try to calculate the effect of rain on maximum lift, the effect on the position of separation for an airfoil near maximum lift was studied. An NACA 64-210 airfoil at an angle of attack, a, of 1 2 O was chosen.
To determine the point of separation, a program called ARB, which has been in use at A . R . A . P . for several years ( 6 ) , was used. ARB, which employs full second-order closure models for turbulence, computes the boundary layer
a (using standard boundary-layer assumptions) on an axisymmetric body of ar-
bitrary shape, whether rotating or not, as well as on two-dimensional bodies.
It handles compressible flows, solving for the extra second-order correlations that exist because of fluctuations in temperature and density, as well as constant-density flows. Since the present study concerns low-speed flows, the constant density mode was used.
A special version of ARB was developed for this project in order to cal- culate the effect of particles (water droplets) as they splatter back into the
*
boundary layer after hitting the surface of an airfoil. A description of the analysis underlying the modification of ARB follows.
4 . 1 Motion of a particle in a boundary layer In analyzing the motion of a particle in a boundary layer, it is assumed that the variation of the mean flow with x, the streamwise direction, is neg- ligible. It is also assumed that the flow is two dimensional and that the
*
effect of particle motion in the third dimension can be neglected after averaging over a large number of particles.
a
r
Consider then a particle moving with velocity P = (up,vp) in a mean flow + v = (u,v) (functions of y only).
In terms of the relative velocity, -+ + - + v = v-v the force on the particle is r P ’
a
where C is the drag coefficient, A is the cross-sectional area of the par- D P ticle, and pa is the density of air.
Then the acceleration of the particle is
a
where m is the mass of the particle and P To evaluate the drag coefficient, the particle is assumed to be a sphere
of diameter d . In terms of the Reynolds number
P where v is the kinematic viscosity of air, CD may be approximated by (See Figure 1.5 in ( 7 ) . ) For a sphere, P 4 w - - - - d p &c 3 Pa where p is the density of water, and so W The acceleration can be rewritten du = D (u-u ) dt K P dv P = DK(v-vp) dt where n " L e D - .
- -
jvrj = -
' K - kc In order to evaluate the effect of the particles on the air flow it is neces-
sary to know 2 as a function of y. Since
P
e
the acceleration equations can be written du D 2 = (u-up) dY vp dv DK
2 = - ( v - v , )
dY vp
e
Thus if the initial values of u and v as a particle comes off the surface P P are known, and if the flow conditions are known as a function of y, these equations can be integrated and the force evaluated.
If there are n particles per unit volume so that the liquid water con- P tent is p = n w the x component of the total force exerted by the particles P P P’ on the air, equal and opposite to the drag of the particles, is
a
per unit volume. This term was added on the right side of the mean momentum equation ( x component) in ARB. No way was found to incorporate the effect of
a
the vertical component of the force from the particles on the fluid in the context of the boundary-layer assumptions on which ARB is based, since these assumptions imply that the vertical component of the mean momentum equation is automatically in balance and the vertical component of velocity is computed I from the continuity equation.
However it was recognized that if Q is very small, a particle will give C up its momentum to the air almost immediately so that the effect can be simu- lated by applying a blowing boundary condition with the velocity at the wall, v given by W’ The results of these runs are shown in Section 5 as limits for Q /c equal cm zero although they were actually run with Q /c of the order of 2 ~ l O - ~ .
cm The rate at which the wake of a particle receives energy is Therefore the rate at which the fluid receives energy per unit volume is It is assumed that a fraction, e, of this energy produces turbulence, the rest going into heat. Therefore a source term equal to
e
-- -
was included in the equations for u'u', v'v', and w'w' which together con- stitute the tarbilent energy. For all the runs described in Section 5 , e was taken to be 1/2.
4 . 2 The variation with streamwise distance The analysis above assumes that the variation with x is negligible; however as the boundary-layer calculation moves downstream from the stagnation -c point the flow velocity v(y) changes (partly as a result of the force calcu- lated above) and the characteristics of the particles splashed back also change.
Let v be the normal component of the velocity of a raindrop as it in- pinges on the airfoil surface. Then, neglecting the settling speed of the raindrops in comparison with the aircraft speed, u-, v = uoosin 8 where 8 is the local angle of the airfoil surface.
a
The particles are assumed to leave the surface normally at a given proportion of vo, say p. Then u ( 0 ) = 0 P = pvo = puoosinO = v vp(0) sine Pm where v pm- - pu-.
It is assumed that no particles are splashed back where 0 is zero or negative.
It is shown in Section 2 that II is inversely proportional to the square C of vo so it can be written ! I c = II /sin20 cm where II is the minimum value of IIc.
cm 4 . 3 Pressure distribution ARB requires the free-stream pressure distribution or the free-stream velocity distribution (from which the pressure distribution can be calculated) as input. For the runs reported in the next section the inviscid flow over the airfoil was calculated by the method developed by J. L. Hess. The results of these calculations were found to be a little noisy so they were passed through a smoother before being supplied to ARB as outer boundary conditions.
The distribution of ue, the free-stream velocity, as a function of x , measured along the surface from the stagnation point, is shown in Figure 4 . 1 .
Figure 4 . 2 is a diagram of the airfoil at a = 120 with the stagnation point, the leading edge (defined here as the point where 8 = Q O O ) , the point of mini- mum pressure (maximum u ) , and the point of zero slope ( 8 = 0 ) indicated. It e will be noted that the point of minimum pressure is almost at the leading edge and that on the upper surface a region of large adverse pressure gradient oc- curs between the leading edge and the point of zero slope. In this region we might expect a very heavy rain to cause the flow to separate and hence cause premature stall of the airfoil.
2 1 I I e 0 0 *I 0 d ?
-I o m - X I 0 0 a - d a i 0 P > - - d o
i
c u m
0 II -
>
g a . 2
. d Y Y C t u C v) 4J C -4 a (d V d 4J & cu 0 0 L (u II QI 5 . Computations and Results Many calculations were made of the behavior of the boundary layer for a number of conditions of rainfall rate (represented by p ) , particle size P (represented by II ) , initial particle velocity (represented by v ) , and cm Pm Reynolds number. Figure 5.1 shows typical behaviors of calculated skin fric- tion coefficient, cf, for three values of p /p with tcm/c = 0.001, v /uoo = P a Pm 0 . 4 , and a Reynolds number of 3.5~10’. This is a Reynolds number typical of a 747 aircraft near landing speed. The curve for p /p = 0 shows that, as ex- P a pected, the aircraft does not stall and the point of separation (cf = 0) is near the trailing edge of the airfoil. The curve for p /p = 0.1, repre- P a senting a high rainfall rate, shows that the separation point has moved forward somewhat so we might expect some drop-off in CL but the airfoil has by I no means stalled. Note however that the dip in cf in the region of strong adverse pressure gradient has become more pronounced. The third curve, for a still higher rate, p /p = 0.32, shows that the separation point has jumped P a forward to that region and we would expect that the airfoil has stalled.
(I Figure 5 . 2 shows the value of x/c at which separation occurs (plotted horizontally) as a function of p /p (plotted vertically) for the three runs P a shown in Figure 5.1 and other runs all at the same k /c, v /us, and Reynolds cm Pm number. It is clear from this figure that the critical value of p / p , that P a required to cause early separation (which we assume is equivalent to stall), is 0.113.
Repeating the process described above for different values of II /c, cm still at the same initial velocity and Reynolds number, we obtain the results shown in Figure 5.3. The open circles are the calculated values of p /p re- P a quired to trigger early separation at various values of I t /c. When results cm such as these were first encountered they caused some consternation, for cer- tainly we did not expect the rainfall rate required for early separation to increase as the momentum defect adjustment length, Qc, decreases.
The ques- tionable points are connected by a dashed curve in the figure. An analysis of e P E m
*
w - 2 0 II
c o +I+ 8
e o c a c a 0 \ e a > rl II
*
\ B N e rl In 4 I I # 1 . .
a t N rl L , 0 0 a 0 M 0 0 e e e Lu m a m U E cna m c u p 3 a & a M v) d 0
a7 2
a
L m , c u \ E V ol P C I C o m N r( a m 0 c W O
I'\ \
I I I N 4: 0
a
a
this phenomenon showed it to be related to the fact that as II becomes small C compared to the boundary-layer thickness, the effect of the particles on the vertical momentum in the boundary layer needs to be included in the calcula- tions. As mentioned in Section 4 . 1 , this is not possible in the ARB code,
a
but a way was found to get around this problem, and it certainly is a problem, by using the surface gas injection feature of ARB for the limit E = 0. The C result of such a calculation is shown by the filled circle for E /c = 0. For cm the purposes of estimation, the authors, for want of any better method, have
a
faired a smooth curve shown by the dot-dash curve in Figure 5 . 3 , between this point and the other points we believe to be valid.
Curves such as that shown in Figure 5 . 3 have been calculated for a number
a
Summary plots of this of Reynolds numbers and two values of the ratio vpm/um' nature obtained from many runs are shown in Figures 5 . 4 and 5 . 5 .
Additional scales have been included on these plots. I f equations 13 and 14 of Section 2 are combined. we find where II has been used instead of II because the value of V meant in the cm C definition of N D ( = p V2c/aw) is here taken to be the speed of the aircraft, a (See Section 4 . 2 . ) Taking the value of f , the portion of the kinetic
-
energy of the drops impinging on the surface converted to surface energy, to be 1/2, we can write Values of N thus calculated are indicated as additional horizontal scales in D Figures 5 . 4 and 5 . 5 . The additional vertical scales are rainfall rates, ob- tained from p /p under the following assumptions. The settling velocity of P a the raindrops is taken to be about 5 m/sec and it is assumed that half the ?
n cn a m r( 3 Q cu c ?
Y) 0 R al 0) 0: d c c , Q -4 Q V -4 c , -4 Ll J u ? Q, c c , cu C -4 c , Q Ll
s
I I w L ?
" I n a l Ll en 0 crr .N L a al CI Y D al m E (D la N L al Y 0 a a a L II la a 4 v) 1 \ a a E Cil L a C -A 3 Y n a L c E c, a .r( L c 0 3 D a m A (u Q ...I C h I1 pi cr: a c c, c, la .-I
\
\
\
a M water impinging on the wing is splashed back. Then it can be shown that at standard conditions the rainfall rate in mm/hr is approximately 4x104 p /p .
P a Examining Figure 5.4, for v /u- = 0.4, we note that, as expected, for a Pm low Reynolds number ( 6 . 5 ~ 1 0 ~ ) the airfoil stalls even when there is no rain- fall. A s the Reynolds number is increased to the level of the NASA tests mentioned above (2.4x106), the rainfall required to cause early separation decreases as ! 2 /c decreases, i.e. as N increases. For Reynolds numbers cm D typical of transport aircraft ( 1 . 8 ~ 1 0 ~ and 3 . 5 ~ 1 0 ~ ) the calculations indicate that the increased Reynolds number offsets the effect of pure size on N so D that a somewhat larger rainfall rate is required to induce early stall for a 747 aircraft than was found for conditions of the NASA tests. This is an in- teresting result and is at variance with the pessimistic results of the authors’ first extrapolation of the NASA data (see Figure 3.2) = 0.6. It was anticipated Figure 5.5 shows similar results for v
*
pm/U- that this parameter, which relates the normal velocity of the incoming drop to the initial velocity of the droplets resulting from its fracture, would be a very sensitive parameter in the interaction between the droplets and the boundary-layer. A comparison of Figures 5.4 and 5.5 shows this to be the case. For example, the rainfall rate required to induce early separation for the conditions of the NASA tests dropped from about 1400 mm/hr to about 800 mm/hr .
c 1 )
a
6 . Discussion and Recommendations The computations that have just been presented are believed to give the general trends associated with the phenomenon of premature stall due to cou- pling of the splash-back droplets with the boundary layer. Clearly the numerical values of rainfall rate to produce premature separation of a given Reynolds number and mixed Weber number (the deposition number N D ) are not exact. However, we believe this to be of the right order of magnitude.
If this conclusion is correct it follows that for rainfall rates ap- proaching 500 mm/hour or greater, commercial aircraft might be subject to rain induced premature stall. The calculations also indicate that both Reynolds number and scale are important in pinning down the values of rainfall rate that will result in early separation due to ejecta coupling with the boundary layer. In view of these facts and in view of the fact that other extremely complex phenomena also have an effect on boundary layer behaviors in the presence of heavy rain, it is clearly necessary to conduct full scale tests if a quantitative assessment of heavy rain hazard is to be made. This is true for simple wings and is certainly even more true of the flapped and slotted wings typical of commercial aircraft.
In regard to analysis, although it is believed that certain trends that have been calculated with the extended ARB program are correct, it has been shown that, if more exact computations are required, a method more powerful than conventional boundary layer theory must be applied. Since such analysis will be expensive, it seems desireable to hold off any further development of analytical capability until such time as both full scale and model tests of the same wing configuration can be completed so that the importance of heavy rain effects on wing performance can be evaluated more precisely.
References 1 . Rhode, R. V., "Some Effects of Rainfall on Flight of Airplanes and on Instrument Indications," NACA Tech. Note 803, 1941.
2 .
Haines, R. A. and Luers, J. K., "Aerodynamic Penalties of Heavy Rain on a Landing Aircraft," J. Aircraft, Vol. 20, No. 2, February 1983.
3 .
Dunham, E. R., Bezos, G. H., Gentry, C. L. and Helson, E . , "Two- Dimensional Wind Tunnel Tests of a Transport-Type Airfoil in a Water 0 Spray, AIM-85-0258.
4.
Donaldson, C. duP., "On the Scaling of the Effect of Heavy Rain on the Maximum Lift of Airfoils, A.R.A.P. Technical Memorandum 85-33, August, 1985.
5. Bilanin, A. J., "Scaling L a w s for Testing of High Lift Airfoils Under
a
Heavy Rainfall," presented at A I M 23rd Aerospace Sciences Htg., January 1985, Reno, Nevada, AIM-85-0257.
6. Sullivan, R. D . , and Varma, A. K., "ARB: A Program to Compute the Turbulent Boundary Layer on an Arbitrary Body of Resolution," A.R.A.P.
Report No. 317, January 1978. Also: Sullivan, R. D., "ARB: A Supplementary Manual," A.R.A.P. Tech Memo 80-16, September 1980.
7. Schlichting, H., Boundary Layer Theory, Fourth Edition, HcGraw-Hill Book Co., Inc., translated by J. Kestin, 1960.
Standard Bibliographic Page . Report No. 2. Government Accession No. 3. Recipient's Catalog No.
NASA CR-178248 . Title and Subtitle 5. Report Date March 1987 The Effect of Heavy Rain on an Airfoil at High Lift 6. Performing Organization Code '. Author(s) 8. Performing Organization Report No.
ARAP Report No. 597 Coleman duP. Donaldson and Roger D. Sullivan 10. Work Unit No.
I. Performing Organization Name and Address A.R.A.P. Group 11. Contract or Grant No.
Titan Systems, Inc.
NAS1-18088 Princeton, New Jersey 08540 13. Type of Report and Period Covered 2. Sponsoring Agency Name and Address Contractor Report National Aeronautics and Space Administration 14. SponsorinK Agency Code Washington, DC 20546 505-45-13-01 5. Supplementary Notes Langley Technical Monitor: R. Earl Dunham, Jr.
6. Abstract Although the effect of heavy rain on aircraft performance was discussed as early as 1941 by Rhode (11, no serious studies of the relationship of heavy rain to aircraft safety were made until Luers suggested in 1981 that the torrential rain that often occurs at the time of severe wind shear might substantially increase the danger to aircraft operating at slow speeds and high lift in the vicinity of airports. While Luer's ideas were not published (2), appropriate measures were taken by NASA to study the until early 1983 effect of heavy rain on the lift of wings typical of commercial aircraft.
These tests, reported by Dunham, Bezos, Gentry, and Melson ( 3 ) , were the subject of a number of discussions between the senior author of this report and Mr. Earl Dunham of NASA during the fall of 1984. One of the aspects of these tests that seemed confirmed by the data was the existence of a "velocity effect" on the lift data. The data seemed to indicate that when all the normal non-dimensional aerodynamic parameters were used to sort out the data, the effect of velocity was not accounted for, as it usually is, by the effect of dynamic pressure. Indeed, the measured lift coefficients at high lift indicated a drop-off in lift coefficient for the same free-stream water content as velocity was increased.
17. Key Words (Suggested by Authors(s)) 18. Distribution Statement
Unclassified - Unlimited
Heavy r.ain Airfoil Boundary Layer High lift Subject Category. 02 21. No. of Pages 22. Price 19. Security Classif.(of this report) 20. Security Classif.(of this page) Unclassified Unclassified 34 A03 For sale by the National Technical Information Service, Springfield, Virginia 22161 NASA Langley Form 63 (June 1985)