Escolar Documentos
Profissional Documentos
Cultura Documentos
-va-S7Zw 7/,40
MEMORANDUM
REPORT A UFAIRS.tV
PUBLIC E
Ohio
WRGT I DEVELOPMET CENTE
I.'. -1
WRIGHT AI DEVElOPMETCNTE
Reproduced From
CLASSIFICATION: UnclassifiedAvailable Copy Best
Lloyd E. Schumacher
PUBLICATION REVIUW
Eo M. GAVIN
Colonel, USAF
Chief, Flight Test Division
WADC-TR-52-71 iii
TABLE OF CONTENTS
Page
Introduction 1
Conclus ions 17
Appendix I 18
Derivation of Longitudinal Dynamic Response
Equations
Table 1 23
Calculation and Summation of Scalar Products for
the Normalized Equations
Table 2 30
Method of Solving Normalized Least-Squares Equations
With Four Unknowns
Table 3 31
Sample Solution of Normalized Equations for
Stability Derivatives
Table 4 32
Sample Solution for the Pitching Velocity Transfer
Function Constants
Table 5 33
Sample Solution for Angle of Attack Transfer
Function Constants
Table 6 34
Sample Solution for Normal Acceleration Transfer
Function Constants
Table 7 35
Transfer Function Constants Computed from Flight Data
Table 8 36
Transfer Function Constants from Flight Test Data
References 46
WADC-TR-52-71 iv
LIST OF ILLUSTRATIONS
WADC-TR-52-71 v
INTRODUCTION
The method described in this report adopts the approach of activating all
of the variables of the aircraft motion simultaneously while retaining the oper,-
ating conditions as constant as possible. Values of the stability parameters
are then deduced in accordance with the theory of probability0 It is necessary
to use an equation describing the relationship between the response and the
stability parameters. This relationship can be expressed in the form of equa,-•
tions of motiQn or as transfer function equations0
WADCTTR-52-71 1,
Two possible techniques for exciting the variables of the aircraft motion
under constant flight conditions are the transient method and the sinusoidal
oscillation method. In the transient method the aircraft is disturbed from
a steady-state condition by a step or pulse type motion of the aircraft con-
trol surface. The oscillation method consists of obtaining the aircraft
motion by a sinusoidal variation of the control surface about a trim point
at various frequencies.
The transient method requires a minimum of flight time and has the fur-
ther advantage of being applicable to flight conditions other than straight
and level, such as climbs, descents, or turns. However, the evaluation of
the variables of motion and all required derivatives with sufficient accuracy
to permit solving for the stability parameters does not appear feasible in
the time plane. Ir the above example of the transfer function constants for
the rate of pitch, 6 can be measured directly. If the time history of the
rate of pitch is quite smooths the second derivative of the pitch rate can be
obtained with fair accuracy by standard numerical or graphical methods. How-
ever, if the third derivative is then obtained similarly from the second, the
inaccuracies are no longer tolerable. For those elastic aircraft having bend-
ing modes that appear on the record in addition to basic airframe response,
even the first derivative that is obtained by graphical or numerical methods is
of doubtful value.
The aircraft responses obtained by the two methods above can be related
analytically by Fourier or Laplace Transformation theory. This provides a
means of taking advantage of the best features of each technique. The flight
test time can be minimized by recording measured responses in the time plane
and transforming these responses to the frequency plane to take advantage of
the simplified analyses procedures available. The transform theories assume
that the principle of superposition applies, that is, that all components of
the motion are additive and that the system can be represented by a set of
linear differential equations with constant coefficients. Therefore, any non-
linearities existing in the response characteristics can, possibly, lead to
prohibitive errors in the transformation from the time plane to the frequency
plane. The validity of the transformation can be checked by observing the
agreement in several frequency response characteristics obtained by analyzing
transients having appreciable differences in frequency content. A numerical
method for accurately obtaining the Fourier transforms of arbitrary time
histories is given in Reference lo
WADC-TR-52-71 2
SYMBOLS AND NOTATION
m/iVS, second
0 phase angle
dC/cc rate of change of downwash angle at the tail with angle of attack,
4 determinant of coefficients
2
CL lift coefficient , 2mg/f V S
WADC-TR-52=71 3
C rate of change of moment coefficient with rate of change of
angle of attack
WADC-TR-52- 71
1H EDRY
This equation relates the lift forces to the vertical acceleration at any
instant of time, irrespective as to whether the motion is any arbitrary
transient or a steady-state sinusoidal oscillation. In the latter case, the
variables can be considered for any given frequency to be constant amplitude
vectors rotating in the complex plane with fixed phase relationships0 These
amplitudes and phase angles can be experimentally determined for a number of
different frequencies at a constant flight configuration0
If these data were obtained without error of any form and the above equa-,
tion described the aerodynamic and inertial forces completely, it would then
be possible to solve for the unknown derivatives directly. The variables
could be resolved into components along the real axis for as many frequencies
as there are unknowns. A simultaneous solution of these real equations would
yield the accurate values of the derivatives. Since neither of these assump-•'
tions is true, this approach leads to large errors in practice0 Random
errors may be present that are larger than the effect of some derivatives at
the particular time in the cycle selected for taking real parts, The effect of
a neglected stability derivative will also be large when its associated vari-
able has relatively large components along the real axis as compared to some
included variables. Standard curve fitting theory for scalars can be extended
to minimize these errors when the variables are rotating vectors.
v = residual
Since all terms on the left are vectors, the residual, vi, will also be a
vector at each frequency0 It is now required to find values of the de-
rivatives that will make the amplitudes of these residuals small. The
WADC-TR-52-71 5
relative phase angles of these residuals that result from minimizing these
amplitudes are of no concern at this point. Later 9 it will be shown that
these phase angles can be used as an indication of the validity of the
equation being fitted.
It can be shown (Reference 2) that the most probable values of the un-
knowns will be such as to make the sum of the squares of the residuals a
minimum. The square of a residual vector is a scalar equal in magnitude toq
the square of the amplitude of the vector,
= +2 + V 2 + 0 0 0 V 2
+ CL i + CLni) 2 (3)
=+(c
ff(C CL) 1, 2I,3, o m
The condition that this function be a minimum requires that its partial
derivatives with respect to the unknowns be zero0
+ 0 ° C~m•m + CL2(Cm = 0
or C i2+ C di - W-7CLnio. ic i 1, 2, 0 0 0 m. (5)
Similarly, when
af =0
4)cL (6)
C 1 ii+ CL2,no, i 1, 2, 3, 0 0
Equations 5 and 6 are the normalized equations for the least-squares solution
for two unknown coefficients of variable vectors, The indicated products of
the various pairs of vectors are defined as the scalar or dot products
(Reference 3). The scalar product of two vectors is numerically equal to the
product of their amplitudes times the cosine of the angle between the two
vectors0 This is equivalent to multiplying the vector of the variable being
normalized by the respective components of the other vectors along this
normalized vector.ý In this way each individual equation is normalized for
each unknown parameter at that unique angle at which the associated vector is
at its peak or maximum value0 It is apparent that each parameter will have
the maximum effect on the response when its vector is at its peak value0 It is,
therefore, logical that the parameter should be so selected that it yields a
minimum error at the point where it has maximum effect. Basically9 in the
equations given above, this amounts to resolving the vectors into real parts
with the angle-of-attack vector lying on the real axis when normalizing the
equation for CT and then resolving the vectors into real parts again with the
elevator.deflection vector lying along the real axis when normalizing the
equations for CL 0o
WAPC-TR-52-71 6
Using the definition given for the scalar products the normalized equations
can be simplified.
Q&6 olIcos %
Similarly, the remaining scalar products can be found in terms of real and
imaginary components of the complex variables. Equations 5 and 6 become
2C + 2 + C A
AAiA + BCB() CL(+BAn + BnB n (9)
Li X
hiii
2 2(A:iL An + •B B (10)
A +Bo B) + (i2 +CBi 2 )z ECL( n(
The amplitudes of all the vectors are usually given in terms of a unit input, and
the phase angles are all referred to the input vector which, in this case, is the
elevator deflection° Therefore, the elevator deflection vector will have a unit
length lying along the real axis in every case, Then A 1 and B = 0 and
the normalized equations become i i
(•.+•)+ C AiAn +BBni) (11)
Lk iY nii i
The original vector equations can also be separated into real and imaginary
parts and these scalar equations normalized by the conventional least-squares
method. When all the real and imaginary equations are summed in the normalizing
process, the same equations as given above will result. This approach, however,
makes it difficult to visualize the basic validity and effectiveness of the
method,
The least-squares method for vectors can now be applied to obtain values
for all of the moment and lift derivatives included in the equations of
motion in Appendix I. It is first necessary to resolve the linear dependency
that exists between the variables of motion, as shown in the derivation,
WADC-TR-52-71 7
before normalizing the eq-ationso In the simplified theoretical evaluation
of the stability derivatives, it is found that the ci and 6 derivatives are
directly related by a constant equal to the rate of change in downwash angle
at the tail with change in aircraft angle of attack (Reference 4). This can
be expressed algebraically as
c KCM (13)
C X KcL 4 (1 4)
dE
where K = d, the change in downwash angle at the tail with change in
aircraft angle of attack.
Substitution of equations 15 and ]1
4 into equations 7 and 9 of Appendix
I will reduce the equations to three linearly independent variables.
CM+ C + Cm (6 + F = h6 (15)
a @r
[ (+C+CLn = or h6 (17)
c
C•; J = morL
1
The normalized equations for m data points can be obtained directly by scalar
multiplication of each equation of motion by each element in the matrix of
variables in turn and summing the results.
or
or
:T
I
WALn c-
10 1
j - L [hi5 (9 ++ j W_M
ýE CLn (9 ++
WADC-TR-52-71
Expanding and simplifying the scalar products and letting 4- AC + iBC
1, the normalized equations become
(19)
1:-CL(A An + Vn)
E-CLAn
[ ~ RcA 4)1
or i- %
B,
L 2 hK(A Aý + kxBZ)J
&A =m
Similar solutions can be made for obtaining values for the transfer
function constants given in equations 12, 13 and 21. Appendix I. The
normalized equations for the constants in the equation for 6/1d will be de-
rived first. Rewriting equation 12, Appendix I, and transposing terms to
eliminate fractions gives
0 2
AO0 + A1D6 + A2D = BO0 + BDdo (20)
The normalized equations for m data points can be written directly by the
method given previously.
WADC-TR-52-71 9
2 1!
1. -
a-sJoiWQ
• iA w-40iW6 C-iuJ)
j-•,,(%) 2:Q
a(o)2
(-iL4) A1i
Ao (l
L~J 1o-
d .-
•,2Ao6 ,
Expanding the scalar products under the summation symbols and letting
J= A + iBd = 1, the normalized equations become
ZWB;
m
0 ~
,V- Ai
6
£-A
2
0
2
Ae
B0
E.B§
;rh9
_W
2 A;
-4
2
62
~~~
i-A6
0
62
L1k6II(4
So
0
ZVAo
[0 0
,, _ o262
Similarly, the normalized equations for the constants in the angle of
attack and normal acceleration transfer functions are found to be
m 0 Z( BMS -- Ax C(25)
0 2 k t.2 I Bl 6 Ig) B
:9uj%, £ 4Ocx 0 A1 0
ý-An X n X n2 0 X4 2 An 10
:E jB :1E-?An 0 X w2n 2 E _10B A!
.E=JBn
2n2
0
An
WADC-TR-52-71 i0
APPLICATION OF METHOD AND DISCUSSION OF RESULTS
The least-squares method for vectors was applied to actual flight test
longitudinal response data to obtain values for the stability parameters. The
test aircraft was a twin-engine light bomber-type B-25J and is pictured in
Figure 1, Appendix I. A complete set of frequency response characteristics
for the air speed, altitude, and center of gravity range of this airplane has
been measured in flight. These measured responses are intended to provide a
store of flight-test data for the development of data reduction methods and
to investigate the possibility of extrapolating the results of limited test
programs. The test methods and instrumentation are described in Reference 5.
Table 1 also presents a procedure for obtaining all of the scalar products
and sums of products required for the normalized least-squares equations for
the lift and moment derivatives and the pitching velocity, normal acceleration,
and angle-of-attack transfer function constants. When only the stability
derivatives are required, this table can be redued from 5) columns to 40.
WADC-TR•-52=71 11
of intermediate steps are retained in the machine, thereby eliminating writing
down the results of each step.
Theoretically, the lift and moment resulting from elevator deflection are
related as follows (Reference )4)-
CM it X CL/C = 2o57 CL
where C - the mean aerodynamic chord = 9.69 feet for the B-25J
It can be seen that the results in Table 3 where Cm =lo_1..l and CL( 0O556
are in good agreement with this equation0 The same relationship should hold
for the pitching velocity derivatives, however, the agreement is not as good
here for Cm,4 = -0o270 and CL, = Ooltlo
As pointed out in the previous section, the effect of the rate of change
of the angle of attack is included by an assumed linear relationship with the
rate of pitch derivatives. The derivativeP CLY is so small that it can be
neglected without appreciable effects on the ýilues for the remaining lift
WADC-TR-52-71 12
derivatives. This was verified by actually carrying out the solution and ob-
taining values of 5.061, 00526, and 0o194 for C C-L, and CL. respectivel-y0
When the lift caused by pitching velocity is not included, the value
found for CL is considerably reduced (0.287), although the relative effect
on C is miuSh less (5.206). The value for the slope of the lift curve when
the angle of attack is the only variable included , is 5.176.
One of the reasons that the value found for 0C is only slightly affected
by neglecting the smaller lift derivatives is that~fhe sums of the products
in the normalized equations are predominately composed of the responses in the
low frequency range, At these frequencies, the angle of attack effect is large.
whereas the amount of lift from the elevator deflection and the rate of pitch
are approximately equal and 1•0 degrees out of phase, thereby having little
net effect. At higher frequencies, these latter derivatives become increasingly
important as their associated variables are relatively large and their relative
phase angle changes,
The angle of attack stability derivatives show consistent trends which can
be attributed to the angle of attack acceleration derivatives . The variables
a and c are vectors along the same line but 190 degrees out of phase. The re=
sponse caused by the & stability derivatives will either increase or decrease
the effective value of the •c stability derivatives depending on the relative
signs. The remaining parameters are surprisingly constant throughout the
frequency range, considering that the theory of unsteady flow indicates varia-
tions of these derivatives with frequency0 . The scatter in the low frequency
range is believed to be caused by errors in the data resulting from slight tur-
bulence, although variation in the speed of the aircraft during the oscillation
cycle may have contributed also0
The normalized lift equations for these limited frequency ranges were also
solved neglecting first the CLo and C derivatives and then also the CL deriva-
tive. The results are plotted in Fig~e 6o The importance of including these
latter derivatives in the solution at high frequencies is clearly shown in this
graph0 Values found for the slope of the lift curve are more than double the
theoretically expected size in the high frequency range when these derivatives
are neglected,
WADC0TR-52,7V1 13
motion to the original data can be examined by plotting the phase angles and
amplitudes of the residual vectors for all frequencies. If the equations of
motion include all of the forces that act on the aircraft, the residuals should
consist of small vectors with random phase angles caused by the normal dis=
tribution of errors in measurement. If consistent trends do appear in the plots
of the phase angles of the residuals versus frequency, it should be possible to
find another stability parameter for the equation of motion. This parameter
should be nearly in phase with the residuals. This effect is demonstrated tn
Figures 7 a and 7 b, where the phase and amplitudes of the residuals are plotted
for several lift equations. The lift derivatives are obtained from Table 5o
The top graphs in Figures 7 a and 7 b are the amplitudes and phase angles
of the residuals, respectively, when only the lift from angle of attack is con-
sidered. The amplitudes are quite large and the phase angles are mostly in the
vicinity of 200 degrees. Now, since CT, acts along the 0 or 180 degree vector
line, it would appear that a value coul1 be obtained for this parameter. The
effect of including CL in the equation of lift is shown by the residuals in the
middle graphs, Figures 7 a and 7 b. The fit has been improved, as shown by the
decrease in the sum of squares of the amplitudes0 However, the phase angles
still show a fairly consistent trend. The inclusion of the CLo and CW deriva-
tives reduces the amplitudes of the residuals still further lwhe the scatter
in the phase angles becomes considerably more random0
The transfer function constants were also computed from flight test data
using average values of the flight conditions and flight test values of the
stability derivatives. These computations are shown in Table 7. The computed
values of the constants are compared in Table 8 with those obtained directly
from the normalized equations0 The agreement is good for most constants al-
though there is appreciable difference in several cases such as Bo0 The effect
of these differences was examined by plotting the transfer functions for the
pitching velocity in Figure 8. The results are in very close agreement throughout
WADC-TR-52-71 114
the frequency range. The actual flight test results are plotted in Figure 9
also to show the OfitN obtained by the two methods. The transfer function
equations fair the flight test data very closely in all respects except for
the amplitude ratio in the lo frequency range.
In the normalized equations for the transfer function constants the results
of high frequency oscillations comprise the largest part of the coefficients as
a result of the high powers of the frequency that are involved0 Since the re-ý
sponses in this frequency range are small these results are probably less ac-,
curate and therefore should have less weight. That this might be the case is
indicated by the comparison of the plots of the transfer function equation
using constants from the least-squares solution and of the actual flight data
(Figure 8).
The effect of increasing the weight of low frequency data points was in-,
vestigated to determine whether the "fit" in this region could be improved.
This was done quite simply by including only the first 17 data points in the
normalized equations. These results are also included in Table S. In most
cases these values are in close agreement with those obtained from the previous
solution0 The frequency response characteristics for the pitching velocity from
these constants corresponded so closely to those already shown in Figare 8 that
it was not possible to plot the results as a separate curve0
WADC-TR-52-71 15
The method of least squares for vectors also provides a possibility of
solving for the aircraft longitudinal moment of inertia. The transfer function
A2 is directly proportional to the moment of inertia. All the remaining con-,
stants are divided by A, in the solution of the normalized equations. ohereo-
fore, the values found or those constants which are not a function of Iy before
division by b will be inversely proportional to the moment of inertia. These
include B B andi E0 By keeping all the other flight conditions
0 0. ' 0
constant and 6btaining data at two values of the moment of inertia differing
by a known increment, it is possible to arrive at an explicit solution for the
total moment of inertia for each of these constants0
WADC-TR=52-71 16
CONCLUSI ONS
The lift caused by pitching velocity must be included for conventional air-
craft in order to obtain reasonable values for the lift resulting from elevator,
deflection.
The use of vector algebra greatly simplifies the numerical computation re-
quired to obtain the normalized equations.
The equations for the pitching and normal acceleration degrees of free-,
dom can now be derived in the form that was found most convenient for analyzing
flight-test response data,
where M is the sum of the aerodynamic moments about the lateral axis, pound
feet.
axis,, slug-feet2
is the aircraft moment of inertia about the lateral
is the aircraft pitching acceleration, radian/second 2
Fz is the sum of the aerodynamic forces along the vertical axis, pounds
Equations I and 2 can now be written in the form of moment and lift
coefficient equations,
WADC-TR-52-71 19
c1~~fv~sSC
+ f~ cs + Cm 1 V2 SCQ + C 2SO*
vN (3)
d 0
CIft& 0C
SOCKJ
+LL f2. S
jVS C 02M
If2,VtS = -mgn (4)
where Cm or CL.,I i 019 d 00 or'c are the respective moment or lift co-
efficienrs or, more generally, stability derivatives.
Transposing the common terms to the right hand side and letting
CL - 2mg/IPV 2 S
the equations become
B0 CL C - C (17)
C CC C cc - 27t (19)
0
C0= C MCL C C -2TC (19)
c hCLj (20)
n V Dx 2
Z-. - (T- -)• E0 + EI D + EýD
2 (21)
21
A0 + AID + A2 D Sg
where E0 = -VB0 /g = V(CmCL - CLC )/g (22)
0
V(C 0 - Bl)/g = V(C CL C
C L C Cmo + Cj Cm S-
)/g (23)
The amplitude ratio and phase angle lag of any variable in response to a
sinusoidal oscillation of the elevator can be obtained from the transfer
function equations. These will be derived for the pitching rate only, as the
procedure is the same for the remaining variables. The elevator angle is
expressed as a function of time
WADC-TR-52-71 20
W is the angular velocity of the oscillation in radians per second.
Then 9 B
B+BID o2 • sin Wt (26)
A0 + A1 D + A2D
A + iB M reio
wherer B (29)
0= tan-IB
01 + 02) (29)
) 2+ j 222 sin (Wt-
S 7(B0 A 2
where 01 tan- 1 El
-1jB0
02 - tan _
A0 - A~kF
____
The condition that must be satisfied in order that a unique set of values
will exist for the stability derivatives in equations 7 and S is that the
coefficients of the derivatives must be linearly independent. This is expressed
mathematically by stating that the determinant of the coefficients of n equa-
tions with n unknowns shall be non-zero, The determinant of coefficients that
must be considered in this case is the same for the moment and lift equations
and is
I J2
0'c3 (50)
WADC- TR-52-,71 21
given by equations 9 and 9. Solving equation 9 for c gives
%ni~
Lc ýJi4 1 - 9it 4x (31)
CL L CL. 0 cJ
S UL ~ 0
CLb
3 9i + KK i
K1ni + Kdi + KK
where-whre
K,
1 = --CL -;ec; etc. i11,~
- 19 29 39~h 4-
CL
WADC-TR-52-71 22
G\A
CU Uf)\0 0 E--'6 Cr\O 1,K\-t \ Zt\ 0 W M~C0
KN K
oa 0 0 0 . 0 0 0 0 0 0 .0 0 0 0 0 * 0 0 0
ON 0 - : ONK\O Y NC,-IL-NW Z NC-Kýý
L~ : w0
'CDrIr- HH CO C\j uN(jN mCUiNcuNC\jj j
~ Lg . 0O H\CQ CQ 000 O\
-a a0 0 0 .0 . . . 0 0 0 0 0 0 0
(m
CU - ~ L\-ft
LC') (, O (. 0N0,\oC'\ oOa\ t-O Nt
G\ N\
0-10(jroK\L- '0 (-C~ fo 0NCC~(C~H040 V-- 8-O-1OLr
o 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0000 0 0 0
E- E_ rH0C--.C H'CCCCUUHHHri
co \0 0 l0\0 0 10 0- \0 0z 0~ 0~ 0 0 0 0 0
H H- H- H HA 0
H
r- o
0 0 ' n 0 0 0- 0- H 01 0 \0
C. 0 0 0 0 00 0 0 0 0 0 0 00 0
C)I00 0 0 0c 0, 0 0 0. 0 0 0 0 0 0 0. 0 0 0
CUj O"CkJK' -IUH K V'i'C
\D ,D C\U.4tL(LNnC>-H G\
.-Zo-Tc-i
CK~\
NS t--WO ON,0 H CU CUj :c
L(N'\.O
0
4 týb
"N'.- -Z Lf\ UNý nf'n' LV..\
\10
H a a I H H iH HH HH HH HH H H
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 04 04 0 04 0
N CH H ONw ri
U\ -: -I K'\CU CUHHH AHH
WADCO-TR-~52~-71 23
UN 1r- - ~j 0 00T\J N o o-\ C jt . 4 Or\ \ 0 L C\,
III Ifi l Ir-ir'4r41 I hII I I
1-4\ a\tia\Cjr- .
0 H 0 0 0 0A 0 0- 0- G\ 0- 0 0- 0 0- 0 CU 0) 0- 0~ 0
0 0 0 00 0 0 0 0 0 0 0 0 0 0 0 0 0 o 0 o
I ~ ~ r- H Hr4 WHH
iC'r\
L(,\fN IC'g 0S ~ O ' O - 0
H l - N - \K aj -- -60 ci cm LN--%-
"D
0 0 0%
0 o. 0 0 0 0 0 0 0 0
k- -r-i' LrN-iCM
Oi mcmty- O i m H M IiIII
HH
50-NO C
10 0 K-\z(R.Zj0 O 0P I'C',I
K\60 CMKO K-1 10 CM CM
r'- 0"-C
H~ C~ \
C~j--I ON % 10,N1
1-0- 1E wN m~zNN C\1MI I- I,-, WONC-0-
H cm
HH0CN~'4> V.0 t'l M -tm o\
r 0
0 . . 0 0 0 0 0 0 0 0 0. .
4." 0 H CMK
tc~LC"'\ E-O ONC\O H CM K-\U0 -~ OOHC
Pý ý N -4 -1 r1 r4 ' H, N aM
CMC
WADC-TR--52-71 24
N &WMO Z 0 1,0 -4 W K Ot H --I w 0 H
UN ~~~ C\ ~CA)-w% 0 \
Cy\ r4j K- 0 "
\0 0- 0 0r 0~ 05 N 0 011 0 0N o
r--IM Cm 01 cQ
W~~ IP7t
El I HH~
1A O,-:1-:fIo \Z 6 'O oc~-t-o
r -0C-0-*-r-O -1 0
0
H
(\J~~~ 3- 0 0 0 0
N~
H~goC0Pco\
0 0 0
0
30 . . 0a 0 0 0 0 0 0
CCM
-o w 0 K-\S C rr-0 H tc
_4 U-, N, C3
o 0 o 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
H- \ r.-zt
wN -1\D I-- (3\ H U\O N r-4 ON UN CM I. 0 0" 0
r- H CO0K' \0p0 E \O o
rA ID\
W
LIN \K 0 NN N
0o~ HUHP PC\\WCC\o~ K-
0- 60 ý% s0 f-
+3 CM Lf\' C 0H MH 0 '0 H-
oo . 0 0 0 0 0 0 0 0 0 0 0 0 0
w 10OO ýrUNHLr%
1\ O\ N0 NfCt,&
UN
CMCH~ N 18 -'OC OLr{'ý 0
CM
Cr) 0- 0o "0 0 0 0 0 0~ 0 0N N
0\ 0 0 0 0 Ht0
u 0 0 0 0 0 1
0O\Lf0 -
4
0 t- 0w\ ~-cy\t 0WLfH
H wcQ t- L% KI
v-I
0 ~ t-v
w
0CQ tK- cý
0
0\U\
0 0 0
A ~tow0
00 0 0
HNK,CMj
H2 H
0
H
0 0
0 0\tt-I0--'0 0 0
0 0 0 010
W~~~nG- IT-5-712
Q~ - Cm U\ 0 \0 0\o\ E- Lf sC W E,-o0 0
ROto0 LrAU O ~
cO0C\ El 0I 0r
CMJ0
3 -0'
0
r-4
0 0 0
mT0rIr
o
0
HCJC4r-,
0
\ ..4
0 0 0 0 0 0
o
LfV.O'
0 0
\OI--
0
t- ON G\ HCcm K%
0
(M
o-_:
0 0
Lrl%
n
0
0 1 HH- r4 0
00 0 0
fN 0 (,0 - ) Y0 r- 0 t- 0 o , w H 0 0 0( E- 0
C)I , r4w r.HHc\J w nC\Jo - H 0
00
00c 0 0' 0 00 S 0 0 0 00 0j H 0~ 0 0 0 . 0t
W -4 Lc,%r-i
Hy Y , Lfr\C~j0\rNLCC
LPA %LC\ EC
- <I\ 0f Gt\I-L Y
0 0 0 0 0 0 0 0 0 0 0 0 o 0 o . 0 0 0 . 0 0
I rI f-l -I -I H r H H H- H H H H Hi
rI fo\ W, INK \ O 0
H00 \
bCbjt- fkC-- G-\\04 br- - K ON\
CiM~CN-Ob H ~OK- MfNýLrN~\
L- o Cis-\U-\U-\ LC o \HNt
LN L HNL'
WA0T-27 26 .. 0 o*
'o Lu- WNl- r ir*- CUlN M~\~k
'o ~ ~ r t O~-- l r-r-CI
+-4ON 1O~~ - N 0 tlOON (\ E- 'Klf\MCuHHH4C- C4O
N 0i 0~ 0 0 -0 P 0 0 0 0 01
O OL,\\O
ONu U\0 C'-C\ t--CQ LCrHH Hr-A r-I r-HH 0N1
+ W~uH c l H - N0 (\L' l HOA
-C- ONt
oND0 'OO4uuNt-4Cu00000 P CO
Nu CUjCu Cu H rir-
- C' 0
rUNOC- - L(,.N)cu QLC\NQfO iWE- UN f
ONUH u 4cu W'
w"400 NNCJý (-0-& r-- ~\0'- U' C
-1'0\ K.5OCuiN\-., N t--K-N\ -- 0 HONONC),Lr C' -LC- ý
H CUjCU Cuj\0\0 LCA
-- I--\0 H.C (MONw\0"LS-\
HC 0 o \0 0 UN"0 ir04
0t- 0 Cu HII\ 0.t=uO
a 0
IH\D[- '\Lf\-OONUý\ Lr\
I- K" Df\t-(\HK
~ C-0 CUcýGu ULi'N\0 O 0-4j N
IH 0H~~
H ýUw1- Cuj Lf'\ N 41A 0\
~
f O-:To j ON KN4~t-~ \JD~-
K%~~tCuj
\0 '- ON r-4 O0\ - uN C7\ -NHLCu0CoLrA aHj K\
OH -ztQuwtbo -4
o 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
H- H- H- H H Cuj Cu Cu Cu
Cuj CuCu NN C %N.J
I l\ n
0\O0 (3 ýj \Dc 5 \D 0
-o6O0 \I Lc~--0 --.
400 L--
U0,0CjC\
-6"0O ~ 0uL\ NH W \O~H'NN'
rl %
O\UouuONoO Oj Nq\0KCl ONO c-H~Z H-
0 0 0 0 a 0 0 0 0 0 0 . 0 0 0 . 0 0 0
KN Cuj 9hNV-K,0-1 KNL' CU ý bO El-O b 160
KN cQ 0SNI\ 0
rl61
W \O IN WCuM" 0
H H E-Cu \fCK\L\O Q
-4 H8ý.,Ou~H-'OCW0t- W\ N\
0 i
Cu-7C H
Cu ~ 0 0 0 0 .0 0 0 0
!4 ri 0
0 0 0 0 0 0 0
H
C\JCuJCuQuK% n C1 CC ' O uHl0--1 N
Cuj \0 $~c ~ONW f\C.--! \0 \0 k>-- '-0
X ýil r% L- c-\ r~- 4
WADC-TR-52-71 27
00D o a',\H LVC'C~jc-. t- CUO tN
m~~~~c '~0 to~~O~~
0 ~~ * a 0 0 0 0 0 0 a 0 0 0 0 0 0 0 0 0
3- HHr4HHHA -IH
HHHH r-IO-
0 H
Q) w%~r<,
CUjKN~ K-\ CUjCj H4 H
O 00 HU Cj rt WKI H N\t- l N KI CU
o . 0 0 0 . 0 0 0 0 0 0 0 0
bO V- Lf,%tqC\ r I H CJHUCXJ
NH r-4 H HH r-4 -I
"-4.
U~j~'\
--r<-
a',
c0
Hi rt~-.z6
t- a' 6- Lr''.aý-I
_: w ' 0 CUjH Hsr.- ON
0\ 0x~~'
U'N\\ n-'(bvUNr ic'\
r-a-.r--- t0
LCI
C--K\%
w WKNNtC- -H4.\0 r4 H H -ZI H '60 00
0i 0e -0 0M 0*I 0 0 r0 -Z 0 0 0 00 0~ 0~ 0- 0- 0- 0 0- 0
r- H H
I
-- YIp T--
H I
- ~ - II 0r
I I 1 ti IC
0o . 0 0 0 0 0 0 0 0 0 0
Q~ :~t0E-
C~0 tC-- r- 4
ý - 4IC -:
.H
NW
\0Hri\0
~ -
c)-\
,
00 V'N 0 0 0 o 0 0 0 0 0 0 0 0 0 0 0 0
0 0 O0
(% K-% 0 \O.Uc
--I C\J \H lI I 12 1 i
WADC-TR-52-71 2
0 \ 0r
0 0 0 0 0 0 0<, 0~ 0i H
0i 0 0 0 \ 0 0 0 , 0
B e - N4 - s~OO
o PC,PC~N
YO KIN
L K ND ) - \ O O-\Cj w t--
N4---z~~' r-
0 0 0 0 0 0 0 00 00 000 0 0 0 0 0 0
.4.4 H w r1H
0 t-- KH I '0 6 - ýOt' 4 r ,8
r-4~ 4OH
r-A N~ Co.,~ H~N Hr-4 NO
\0 ~
\0ON- 11 y\D\)
1 L\ ~' o0 : ý~
CM)
.40r- cr .4'.OU ON
xIUN z\O-
C-K WNr\q
00W\ \
0k H -4 '\ 0 0t
H+ 00 7'u~ KN tC- W\boO \JK"HCMj rlfN\-LfN'~C\Jt -o -It -K
Q.Jm. ON-
m 0 0r, 0c 0 0- 0- 0:: 0~ 0~ 0- 0 1 0 0 0 0 0 0 0 0 0
VZO (\j r-4 o~j311 MO re CMAH ON
.4 O\ aN RH0 tQ
O OU\ \
06 .4ý r- -
LI& +~ ONý VNC--OtCM !l,-) N C-- 0 CAvON--I
ON\D
I s 0--0 0c 0ý 0- G\0
0N 0 0f 01 0t- 0" 0 t- 0 0 0j
\ 0 P
.4 > O ý --
N'- zt-
Nr o-j4() E--OO~C t-b 0~ - -
0+,
0 HMC.Lf fV.O0 - CiOON -A0 \0HCM LvO\_0 tr--bO
ýe(\4 0 0H C
Hr-4
HHHrI H H H \ CMCMCN
WADC-TiR-52-71 29
C'JCM
A Cuj
HICu m Cuj
4)~C
oc Cd C u
000
H~~~ C :HrIC~
H- CM , -4
_z
C~Cu
cuy Cu
Cu~
H- I ~
11" aU1 I I a 4
N N t.
H~u
r-i~c~ NC ~ f 0\
I- I- H
H~ ~~ d \U Hl
HOCO Cd~ CudC
0K
HNKNH( 0
0 H+ + + C -j Nc
coIC HC (\j N CM \ Cu
+ + + + ~ -
H
H HH H HINCH
H CH
r,H-4 H- _q x H Cujc PI\C'0c
.H K\ ' - 0 ___4
WADC-TR-52-71 30
TABLE 5
SAMPLE SOLUTION OF NORMALIZED EQUAITONS FOR STABILITY DERIVATIVES
G (9 G
(See Theory, Equation 19)
Augmented Matrix:
C = o14i25
CL = 29699 - ol125 x (-1,90670) = o55631
CL 5.17654 - l1425 x .86357 - .55651 x (-o10165) 5ll094
Cm = -. 26959
Cm -°90383- (-,26959) x (-1.90670) = -1h41786
C -o642o6 (-,26959) x 96337 - (-1,4178>6) x (-.10165) -o55342
If C is assumed to = 0, then:
CL - °28699
w
Cý, 5o17634 - .29699 x (-.10165) 5o20551
If CL6 and CLa are assumed to = 0, then.
c 5ol7634
WADC-TR-52-71 31
TABLE 4
(-@ 0
-O -@ 0
(Reference Theory, Equation 24)
Augmented Matrix:
Auxiliary Matrix:
K0 = 4.00456
WADC-TR-52-71 32
TABLE 5
Augmented Matrix;
Auxiliary Matrix:
10 h4,16667
WADC-TR-52-71 33
TABLE 6
0-0 -0 0 -0
-0 -0 0 0 9 0
0 -0 0 5 -@ o
_@ 0 G -0 Q G
(Reference Theory, Equation 26)
Augmented Matrix:
Auxiliary Matrix:
E2 - o7994o
A, 2.,6051l
A0 = 3o95453
R1 = lo19013
9 609ohgo
0
WADC-TR-52-71 514
TABLE 7
07
0 (c-m -CICm)/A2 =-6o359sec
2
= V(CMYCIo C CmC) 2 a 52<334 sec
to
WADC-TR-52-71 35
TABLE 8
2)
f1D)/(Ko + -1D + D)
(fro +
i h4167 4o174
0
Al 2o915 2.959
0 -7°599 -7-732 -7-752
n/- ( 2 2)
(to 1 D+ D)/(A0 +A1D )
A 3-954 4o189
0
A1 2.,605 2.99
2 06°9o5 51.211 52o354
S.lo90 0°L35 0.527
WADC=TR=52-71 36
WA -T 37
-5-7
4 H
.44-
r++
tt
hill I
:m MM-1 f m. I
+ I MH
4 tmff-t: wt
44-f-111Tmý IF.
all$
g g fM Mis TT
T T
+ IT
®r
tMI, MIM
EM
I iH
4+
NOR
41 .. .
M Mv fim: +
4 1
A ým
Ira H 44 4 4T T-:P:I+
#69
T; tl-+t
+1ý11t
1; + ......
4
w
P11 4ý
Rin
Ti
194
tT
ig.
1+ 44414
4MM
21 ,Ml
w"t N111-T
IRIMIIIIZIP, I f, i ll, a
ifl
44
41
ILL L
WAD-C-TR-52-71 39
I ~I
14 I.
K: ~ 7rVt
-~rtt i
A.:.... 1 . .. f ..
Ft - TJt'
4
i Ji
4L -7-H-j;
WALE-T-52A7 -!*A9
T
:I.:
wt 41r,
.......
. . :z
Olt
t,
tW
i7
f
Itt
W-0 ýT,
7
JL.;-
-5 ' ...... .
..
Tr
4 44ý
4j-
jr
L
11
.114
4 1,J14
144-
7T=r
44, 'r
llf + 1H
ff,-, AR 4j;
fi% ttý 4r
M 4 Ill
4f- Mr
14W ill it-
MW
.rtfl
41
41
.14
ýWADC-TR-52-71 40
i7j jji77 -,77177
-V
42~47~1VL7Ij
h
j44L~+
_71,
yfi7 47
7_ -j IL -i~I1
:- 7.
4-I- ± J-p~
11j774 - * i1 j.'
7 "tT
ifi77i};f;; .1.
7"-7T~
t7- t
T
I',~~~~7 -77i--.i~-~I
V77~ ii..Ii+2....J iw...
.
.I
t1 :JAt
-4
-'I.-7 ,
~ __ ______________L~LLLiL iiq
527 aJD TR
F - -,7 -71 ,7 =1 4----
I4 L_ 4
K~7
I i -pCý~{
i _7 7-
I7 7
I~ Tf
Ft-_ ..777
7T7
r 7
- ~ ~ I"it
:1 ~ F 40
L- 7 11
--. -- -T 1 r7+
L ~~ -IC)9I ~ ~ LLA
'1:-
A -C T;- -j
__ 2'.M
Il TH
I :f
Ptf aft. W I
44
I41
FITA
i4 -fl- -ff
44"4
~ 4M;
V;A4N
1. -
JAW TRo2711
17 'F E., 1,1- li!j
TIM
TF
I I
it ....
Fýrý 177 L
.. 7 TI; il 1
11
4"t Mt T419-321 I
C, V:
t! '4qd; I 114, 1 +
7i
T
"1"
444 i 4'l 4 frj U 4,44,
r! . .....
t ji t
ff 4 ýTj
4+1
'it
!j 1, 411 1t
ti tIA- 'Hi
4
ý'Pl 4- 4"
jrý III t. 1
INCH M
t v
L6 t H- pi I
f
4 AA
90, Rký
# .
.. ..
i _,11T i 1, 4- _11Vt
+ ... T m :1
1 44-
#
+
I Ar M.-
ti I Mwýma gl
r tT
4f,,Wif: +
I +i+
+
7- + +
Tý 11`71- T-1
4 44 +
ýT
IT
1:1
Ur MU:
M m 441 7l; 4--,-rq
ý1,
ý4
41t
M V,
_ff"t_1mq#W-V 4 p T
Tf
Hý ý4ý4 ++ iý,,: -
if,; 4 T T_; W IM
t,
41h i
;;I- tiýýj
iI I; TM , 'N -4 -1
+ 4i 4-+
XX i
7r
TR-C-2-71 44
1 4 -
1q, 11
I` IMM
1.'ý
RT TV
T,
7
11
IN 114, jj 14 tt I RJ'ilf tit i
Ij
..
til itk
j. i
. .11 1 , -11- 1 J,4
4
ttttttt
4
1 11ffl
p
417
4 4 Wtý I
if
TrI,
ýj4
AV 91
-4 Tflll
#fff
'w,
ý4: AM ýd
I- z
14V trRt
lN
itfill 1;i-T111
It ig: 1:1
-F,
i-I
r, HU T P-- :l4 I1 4114:1
"L4 4,
444
...........
PIT -- :1 :
I-iq
#i
I - ; :-, -ý,
T71 tR
TH 14
ý4 t V,
HIM
Tt TT
I - - - ir -, .11 4, 1
4t ft 1 ' - :
A f Ila
'17 1:4 1TO,4 J,ii- 41
Mq 4,1
Tlý 11ý ll W.
S M 44 --
-
r it 4-
Jý 5 WN Ti 1,111
4111 IT :4
_4
+ +
t+4
;TIT
14 -:T4ý - 4 iV
z, 41 Cý 41
;7ý
f", Ht +I 41
+4f i+
4ý t4t1i .
Mlý VT
7 F.'t :ý .. ;T4
44 ir::I
77
4_1
, ; T - fH
+ 111t-ý -1
ELI -Itul-Tri ITT
WAX-TR-52--fl 45
REFERENC ES
3. Phillips~, H. B., Vector Analysis, J, Wiley and Sons, Inc., New York
WADCTR-52-71 46