I
i
TI CflX CAL XgMOBANDUMS
NAT±ONAL ADVISOBY 00MMITTF
?5B AROAUTICS
ANkIYSIS 01' xP
IN!1AL IN
5TI(ATtC)NS 01' TPLANIN
PR00 SS ON TB SUR1'A0 0?
By W. Sottorf
ahrbueh 1937 der Deut.chen LuftfahrtforachUng No. 1061
Wae)4ngt on
March 1944 TRMDOCLJMt«
FILE I rMAR 30 1944
NACA
LIBRARY
LMLEY MEMORIALAERCNAUTICL
L&5OIATORY Langley Fie1d, V
Delft University of Technology
Ship Hydromechanics iaboratory
Library
Mekelweg 2 26282 CD Deift
Phone: +31 (0)15 2786873 E-mail: p.w.deheer©tudelft.nl
i.
i
n-- NAP IONAL .4DVISOT OGUXI!TBE "OR LRONAUT lOS
TEOESXQAL MEXORIJDtTM NO. 1061. .AJALYSIS o
xxpu nrvzsTIGATzoNs
6i
Ta
PBOOESS ON TE3 SUBJAC 07 WATEB By W. Sottorf
Pressure diotributio.n arid. spray measu.venints were carried, out on reotangular flat and. Vbot.,tam planing
surfaces.
Lift, resistance, a4 centerofreesujro data
are analyzed and. it le shown how these values niay be
oomp-u.ted. for the pure planing procese of a flat or V-bottom surface o' arbitrary beam, load, arid speed., the
method being illustrated. with the aid. of an example.
SUMMARY
Plat
and. Vbottom ongitudinal1y straight planirig surfaces are investigated.. Por such surfaces the total resistance nay, for a given lift, be separated. into a normal and. frictional component. The analysia of thé tests leads to representations of the lift and. centerof-pressure position aS functions of the aapeot ratio of the pressure area with the Proud.e number, which characterizes the effect of gravity, aB parameter. While the fric-tional resistance coefficient Is equal to that given by Pran&tl for theturbulet boundary layer with preaeding laminar layer. Por high 3'roud.é numbers, for which the.contribut ion of gra.vity. to the lift is negligibly small,
there is shown to be an..agreemerit between the planing
surface. tests' and. the Lift on flat airfoils on the under aide. The deviationà from the Wagner theory for short plates at large aspect ratios are considerable thougb,
justified by- the cond.itiona neglected. in the theory.
With the aid. of a working chart a number of practi-cal examleø are computed. that provid.e the answers to several important quastíon. Por the caøe of the flat plate the width of the plate, which is always the full yidth of the planing. surface hae an effect on the
re-istance-].oad. ratio in that the latter' becomes more
favorable rith increasing width.
tAnalyao oxperimentellei' Untorsuchungon fiber den
].elt-vorgang an der 3ahrbuch 193? der
2 ' N.Ø.L T No. 1061
With regarto
the
]oád.affebt ,ot 'ac0unt of theimpairr.ient in the aepect ratio there is an impairment in the rciBtaace].oad ratio with increasing load..,
1ith regard t the, efact of the speed., however, on account of thé improvement in the aspect ratio with increasing dynamic pressure above, the first reeistene maximun, there is an improvement'in the resistanceload, ratio zith increasing speed..
The teet for he.ac.leeffect sh6w that there'is
similarity of thè preestre eurface and. presure
d.is-tributionathát"th
ffet of the.écale. 1B.given onlyby thQ d.epend.ne. of ta frïctioii coef1'1cfrxt 'on the Beynol,d.s numbq'. Ouiy forvery small d.ins1ons and.
loads of the planing surfacés does onsimiIarity occur In the' flow oond.itions because of the effect of surface t ene ion.
The Impairment In the resistanceload. ratio by the effect of the V bottom is a].Bo shown. With regard. to
the effect bf the width it is found Zo.r the V bottom that at rn,11 load. or,high dynamip preesu'ean. with 'he "natural width," :vhioh is ]ese than that of the p1aing
surface, .n optimum width oc1.trs which Is to be deter-mined for each particular case.
INTRODUCTION
when, with the
eve1opment of aircraft, the
invti-gation and. d.evelopment of seaplane floats 'were added tothé'usual problems of he towing tank, theré was a lack of theoreticl as well as expérimental underlying bases for Ovaluatin.gtheteat results. A first expöriment with
a flat planing surface was carried. out in 1912 In nland.
at the lIillia,m Yroud.e Laboratory and. a report was
pre-sented by Baker and Il1ar (ieference i), who, however, did. not continue any further work of fundamental invest
i-gatlon. In order' to create a sufficiently wide basis, teste on planlng ,s'!lrfaces ha'ebéen conducted by the author since 1928 at' th.Rambüg Sh.p Conetuction
xperienta]. Inßtitute. Starting with testa' o a 1át
rectn.ular plañiug,B'rfe ybich served. for the tu_y of the planittg pocesa ánd.' its effect o the fli{
NAOA Ti! &o. 108]. 3 could. be furnished, to the moet inportaxit restions
arie-.-Ing for the oonstructoi and research engineer: namely, the effect of the beam, loading, speed., and. V angle on the resistance and. the effect of the sca'e in applying the model test resulte to tb6 fullscale design. Thsse
uestlone nay be answered. from the resulte of planing surface tests Insofar as the pure planing condition is being considered, this oond.itiou being ciaracterised. In a float by the breaking away of the vater at the step
and. side edges of the port ion of the bottom in front of the step and. also by the breaking of the water coxzta&b
at the stern, (See reference 2.) The results vere
partially published in the following years (reference 3). Meanwhile, sirias a theoretical treatment of the actual planing process that take all conditions luto account
does not appear possible, Wagner (reference 4)concerned himself with the limiting cases of the planing process (high.. speed planingneglecting gravity, and. infinitely.
small trin angles) and- by the application of the airfoil comparison, presented an approximate experimental method. that enabled him to determine the forces and. wetted sur-face of the flat plate with small aspect ratio aleo for finite trin angles. The comparison of the first teat
re-suits of the aut'hor with the Wagner theory showed that in the range of small aspect ratios rough agreement was ob-tained while for large aspect ratios the deviation from the theory was considerable. Wagner' aecrlbed.this
d.evia-tion to the effect of gravity. Sambraus (reference 5) carried. out supplementary tests at higher speeds at the Prussian Bxperimentai Institute for Ship Constitution in
order to test the Wagner theory also with neglect of -gravity for higher aspect ratios. Prom his result
Sambraus concluded. that the Wagner theory ho-ida true at
high planing speeds for all aspect ratios. Shoemaker (reference s) likevIe. extended the range of investiga-tion by totiing a series of flat and. Vbottom planing
surfaces in the SACA tank.
-Th the takeoff of a seaplane the lowspeôd. planing stage, during which gravity exerts a considerable effect,
la of oqua,l importance with the highspeed planing stage. The analysis, preBented. in this paper, of all the
authorts own teats as well se those of Sambraus and
Shoemaker, leads to a clear explanation of the effect of gravity rind, hence to the representations of lift, re-Bistance and center of pressure, and with the aid. of this
analysis these values may be numericall' determined. for flat ririd. Vbottom planing surfaces of arbitrary beam and.
4
ACA M No. 1061
arbitrary load.iug ad speed. for the cond.ition of pure planing.NOTATIoeT
A total lift load. (kg)
Aìyn dynamic lift (kg)
stat static lift (kg) G total weight (kg) W resiBtance (kg) normal resistance (kg) frictional resistance (kg) S :po:peller thrust (kg) Z to'ring pull (kg) N normal forces (kg) T tangential force (kg) 3 V static displacement (m ) load. roduction (kg)
Mhst mouent about the lateral axis through step (t,ail-ing edge for flat surface, after end. of keel for Vbottom surface) (mkg)
y speed. Cm/e)
mean velocity of the water along the planing
surface (rn/B)
mean vertical velocity of the mass of water
dis-placed. (in/s)
roude number referred to a f ixd. dimension of the body
Ca* =
b lift coefficient of the pianiig
eurface
derived, lift coefficiint
of the planing eurfaceA. cBa = A e
- -
load, coefficient -.2Dq,
Cf, Cf1 friotiona], coefficient - lih St Cmh Y b4 load, coefficient moment coefficientp/2 v.
d.yuamio preoeure (kg/m2) p - d.enaity (kg/in4)
'1 specif,o weight '(kg/m5)V
kineaatic
vieooeity (ma/e)L0.& TM NoS. 1061 5
r' =
v/h11v)n1
'rou.d.e nber tefexred. to a length correepond.ing to the loadR . vi/i, Beynolda namber
= Planíñg Number1 reeletanceload ratio
'A
Cat
?b
load. coefficientca lift coefficient of the airfoil
S S
e N.LcL&
:
No. 106].
P PdYXZ.3 at.at .re.Bsue,(kg/in2)
mean pressure on "pressure area (kg/rn5)
mean length of wetted surface Cm)
mean lenth of pressure area for Vbottom planiiig
surface (nc)..
distance of center of preasure from trailing
edge
of planing
surface (in)-local elevation (in)
Y ,stted auiface, also pree sure areá for flat
plan-ing
surface (in5)projecion of pressure arca for Vbotom planing
surface (in2)7
wing
area(in2)
b
beam of planing surf aceb5t 'beam of float at step (in)
bnat
naturalwidth of
pressure area for Vbottomsur-face (in)
o distance of fountain from tiailing edge of
plan-ing surface (m)
ti im
X scale of model dead. rise angle
NACA TI
No.. 1061
XPEE IMNTAX -PROCEDURE Measurement of the Porces
Heaeurementg were mad.e on the lift, reeistance, and. noment about a lateral axis. The threecomponent measuring gear is schematioa].Ly shown in figure 1. The resistance tension vire leads forward. from the point
Po, through which passes. the axis of the model head, to the dynamometer. The resistance rneasuremnt -is made by means of a recording pen and. coarse weight, the
spriig- force being. determined from the recorA of the spring extension, the -spring having been previously calibrated. A constant weight serves to maintain the tautness of the tension member. The calibration wire,
in addition, takes up by means of a stop the inertia force that is set up on the model on braking the tow-ing carriage. At equ..]. distances from po there are attached. the vertical tension wires leading to the moment beam. With the aid of a sliding weight the moment appl-ied to the model and hence the trim angle may be varied. At po there is attached. an ada.itiona].
vertical wire that leads to the unloading beam. By
varying the size or the p-osition of the eliding weight the load on the model is reduced so that the remaining weight corresponds to the loading desired. One of the reversing wheels is formed aS a gear wheel and. to this
io attached. by means of a. coupling, a winch by moans of which the model at the end. of each test run is lifted from the water in order to remove it from the following waves and. utilize the time of the return trip of th carricige for quieting the water surface.
At the start of each test run the. model is reL. leased only when the 'dynamic lift is sufficient to -S.
avo1d. an undercutting of the leading edge of the plc.n ing surface.
The draft and trim of the model were determined. from two draft Bcales attached. to the wheels of the
moment- -ba1anca.- - The setting for a given trim angle is obtained by displacement of the sliding weight of the moment balance and. reading of an angleindicating device. Damping when required. is exerted at the d.ynamometer
lever (damping of lon.gitud.inal oscillations) and at the slide of the moment balance (damping of vertical
-8 -
iL M NO
-1061The relative a.r velocity at the location of the
model Is, on account of the closed, carriage and two air
scoops, practically equal to zero so that the air
re-aistance of the inode]
,s ngeliglbly email aleo at high
speed.s.
The measurements are therefore purely
hydrody-iianiical,
The preliminary testa were conducted. with flat
glass platee In -order that the forward boundary line of
the wetted surface might be determined.
This boundary
line under ail speeds, loads, and trim angles had. the
shape of a very flat arc.
In -the subsequent teats
there 'rere therefore employed plates, some of aluminum
and. some of wood, glass etrips being inserted at
one-quarter width so that the mean value of the wetted
length could be read off.
In the case of very wide and
V-bottom surfaces several glass strips were used.
The model was su8pended. without crowding so that
all the errors in measuring the draft, trim, and. lougi-'
tu.d.inal displacement of the mode]. were negligibly smäll.
insofar a they were not eliminated by the calibration
itself or by having the vires run on circular segments
where required.
Pressure Distribution Measurement
Por' taking the pressure distribution measurements
the planing surface was provided, with about ninety
measuring stations which were arranged in thrQe longi-'
t-u.&inal sections and. a number of transverse sections.
Pigure 2 shows a measuring station in. cross section.
The free opening of the orifice is 2 mIllimeters In
diameter.
The glass tubes over the orifices ae held.
fixed in position by means of a short pipe construction.
In taking the measurement the planing surface is held
fixed at the trim determined during the resistance
measurements, the measurements being cond.ucted with the
water surface smooth.
The heights of the water columns
are marked,
It was checked to see whether there was a
poible interference effect of adjacent orifìcês by
closing the QrifIces 'with plasticine in the cheek runs
an& keeping oiily- one orifice open.
o apprecia.ble
dif-ferences were observed.
Check measurements with 1
millimeter free opening of the orlfiâes showed, on
.a-count of the stronger throttling of the liquid column.a,
a smaller fluctuation.
jCA TM :No.
I Oi
9Yor obtainin
the preasurs distribution, lt was
- necessary ,to make aeverar test- run
at each trim, angle,..
alight speed. differences which 'àìéäted. th61áad.ig on
the adjusted plate bei.ng foun4 unavoidable.
The
meas-iiring accuracy is to be estimated at about
5.pereent
of
¿h.
THEOETICALBÁSÍS 01'
STSl'or a surface in steady plaxiflg motion on a quiet
frictionleas fluid. surface (fig. 3) the Bernoulli law
may beapplied. tg all the streamlines including the
d.iytd.ijg stréamiine.
.L.
+ +g z = constant
(i)
The pressure
p
is assumed 'to. be composed. of the
static pressure
stat
due to the weight of the fluid.
and. the portion
Pd.ydue to the dynamic effects.
The
staic lift is therefore'
stat = COB
(LfPstat
dl'
(a)
an1 for a flat rectangular planing auifaae may
approxi-maely be set equal to
-2
(3)
64
(os v'
" / I4
if the wetted length
Zle determined during the teat,
it being assumed. that the rise in level at the
neighbor-oo& of the plániúg eiirfàoe 18 -obnstant.
-The dynamIc lift is
10 .NLC1 TX No. 1.061
and.acôórd.ing tó the momentum law is equal to the time rate. of change. of momentum of the .aBB of vater involved.
d. (in
y0)
dyn
dt
The total lift of the planing surface is thus
A
A2tt
+ Myn
In the case of a iscoue fluid the streamlines
and. velocities, in spite of the bound.ary layer occurring, agree approximately with those of the frictionless flow, differing only in the appearance of an ad.d.itional
tan-gentia]. force T.
Owing to-the effect of the finite width there is set up at the edges a áross hoy with lateral pressure drop which for sufficiently large values ).ead.è tÖ a separa-tion of the water aleo at the aide edges.
In the pure planing phase .haracterized. by- the separation cf the water at the trailing and. aid.e edges of the planing surface the resistance of a flat surface
as i'ol1 as of all surfaces that are longitudinally straight and. without twist is given by
Y = + = A tan a + T
cas a.
Thè normal resistance
= A tan c (8)
is proportional to the loadand. the trim angle but in-dependent of the speed, while the frictional resistance
---- T = Of1
q. (8a)COB (L
-is proportional to the wetted surface and. the dynamic pressure1 the friction coefficient Cf1 being a function
(5)
1,2
and.P = 80 m
çorrospoud.iug to
65 kg/m2
surface loading,
NACA TX No 3dì
11
of the
eynold.e number
R = vs/v
an& d.epend.1.ng on .the.
turof the bound.ary lay-ex' .as..,wa13aa...qnthe
rough-ness o
the su.rface.
.R.SULTS
di
ÌEE TESTS AND APPIiIOAT IONS
(A)h flat P1anig Srface
(i) iethiì
f''he' reliminax'y Tests 'with Plat
Planing Suzface
l'or the pre1iminár. tese on the flat rectangular
planing surface average re1at.ons were chosen with
re-gerd. to boam1 load., and. speed.
ligure 4 shows a
loax'ithiiic plot of the beam at step
b3't
as a
func-t1.on ¿f t1p f.ying boat weight
0',or
G/2
in the' case
o
t
tijnf1oa
seaplane, the load. coefficient
Ca'
bing taken as pars.metei.
Select.ng' from'the thean raÊ.ge
a.beam 'bSt
- 1.800
meters
and. with
G = 5.18b t
and.c'
0.889,
furthermore a mean airfoil lift
= Ca P q. = 1296 kilogralnB,1
.nd.èpendent of the trim
anglo and. assuming.a model scale X = 6,
then the model
beam is
bst = 0.300 mete?, the corroepoud.ing weight
G = 24 kilograms
and. the hpdrod.ynamic load, at
y = 6
meterß pr second. - an aorago speed. in the spoed. range
in which tho se.plane float executes 'a puro 'planing
motión -
18 kilograxnB,
then
= 0.109
and.3.74
The trim a2xg].e range of interest 10
a.2°
'to 100.
Resiatanbe,' inoment, center of preesure,- and. static
2.L t. On figure 5 -are plotted the nonaimons lonal values
C,
Crn',
i/b,
7,,/t,
and.Astat/A
a
functions of
u.
Tan
.r'oprosents the lower
lug' value of the re-
-si'a'1' 'ôfroTh
'
'..i
ccrrdimg-to eq.uation
(7) the fr±ctional resistance
NR,' 'on the assumption
of frictionloss flulod., becomes zero,
Since with
12
NAOA.TM No.1061
por.'tlonof:th.reaie±ance
Vto the total resistance
Y
rapIdly d.eoreas-s on account of the øtrong1
deceas-Ing wottod. area
7,
as
ay be seen fromthe trend of
¡/b,
the
curve approaches asymptotically the value
tan
> 100 may be replaced approximately
by
tan &.
Another -asnptoteis the axisof ordinates
since
Pandhence.
Rapproach infinity aS
-
0.
The position of tho minimum lying between the asymptotes
is d.otormned. by- ther.atioof..tho frictional to the
total rosiatance.
Por avorago
onditions this position
lies botwoon
40 an60
Proaaure distribution and. character of flow. Pigure
6 shois the measured pressure d.intrib.tLon and. the flow
picture for tho anglos
40,
6, and.80.
The etanation
point liös nóar the 1eading. edge of the wotted. surface.
The groatost portion0f the watoretroam flowing up is
thus d.oyiatod. downward whiJ.e the portion.of the water
lying shoed. of the etagnattòù point le pro3eóted. forward.
& spray.
The maximum prouro, measurettat the
stag-nation point remains considerably below tho stagstag-nation
preesuro, .It is présumod. that the ful). stagnation
pres-sure occurs within..sD narrow a rango that it does not
show up in the measurement.
On account of the sharpedge boundary. of the
plan-ing suriaco the flow at tha trailplan-ing edge already begins
to separato at.
Qmp.aratiToly low speed and the water
continuos its downward. mótiøn !ohlnd. the plate, thus
forming a dopreBs ion wh ich is l.mito
sideways by two
wave tzains coling. from the sido edges o
tho plan.iug
surface,
Thwave tral.ne mGet.bohind the planing
sur-face in tha plane of symmetry, and .at 'the point, of int
tersect ion the water spouts up In the manner of a fountain.
The tan8voree pressure drop producoe a cross flow
which £or a sufficient speed. leadB to the.eepara.tion. of
the water also at the side otges.
The separation' :bog1n
in the forward. region of tho pr
suro area evéll
t t.he
smaller spoods.,. on account of the rlative1y high
pros-eur.e or pressure. drap that occiira there, and. with
tn-cz'easig speed continues toward the reax'
Te ide :spray
rises steeply near tesurfco and thòn spread.a out
I
NACA TI! No.
13
aenter of pressure. Zf the trai1irg edge of the
planing surface is chosn as .the. axle about which
mpmente are taken the moments on the planing surface,
according to Dlgure 7 aró
G i
4Z'i
Nnder the aotiói'of the app].ied inome
at the left
of the equation the p.lanlng suifaoö trims to an angle
a
or which the hy&rodrnainic moment at the right Is
equal to the applied moment.
A plot of the moment.
coef-ficient
mh*ahovethe vriatlon of the moment with
the. angle
a,
The greater the tr.m angle
a
the
smaller Is the increment
tNj
In the moment required.
to change the trim angle by an
.mount
a.
The ratio
Cmh
= i/b
Ca'
gives the position of the center of pressure If
A IBassumed to be approximately equal to
N.According to
the curve in figure 5
Lp/i = 0.77
with a tendency to
decrease as the angle a decreases.
The moment of the static lift determined according
to formula (3) Is
Matat
As-tat
.1
COBBO that the di8tanceof the center of pressure is
L1,
- Aetat.
or
T
0.333
The moment of the aerodynamic lift
> Aj.yj * i COB
a,u ii i
i i i. I Iiii .. 11 IiI I-14
N..LC.L.TIt !1o. l06l
and the d.istance of the. center .o, pressure > 0.666.
if
a
triangular pressure distribution with the maximum lift at the lead,in edge is apprximae].y aswned., sinceaccord.ing.tO
the measured presure distribution the center of pressure lies always ahead. of the center of pressureof a triangular. presv.re. d.istributibn tha lever arm of which 'trould be 2/3 ;- .
'rom equatiou (io) and. (ii) the:following may be said. about thecenter of preseurO travel: With
in-creaoin trim angle (A and. y being .conetant) the ratio of static to total lift Astat/. decreases and p/2
increases, as is confirmed. by the preliminary tests. Withincreasiñg speed (i/b and a. constant) the dynamic lift increases with the dynamic pressure while the
static lift remains ooatant. The ratio of the latter to the total lift therefore decreases and
correspond-ingly ip/& inoreasee.
(2) osults of tho Tests with Planing Surfaces
(a) Lift
The lift
Aby
analogy with the similar expressionused. in aerodynamics may be expressed. by the equation
A = a*HBI q
In tho case o the airfoil the lift coefficient ca.
in the nid.dle raige of the angle of attack and. for small aepoct ratios, is proportional to the angle of attack
and the d.orivative dea/da. = oon.stant if i/b = constant.
If, for the experimental range of angles, the same as-sumption is made for the planing surface (reference 7)
thon
c-
=Ca.a.
da.
A0A T14 No. 106].
15where the derivative Is replaced by
Inorder
oeliminate
a.as parameter, as would.be couvenentfor
a suitable representation of the lift coefftàlènt, it is
therefore written
0a. =
A(13)
-
The test carves of fIgure 8 show that for the region
of high 'roude numbers
the assumption is sufficiently
justified as a practical working hypothesis.
Deviatipns.
from the straight line'law may be ascribed to the gravity
effect as will be further clarified below.
If, in
fig-ure 9
cis plotted, as a function of the aspect ratio
/b
(reference 7), then from the foregoing considerations
the Froud.e number
wi].l be the parameter.
The
family of curves gives the results for all the lift
coef-fioiente öbtáined. from the 33 test series indicated on
the figure,
T-o bring oùt more cleàrly the scattering
of the points, since the plótting of ath test points
would. oomplicatè. the digram, fIgure 9a ahow the values
for the test series 2, 14, and.
2 to 33with
= 374
the constant value.
-Al]. valuesare well represented. by
the heavy averaging curve.
In figure 9h the lift coeffieient values
correspond-Ing to the tests of Sambraixê (reference 5) have been
pl'otted.
The wetted. L&ug
i
on the determinationof which
the a,cca.racy of
ca.main].- deèn-d.s had .to be obtained.
by photographic meawurement-s' in:. the Sambraus teat-ø'- since
thehigh epeed. carrt'age
.id not- permit direct
obsrva-tion.-. 8-mce the Z.örward. bound.aryof the pveeure area
at high dynamic preseur"ea and.- sm1'aas,-.euch as were
mostly assumed. by Bambrauo, fluctuated. greatly even for
an almost smooth water surface,..(.øae'Seo. II,. 4 of ref.-.
erence-5) directobservationa as made in thetets 14 tho
ESVA tank, in whib.the f1uotiatione maybe well averae&,
are more reliable.
The scattering, must be greater the
-shorter the wetted. length-that is, the smaller the
as-pect ratio o
the planing surfaos.
Taking theec
condi-tiona into consid.erdtlon.it is poBeib1e.t speak of a
reaoön.ble agrB'ement between the two serles of tests
with the exoopt ions of tests 2,
3, and. 4 the coefficients
.of which aro about l5percant higher.
he c.ef'icients..
16
:AaA. LH No. 1061
higer.on1y.in the smallest aepect rio.-range
/b.:0.7
to 1.2, areo for the medium aspect ratto range
Z/b
12 tá 2.5, ánd. becomé somewhat leso
t the maximum
'as-pect ratiG i/b
3.
In figure 9c there are similarly plotted the
coef-ficients of the testa byShoemaicer (reference 6).
In
the Aniorican teste the
atted length was obtained by
mirror observation of the spray coming off- at the sides.
This method, too, cannot lay claim to the same accuracy
as the determination -by direct observation of the
for--
ward. contour ofthe ressute area through a glass
tr-ip
insert ed in -the p].ate.
Gorresponding1
there is a con'
siderable scattering of the values for smal]. asp
ect-ratios
i/b <.1. the valueslyingon the average below
those of the ESTA measurments so that in this range
the latter values average approximately those of Sambraus
and- Shoenaker.
At ]arge aspect ratios the measurements
agreevorye1l ance in this range an error inthe
determination of the votted-length only slightly affects
the value of the lift coefficient.
-For Proude nuxnbera
8.5
the offèct of gravity
Is negligibly small and the lift coefficient
ca.&eterminod by the empirical euation
-C. = 0.845
(+)
-3/2
In theprevious section it was established
that-the fraction-of that-the -static displacement increases with
d.ecreaszg trim anglo or Increasing- aspect ratio.
The
ourvos jith &iffeent
parameter therefore d.lverga
in tho d-ireotion-Iñ increasing
-i/b.
--
-. As a limit- of the pure planing condition - that is,
when tho sHee of the planing surface begin to bo wotted.
at--the stopthere is obtained--a straight lino passing
through the family of curves
-.
--- 0.375
--= 0.94
Et)-
(if-q
'!)
yh1h at the samo t.me connecte the miiums-f the curvos,
(14)
NACA T1 No. .1061
1'?CompariBoL wjh t)e flat latC
etre.- The
-wiu.tiiune'te'±ì1ed. out by
i'nter tr.eference 6) onflat p1ate of various àspect ratlos povi4e a means of comparison between the 1aning e;faceand ai airfoil. In the table are given the coefflient of the total
--. :, CN COB ß
lift. cooffiofent in tjeir ca.tot as a
furct ion of - Z/b I for.d. = 1:0°. Jom the pressure
measu.rementÇfig. 23 nWinterIs..raper) therp is ob-tained. by integratIon,for
95O
the. presaure d.istributionon te topand bottom slds, the ratio of lift on the two sides Abot/Atop orAbOt. 0.38 Lt0t -
()
If for the angle-of-attack range considered tlie above re.ation in the absence of pressure d.istribu.tioi meas-urenent ct small angles of attack le
BBSIØd
onatantwith change in angle of attack and. aaect ratio, then tha.stnIght. line IÙ figure
9b-CcL
0.91)
(17).giveB the change in the lift oefficlent on the bottom sid.e of the flat rectangular plate in a.r.
ormulaa.
(14) and. (17) .d.iffer -on.y by the value of. the const.nt
and that to such-a email extent that the lift coeffi-cienta in air aud. water may b6-.sat& bo agree ae long ac the Dffoct of gravity is nogligible.
18 ACA 1UNo. 1061
i4 Rectangular
1st.e in kirComparieon with the theory of Wagner and. the re-suits of Sfiinbraus. On figure 9b lu plotted. the lift
coefficient, c according to the Wagner the.ory for the
short plate. The ourve, which Was obtained. by neglect-ing gravity, le to be compared. with the curve iron
formi.1a (14) for which the effect of gravity is negli-gibly small. In the range of email aspect ratios the
theoretical curve lles about 20 percent above that ob-tamed, from the test resulte where, however, lt le to be noted. that in the theory, on the assumption of in-finitely email angle of attack .&bOt hae been set equal
to 0.5 Atot; whereas, for finite angles of attack the
pressure ¡neasurements of Winter give
theratio of
formul-a (16). A com-plete. agreement between test and.
theoy is therefore not to be expected. n the
neigh-borhood. of 1/b = 2 the turvee Intersect and diverge with increasing aspect r.atio.
In the case of long planing surfaces (i/b 3)
there is good experimental agreement with the flat air-foil in air (under the assumption Cabot = 0,39
whereas the Wagner theory- of the short plate does not agree zith the test results also for planing cond.ltione
for which the effect of the earth's gravity is negligible.. -(The theory of the long plate gives greater deviations
from the experimental results as shown by Sambraus.)
P
Q.&Om
28 m/e,.a1OÓ
1/b Ca,tot .: 0.5 0,576
325
1.27 666' .520 '2.93 i . 14 .80 .468 2.64 1.03 1.0 .41.2 2.32 . Ì5 L.51? .338. 1.91 .745.2
300 1.69 .66 2 . 8.6 / 2'37 1.34 .522 7.48 .155 .875 34]..NACA
fl sc-.. 1Ó61
1g-Thoconclsions of S
byais on the long flat
plan-1±giirfaàs (5uimar'y pøtnts i t
3.ii -reference 5) thae
find. no coñf1rmatio
while poi
tcbe
e.--.---...mentad. by the statement that at high Proud.e numbers
'
the Lift in ,the inveetlgat.ed. range at oonBtant aspect
ratio vß.rios linearly with the
rirn ángle
iit tb.t with
decreasing
rotide number thé 1ifí coeffic1.eit as a
suit of the increaèing, favorable, effect' of gravity
increases so that
d.Ca*/&&bea not remain oonstnt.
b-)
eeistance
ProLi formu1, (8a) tere is. cbtaind. for- the friso.
tional coefficient
cf.t
R/
q.
When plotting
Cft
as a function of the Re'nOl&s number. R
vl/
arid,
corn-paring i:ith the resulte c-f pure friction meaauremens,
the fo11oring points -are to be noted.::
1. The normal resistance
Wobtained, by
subtract-in
f.ron the me.oure.
is, in the
rago of minimum resistance- d
about the sa3nO magnitude,
an4 'for higher trim añgies, -couaid.erably1ager than the
frctiona1 resistance
Since the scattering
con-tribution of the measurement on
Wis removed., the
scattoring of
Cft
must be relatively largo and. in
Oreases th' i..rxcreasing trim anglo.
11n the
resentetio
herochoßen -the Proud.e number
= v/Jg(A/)
±úproportionalto the epée&ad
in-versely proportS.ona]. to the sq,úare root of a d..ifl!
which increases with increasing lift; n&nel, the length
of anedfe of the water cube in correspondence with the
hysica]. interpretation..of the'Pr.oude number.
The number
chosen by Sambraun
Pv/.v-g.b
'which fulfills its
pur-pose as regard.e coneiderations of similarity, increases,
however, iiith d.ecreasing wi4th
b1a1so, for exampie.
when,be vetted, length remains .conatant with ß.eceaaiùg
load..
This representation is not suitable for the wórk
under ccnidi'on.
I
1he-
flg, 9b)
is used. for the Sambtau'teats, it,appear,e that- the latter
only- slightly exceed. the
Pbangt. o
our own tes'te. in
spite of .higher.test -s;ees on-account of the
applica-tion o± narrower p.lates,
- - . -.20 NLL TM No. -1061
- :. 2. The. acatteri.ng j.e izcreaeed. by
.tke.
!act that theme.8UrOd vala-of 4he vttpd qurf ace P-. !1ucuat.es more the -smaller the, ya].ue of 3 - that -te. t.he. i.argei the
angle , .. -.
- ...
3.. The xater thrown uj in front as spray parti4lly
vets the bo.tt eif ace and. d.eòxeaBea the iesietance
(reference 4)... TbJÏ thrust a.nnot'bø d.etermin.ed. and. is
thu uot .ako.n into acczun,t. .
The total pressure surface is used in the compu-tation as the wetted.surface in spite of the fact that
the .d.iroction of motion oftbe vater partiçies at the urfaco ahead. of the stagnation line is forward. and. in
the aid.o régfons haB a eid.e
component.-The high preaeurerogiois
uperpoeed. on the frictions]. boud.ary 1-ayer,vit.h the rqßult of ad.ecaaaeein speed. so that vm-<.-v, hia-d.ee.reaao in speed. ieot
taken into account.
The Reynold.s law asSumes geometrical similarity
óf- the. flov his requirement e not eat j.sf ted. since. it
is possible, that tha.wottüd. length J .and. hence also
R ma.be..conant for, .v,ar,ious..load.s; whereas the trim
angleand. honce the presau ánd. ve1ocit istri.button aro d.ifÍ'orent.
In figure 10 are plottea. the cur'es for the erips
of tests 2. to-33. It -may be seen that In--spito of 'the
above restriotions the valu.os. of
oarero].ati'veiy
veli ropresontod.. by ho puvo givon by. Prand.tl
(refer-ence 9) for the turbulent bdund.ary layer with reoed.ing
laminarlayor. . . -.
e = -, 0,455 -. 1700/R
(log (16)
if'it'.s taken' into accout that thecattering that ind.er1ios the'Praúd.tl curve ia-likewise not small.
The rn'éasuiementB vere partly cond.uoted. in completely
quiet vater: namely at: the beginning òf the tests an& after long intervals and. pütly in water yitli a alight
-amount of motiòn. The relative'y mal1' scattering le
RACA TN Ro. 1061
21eing].e stable form of.the boundary layer exista.
This
ie of paricula importanoe for -th-e -model t:est Binde
the larger portion of the towing test occurs within the
range of Beynold-ø ni1mbers 4tu which there might possibly
have been two boundary layer conditions and henoe
fluo-tuatione in the fribtiona1 relBtafl.CeB up to about 100
percent.
oreynold.s
mber
the surface tension
at mean pessura
:.=
20 - 50 kg/ms
(19)
results in the wett.ng of the sid.eo of .the planing
sur-face, thereby producing -a cQnsiderable Increase in the
resietanee..
Under these conditions spray no longer
oc-curs.
The condition for the occurrence-of 6he pure
planing process
s therefore, according to formula (15-),
f...
\.-o.375
c
<
an,d,:accoi'd.ing to formula (is),
Pm >20:- 50 kg/ms.
In the towing test, for
consider-able intervala d.uing the t-a]e-of f process,
p<20_-50kg/m
if the e cale of the model le made too small; then the
model resulta are no louger .traneferable to full scale.
The test series 31 to 33 belonging to a single scale
series lie In this range and partly also In the test
series 10 and 11, for-which, reason In the latter case
the
Cf1values risa unstead11y with decreasing mean
pressure (increasixig
and.s)..
That these values -cf
li-e in the range of the curve for turbulent
bound-ary layer le physically noi justified and. may be
con-sidered as an accidental resu3t.
(c)Ceterof'piaflure
The fun&axaental obseratione pteously made on
-
'the effect
f- the
athTsgrávlt
bn the position of
the center of pressure are confirmed in the family of
curves (fig. 11) which shows
as.-a function of
l/)i
with the Proude number
as parameter.
or Proud.e
numbers greater than 8
0.8 oonstant.
Both with
decreasing r and. with Increasing aspect ratio
(if
V
< 8)the center of pressure moves in the
22 NAD1 T14
o. 1Sl
(3) Application of ths ReSulta to the Determination
of the Sffact of the Beam, Load., Speed., and. Scale
in the Range of Pure Plating
In the following paragraphe a number of
i].uetr.-tive examples will be computed. asregarda lift,
resist-ance,and. center of preestre and. thereby a number of
important questione with regard. to the planing problem
will be clarified.
In the following diagrame therefore
it.. addition to the nwneri«ò.11y computed: carvee there
will also be lnd.icated the test pointe obtained from
measurement so that the extent of agreement of the families
of curves in, the wor]cing charte wi1
become clear.
There will first b.ö described. the proce-d.ure for
the numerical computation.
Let there be given tbewid.th
b
of a flat rectangular planing' surface, the lift
A,and. the planing speed.
y
and. hence aLeo the Proud.e
numbor
I
= 'v/.Jg (A/'Y)1'.'
1t is req.uirad to find, the
resistance-load ratio ratio
and. the moment
coeffi-cient
cmh*aafunctionsoí' the trim angle
a,
Oorre-sponding to a number of suitably chosen
i/b
values
there are determined from figure 9 for the parameter
1'
the lift coefficient
ca,..Prom the eq.uation
-arca-
Athere is obtained. the trim angle
a,
The.Reynolda number
is
C, i/b b3 q.
The corrosponding frictional coefficient
c'
is taken
from fi&ure 10, the curvefor turbulent bound.ary layer
with preceding laminar layer and.
y.
i' b2
E
=-R A b A
is determined,
Therefore
I
&A.L T1No
1061
2?urer,- fo' the intersection point of the
ccurve
and. ti b: iiit Ig'
'f
j1aning in f Igte 9
thèa are detrinined. the corrénponding apoot ratio and.
uimiting angle, réspectively, at which the character of
the f].-ow changes.
The d.ottea
ort Ions of the
i1lu.tra-tlon. curves apply tò con-d.it'ion.-e for vhicti the pure
planing procese has not yet been reached 80 that it may
be expeted. that the
eaeurement reeul-tTs ezceed. the
oteeiu1t on aòòunt of additional resistances.
3rôm f.gu±- 13.
/-
1A similarly detérmined. and.
with HhStpA
= -
-Xi.st
-.L
bCmb..
(j/y)1/3
-
b (A/Y)'/o
(a)
ffct of the Width
Por the interpretation of the teat results on floats
it is -of importance -to determine how., the resistanceicad.
ratio and. moment coefficient vary with the wtd.tb for
constant load. añd. planing speed..
-Pive flat rectangular planing surfaces of various
.widtis, in comparison with, the Initial test with
plan-lug surface
A,
were investigated. according 'to the
following test schedule:
figure 12. the reel tance and. mcixnen.t coeÍ'ficieit
and.'
cmh* '-computed. from .the,nieasured. values have
been plotted. as f.nctione. of
a.with
bas parameter.,
U ...
Planing
surface
Test
num
-ber
. b .. -.kg'
.. C-'vmJe
. --. :0.600
1ff0.0272
6 -3.74
- 2- 23
- 50Ò
18.0392
63'74
1].-- -24
. .400
18-.0612
6.3.74
j
2+25
.300
18.1090
6 - .3.74
326
.225
18.1940
374
-- ..,?? - 18 -435p
.374
24 NL0A T No. 1061', the cònt&n'uou cu.r"vee'
iYing
th c.oefflcients deteriainby compu,t.ation.' The ex'apJe-show .that !nth.e case of the f3t platø for which -the 'u1l wU.th contributes to
the 1i,fA the opt imui rßBiStancßlOe4 .rtio becomea. con-t-inu,ously more favorable with increasing w.d.th and.
de-creasing aspect ratio or wetted area,
In the range of small wid.the the' .impairmn çf the
rat.io due, to the ,increaeing effect. of gravity .ecr0ases.
tf the pure planing process discontinues, bowever,tha ratios again:1ncrea.ee.o account of the ad.d.ttional edge
resistances.
As limiting value of the width at the' steps of sea floats there is given ij figure 4 for the wide float
the value bt
..4 (GJ'Y)1'3 'and. forths narrow floatbBt = 0.7 (G'/Y)J. If for the 'tests under considera-tion, according to the assumptions of the preliminary test, the displaced weight at rest- is chosen as G' = 24
kilograms, then the beams corrèsponding to the beam at the steg in,.the moel are biarge =..O..4 meter and
ba11
'0,2 meter. For 'thoGe values there are obtained. the optimum load.resiat.nce ratioo o' 0.122and 0.154, r the difference amounting to 26' percent.he effect of he 'beam où the' trim anglo becoma
cter if the
.ng].&"is det'ermiuedas function of 'thebeam for constant cmh*. or t1e previously mentioned limiting beams there is obtained é. difference of about 5 percent.
he beam of the planin.g 'surface has a great effect on the in'tensity of the spray. Since the lift coeffi-dent decreases with decreasing aapet ratio; that is, with increasing beam, there, is a reduction in the d.is' placed volume of water and in thé 'wetted side length, at whih the spray ésc'apes. The axsnt by which the
spray formation may dife'r in the t'wa previously de-fined' limiting beams may be seen from figure 13 where the tro models are equally leaded and bavé t1ie same
speed and trim angle.' Figure 14 sh'ows, for the example given, the ratio of the static displacement V to the
tota]. lift V ?/A as a function of c and. it may be
sécu that for the two limiting caser the ieplaced. volume ci' iater' of the narrow flo't exceeds that of
'I
N&OA Tu No. 1061 25
in thIs connection the position of the d.epress ion behind, the planing surface is aleo of significance. In
the first ana. second-pa t.ofthe takeOff process of
"a eeap1aie the aten lies .behl.nd. the fount-a'i on' the:vater thus relieving the load.' on the for ebod.y and., on
accoi.i òf the Bmá].'l Lnit,a1 ti1m'of'the f.oat' resultiu
in a egaÌve trim bp' te atein,'there.. i even set 'up a
thrist that owere the resis'anoe(referénce 2).
he aitance
c' of'the fountain behind, therai1--ing êd.ge 'was meaauréd. in tests 22 to 27 aa well as in. a speoiaJ. test vith..p].aning.surface
A using a l°a
corresponding to the takeoff process. I. fIgure 15
the values + e are plotted, against the trim angle . It may be seen that + e,. is practically independent
of .... ao'that thepoeiti'on of the f Òuntain with change
in t±.im angle is Bhift ed. in hó same sense' as the c6n-tour of thewetted. surface.. The value of + o is
approximately proportional to:
V
and, b'1 + e 1..8Z* b (20)
as may be &erived from figure 151.
'(b) Eff'oct'Df Load.
It. will nov be determined how, for a given beam
and. speed, the resistanceload. rati,o and,, the moment
coefficient vary with the loá.d.
Por constant seed. and.' a large ra'úge o loa&s,
tèsta were conducted. on planIng surface i accord.Ing
to the following test achedu1e:
expression given by Wetnig (reference 12) of the
fountain distano does not agree with the results obtained by the author and. not published, at the time and is
there-fore not taken into account here. Planing áurfaca Tßet n-u.m ber
b'.
.'kg'
...
' . "rn/s -18 '' 0.3' iO 0.0218 10 6.90 - A 19 .3. . .5 ''.0545 10 5.90 -. '16 .3 0 ".lÓ9'' .10
. 20 .3 100 .218 10 4.70 21 .3 150 .327 10 4.3826
r :'
NACA TM No, 106].
_
Ia. -figure l6.temeasued.-valu-os; ' and.-.
cjare
p1ottod. aainet '
wnitlib5
pametbr aid. the
theprfca1ly compued, ci.ves are a,lso 811CWL.
Imay be
seen that
or sivalj. and. averge load,e tb
reeistancç..-load, ratio inoreasen coxaidezaly witi ipereaoing load.
on account bf the ipairment of the aspect ratio.
In
&oubl1ug the -load.1.for eamp1e1.frorn
5 to 50.. kilograms,
opt.:
tzcreases. by 25. percent .
&t. very ]a.rgo loa.s the
impairment' is slight on account.of.theiu
easing-of-fect o' gravity.
-i
"Choc2 'cf the re1tibn
ca/&
cOn'stant
in. the
investigátéd, range for higher
nujnbers.- In figure
for toste 16,' 18, end. 19 the wetted. Iength-
1have
been plotted. again.-et
and. emade to intoÌ-eeot the
straight linos
0.3, 0.6, and. 0.9 metbr,
corrre.
sponing to the aspect ratios
/b = 1, 2, 3.
The
0a
values are also plotted.. 'Tbese lie on straight linos
,which pase through the origin - that is, in the range of
numbors,within whiçh tts.effect -of gravity may be
entirely or approximatoly neg1cted.
d.ca*/d.r
= constant
and. thus the assumption iin.ß.r1ying forni1a (12) is
justified..
(c) Eff
tof.th
speed
-The t&to--off d4agra.in of a seaplane showe two
ro-sistanco znaximune, the f iret in the sango of the transi-.
tion frox the floating to the planing cond.tion, thcr
second. boforo the get-ava.
It will now bo itvostigated.
to what oxtont the formatipn of, these maximwns depend.
on the stern of the float.
This qiestion may bo
an-everód. with the aid. of the planing siirf'aco which is
equiva1ont to the longitudinally straight forobody;
a larger speed range with conßtant
beam and.
load. the rosiatanee and moment coeffiionte wore con-.
Plotting c and. cnih i.e functions of a. with
V as IDaranleter, figu.re 1?, there le found the type of
relation familiar from float iuve8tlgations. Lt low
speeds c increases approximately linearly with a.,
the puro planing procese not yet occurring. With in-creasinC speed the resistance increases, particularly
o at sLlall trim angles so that a maximum resistance is Boon obtained, at moderate trim angles. After exceed-ing the maximum the resistance 'again decreaeos with In-creasing speed. A second. maximum therefore does not appear In the case of the planing surface and. the ap-peararice of auch a maximum for the float Is to be
as:crlbod. to the wetting of the stern by the spray. In
the plot, figure 18, g1vig c and.
ch
as functionsof Y. the formation ot' the first maximum is brought but.Titb particular cleariese. The indicated boundary curveof the pure planing condition:shows that the
latter occurs before the maximum is reached. Resistance snd. moment maximums lie, as aleo In the case of float tnvestigutione, at about the same speeds, so that the greatest trim angle coincides with the resistance maxi-mum if the p].anIngsurf.ce planes free to trim.
llheroas, with decreased. speed. the résistance-load
ratio increases on account oÍ the impairm.ønt of t,ho aspect ratio -and on noglecting gravity there would be obt'a1'nod the up ercu.e
te affect of
gravity in the lower speed range Is to de'crea'sè the re-eista.nco to such an. extent that with tho second curva branch a rosietance maximum occurs. By support of the
stern tho maximum in a float may bo considerably affected as explained, in reference 2. Planing surface bm
Ag
-°BA
0.3.50
0.436.
5 2.63 2.92 31 .3ß0
.302' 6 3.16 3.5 31 .. .3 . 50 .222 .7 . 3.69 4.09 31 . .3 .50. . .170 8 4.21 4.67 31 .3 50 .134 9 4.74 5.28 A. . . 50 . .109 - 10 5.265.4
A .3 50 .075? 12 6.32 7.00 31 .3 .50- .0556 3.4 7.38 8.18 31 .3 50 .0:424- 16 8.449-5
f,, NAOA.TH N6. 1061 27In figure 20 the values of ,
/b, and.obtainod. from the measured values are plotted. aan.pt
a. uith X as parameter. Planing surface i. within
b = 0.600 meter is the widest of all surfaces
investi-gated.. For the scale comparison the fullscale design was c flying boat of beam b = 2.400 meters with A
= 9216 kilograms and. G = 12,288 kilograms p = i).
The planing surface with b = 1.200 meters and A = 1152 kilograms, G' = 1536X2= 3072 kilgrane correspond.e to
the design o' a normal twinfloat seaplane (X5 = i).
Plan
ingsur
face Toat number 'ull .Pul].-acab flying boat scalo bwin float e ea-plane bm CB Akg Xp 1 2.400 0.109 9216 3,74 16.96 2 1 1.200 .109 1152 3.74 12 1 28 4 2 0.600 .109 144 3.74 8.48 2L 2,25,29 8 4 .300 .109 18 3.74 6.00 3 30 .0.66 5.33 .225 .109 7,6 3.74 5.20 4 31 16 8 .150 ,109 2.25 3.74 4.2 5 32 24 12 .100 .109 660 3.74 3.46 6 33 32 16 .075 .1.09 .281 3.74 3.00 28 b:oA .Th Cd.) Xffect of the-ScaleIt ji1l now be investigated. to what extent
simi-larity a:pDliee a regar.e the pressure eurface8 and. centerofpressure position and. whether tho scale ef-fect is given solely-by the dependence o tho frictional
ooeficiont on the eyno1d. number. Byinvestigatin.g a family Of planing surfaces thi,s question may be
an-s'vereci for the Case of the-pure planing process.
Sixplane rectangular plan.ng surfaces of various
beams îcro investigated. accord.ing to the. following test schedule:
RACA III Ro.'. 1061
The. contixiuous curves show .the ar1at ion of the values obtained, by coinputat ion vhic were' determined.
for all conditions for which p
>
50kg/m2.
At smaller preseures at which, ¿n account f the effect of surfacctension', the pure planing process has diecontin.ued. the d.8tetinat ion of the restet'anoe 'is not poas
ible.'
In this
case the dotted curves ahoy the 'ai.ation in the maas
ured.-va1uea. '
Thó
values d.etermined. in
theséries
of testsfor
the aspect ratio
/b
and. the moment coefficient
c,ht
arrange these14"as i
euch a manner t.at no regular
deviation, rom .an avèraging cur' through the, test points (thia cui-v,e has. not
béend.rayn
one
diagram)
can beestablished. -
that int'ha pure planIng prqceea there
is conplet'e siini1.aitr of wetted. sfaoe
and. momnts and
hence also of pressure ditr.b'jtion,. The sim±larity still -holds in"bhoaa'ranea i which the urface tension exerts an effect on the planing condition.The resistancéload' iatios become more favorable with increasing size o!. planing surface on account of the decreasing-resistance coefficient, a shift in
opt
tot-ar,d email angles taking place. On acco.unt of the
fact that
ct
ia' approximately constant :.1n the rangeB i
5X106
the ratios ar practica.1y equal foaverage scales. With further decrease the resistances
strongly Increase o-n account of the eff@ct of the
s'u.r-face t;Bion, with' Opt
gkifting,,c.onaid.erabr'.in the
direction
ô!
higher angi,es,..
-T]ieimpairme't
for eq,ual
Cmh,
i/b, Ei!1d.'for, the flying boat and. twin float seaplane is given"
in the following table:
r. 'g boat
400
ni win floa b'= 1.!lyin
b.= 2.
2 '7.2'
'4
122
se,aplaie
200
merc.nt.
''2'
3.8'
'4l0.5
-5.33'
21,8
8 32; 1216
107
18.5.: 10.66 30.6 '16 4224
90
.30
'&CA TU No. 106L
The above Ligured Ehow that teats .wItb. mod.eli oftoo sall' ascale
(b < o.&) auch as were mainly
conducteà. in Exgland.,for eùmp2e,'(referoncie io) and. In It.&iy, (reference 11) since the, beginring of flo,at in-vestigations,, cannot: be used. or the. determination ofthe. resietance, aince the additional resistances arie ir& under the effect af .he sirfaoe tenion canot .be de.
ternined., Also in the case of somewhat larger .mod.ele the difference in the ratios between model and. full -scale design is s141]. considerable.. L transfer of the -od.el test reeult according to the method. here given to the fullscale design is practically impossible on account of thd.1ffidu1ty d.ur.ng the teät i'n
d.etermin-Ing the vetted. ares b moasuremeüt,' rt la 'only after
the construction oVtowIn tanks with high carriage velocitlés that têsta ön large so.Ìe models became
posible. for which, even with no ac'couñt being takénof
the
cale effect - that is,vith te ésumption:
W = w X3 or E CM
and. or emhE = (21)
results are, obtained that deviated from the true values only ulthin the accuracuival 'In passing from the model to fullacale computatioLna.
The ebove example shows that in passing from the
model to.the fullscale de.sign..accord.ing .to formula (21)
one je on the safe side. On account of the additional
rouhnese and, wave effect (ref erenc'e 13') which occurS
part icularly in the case of riveted sheet bottom of the fu11cale design the actual differences' bétween the
resistance computed. from the rnodel.and. the true one are
smaller than given In the prevlou8 table as is ehown by scale tests conducted on families of floats (references
l an'd.' 15).1
1The-possibili'ty should. be noted that In th model of very large scale the dIffeience of the fric.tion
coeffi-cien.t of the model and ful].ecale. design la so small that on account of the roughness' 'and. wave effect te
re-sistance o the f.].lscale design. may be higher than that obtained. from The model com'putatipn but does not
come Into c'onstd.erat&on in the case of the usual scales
NAOA. Tu ITo,- 1061
(B) The VBottom P1antn Surface
(i) aslc Equatl.one and, Determination of the Lift, Resistance, and. Center of Pressure
or a longitud4nally Gtralght Vbottom planing
sur-face the reaultaxi:t brma]. force on one side is acoord.ing
to figure 21
2 2 cos
or the tota]. normal force
A
Ca =
COB ¿'
The lateral componente Q balance out.
The production of the lateral speeds corresponding to the lateral components Q for equal lift and. other-wiae equal conditions, particularly equal effective beam results lxi an additional exit loss and. hence an increased. resistance uhich corresponds approximately to an increase in the load, and. la equivalent to a lilt increase from
A to Ii = Alcoa . That this actually represents a
usefu.l approximation is also shown by the analysis of the further tests. There -is therefore set
N.
= Aa COB. ¿
h.-)
(23)
3'or the length there is taken the mean length 2
of the ed.geshape surface 7 the forward bounding line of'the preseure-aroa --in -oorre-pondence to teat
measurements being taken as a straight line, and. e.
atagnat.oa of ccnatant hight along this line being
32 GÂ.T! No. 1031
and. at the outer edge of the planing surface
- btafl
4 tan a
Tho full beam of the Vbottom planing surface contributee to tio support as long as
' o. or and therefore If
btan
4tanci
tane
b4tanct
tane
b -
4 tan>_rn
4tana
bthe natural beam bnat of the pressure surface le boloti that of the beam b of the planing surface and.
tane bnt - 4 tan a.
(25)
(26).
(27)
In contrast to the flat surface for which, with increased. beam of the planing surface the aspect ratio o the
proa--sure area also becomes more favorable, since the trail-ing odo of the plantrail-ing surface Is always utilized. to ita
b b bnat (24)
Prom £igue 22 the length of the pressure area a!t
the keo]., equa]. to the maximlim wetted. length, is
CA Tillo. 3.061
33
full extent,
in the case of theV-bottom
surface therea
imitig"baaw
bnat of thepressur-e areavhioh.i
.t
the samp jme the most favorable beam of the planing surface for the given load relations. Theportions of the planing surfce lyingou.teide of the pressure area aro wètt,ed bythe spray which is rotar.ed.
on the surface and. so increa.ses the resistance. They aro therefore, without being utilized. for tho support.
of tho surface, the cause oZ ad.d.itional resistances which aro larger themdre..the width of the planing sur-face exceeds the natural wiath of the pressure area.
In the case othe a:eaplane float this condition occurs according to figu.r.e 23 before the got-away. Since in the outer free bottom strip of width (b
-a pozorfu]. spr-ay is directed. b-ackw-ard., the stern of onefloat is aleo wettod. and. under certain conditions such largo additional resistance may be set up that in epito of the small residual loading of planing bottom the total resistance of the aircraft attains the valio of the pro-pollai- thrust and. take-off is impossible.
The
assumptions
required. for the numericaldetermi-nation of the friction coefficient ori the size of the
wetted. eurface $' and. the Reynolds number R obtainod on tuo basis of test. observations are the following figuro 23a.
Cceo i) loaded beam: b 1f
Caso 2) loaded beam; b bnat if
co s !
tane
<IR
4tana.
btane
4 tan a. -. b I 1m tanòRi'--, whòre. -=-+
.(28).
D. . b4ta31
aae- a)Gad..ed. b.eaz b
21
and.
tanY=
1034
NAOAN ITo. 1061
iero is then abaine
and
sothat
Y fr u (b + bnat) -2 cos I 1Caeo 4) load.od. beam: bnat if tan
4tan
band. tan?
<10
bbnat
(bnat+
(so) cos 'j uIn casos i and. 2 the water escaping from the forward oontoir of the wetted surface and the water running along the surface as a film covers almost the entire width of the surface so that is to be used. in the determination
of P and B. In case 3 the limitj of the spray water wetting is given approximately by t1e dotted curve. In
caBe 4 no further wetting of the edge portions occurs with further increase inwi&thof the planing surface. The limiting valu.e is for tan ? 10
(29) bnat i A
J C
COS ¿ q.a bnat 'n a tt.
In figure'24 the computed coefficients
and
are plotted. against
with
as parameter;
the values of
Cah'
being gi'ven by the dotted. curvas.
The computed. curves are..contlnuous.
Very. goo1aroement
io ohaurn between the computed and. test. values, of. c,
the, as,sumtion lying at th.e..bsis .of 'foru1a (23) thus
being .eonftrrned..
0.ly .i. thcase of th'e sharpest
Vbotton &o
the test 'a1ues. at 5mall .trim angleB. shöw
any &eyiation. from
ho'genera1 shape of th.e,curve
since
in this case the Btatioa,lly d.1eplace&.vo1u.me.f water is
practically equal to the entire lift so that in spite
of free side edges the process is more like that of
f.oating than that eI. planin.g.
CorreBpo1dingJ.y.the
pres-sure .drop at th9 sides ii small, and. heñee the.spray.
for ati&n'.weak.. The .ress,tn i-e- stroug1y keeled and
there is formed. a short distance behind. the planing
sur-face a low fountain extending over some distance.
The
spray development then increases up to a V angle of
beS-t%Ieen 1000 and. 132°.
With fwrther increase in the angle,
howevor, there is a strong decrease in the Spray.
Planing
arface
n'u.m-Teat
ber
A C3 Vm/a A2&25
1800
180..09.
63.74
.734
i.60°'
- 18 -.109
- .3.74
I50
.8.109
63.74
36132°
18.109
6 .374
10
..37
100°
.18
.109
63.74
NACA TH No. 105).
:35(2) .App4qation of the EeBU1tB for the Determ.ination of
(a) The Effect of the V angle
-1heeas, the impact forces in take-off and. land.ing
decreiae with thcrease ±n 'V. angle of the f]oat'. the
re-aistance increases.
The d.epend.ence of the resistance on
the V angle will now be found..
Four rectangular planing surfaces of various V
angles and. constant beam were investigated. according to
e
-36 HACA M No. 1063.
The añgle of in1iimum resistance apt increases with increasing V angle. The ConflOOtixLç-Btr'aight line
opt
has according' to the aesumption made iñ the compu-tation the same trend as on increasing the load, of a flat surface (fig. 16). Por .accord.ing to figure24 thore is an' increase in the drag with V angle given by
100 1 + 40
+ 622
(31)0
-- In the formula for the flat plate'if. i is
re-placoci. by . there is obtained
Cmht =
b (A/'Y)1"3
The agreement of the computed and experimentally .etprnined values of Cm is very gaod. up to medium
V ancles; at larger angles for constant Cmb* the
difference remains less than'1° so that heie tQo. a sufficiently accurate determination of the trim angle is possible.
(b) Zffect of the Beam
In the takeoff of a seaplane the wi'd.th of step of the Vbottom float, on account of the small residual load and. the high dynamic pressure exceeds the natural width of the bottom surface under pressure, in contrast to the flat bottom for which the full beam at step in
the pressure area is utilized. The most fAvorable width of tho bottom depends therefore on the load and. the
d.ynaLic pressure.
'or the purpose of testing these conc.lusions plan-ing surfaces with two Vbottom anglos and. various beams were tosted. according to the following test schedule:
t
In figuree 25 to 28 the measured. values of are
connected. by d.otted. curves and. cmh* plot-tea against .
The continuous curvos give the results of the computed.
valuas.. The differences between the two aro indicated.
by arrojs. Since beyond tha limiting curve, (fig. 9)
no nro planing condition is set up, the computed curves
aro also dotted. - Good agreement is obtained between the planing process determined. from the diagram and. that from
the observed appearance of the free planing condition. In a comparison of the values of -e for pure plan-ing conditions it is found. thai; the test values are
always somewhat below the computed values but otherwiao the curves are similar. This is probably to be as-cribod. to the fact that the mean flow direction on the ad.ditiönal wetted areas does not agree with the d.iroo.-tian, of travel so that the resistance -isincreaaed. 'ozt].y by a component of the additional frictional resistance. In spite of this for most purposes the numerically com-'
e6value* ara sufficient
(greae
ff e iceLc = 0.025).
An optimum beam which is obtained. both oxperimon-tally and. by computation in the above example lies botwoon the two mean beams, the experimental valuo being somewhat higher than the computed. value.
III U