https://ntrs.nasa.gov/search.jsp?

R=19870017487 2018-10-16T15:51:33+00:00Z

NASA Technical Memorandum 88297

Dynamic Stability and Handling Qualities

Tests on a Highly Augmented, Statically
Unstable Airplane
Joseph Gera and John T. Bosworth

(NASA-2'1-88293) D I I i A ffIC S f i B I L I f T ABD B87-26920

B A B D L f l G QUALITIES ZBSTS 01 4 816ELY
AUGCIEZIHCED. SPASICALLY UUSTABLB ALIPPLILEIB
(PASA) 16 p Arafl: IOTIS HC 1102/LIP A01 Unclas
CSCL O X 6 3 / 0 8 6092572

n
August 1987

National Aeronautics and

Space Administration
 NASA Technical Memorandum 88297

~ ~

Dynamic Stability and Handling Qualities

Tests on a Highly Augmented, Statically
Unstable Airplane
Joseph Gera and John T. Bosworth
Ames Research Center, Dryden Flight Research Facility, Edwards, California

1987

National Aeronautics and

Space Administration
Ames Research Center
Dryden FIight Research FaciIity
Ed wards, Gal i forn ia 93523-5000
 DYNAMIC STABILlTY AND HANDLING QUALITIES TESTS
ON A HIGHLY AUGMENTED, STATICAILY UNSTABLE AIRPLANE

Joseph Gera* and John T. Bosworth*

NASA Ames Research Center

Dryden Flight Research Facility
Edwards, California

Y output vector

Initial envelope clearance and subsequent flight YR output sequence for open-loop roll
testing of a new, fufly augmented airplane with an ex- frequency response
tremely high degree of static instability can place un-
usual demands on the flight test approach. Previous YY output sequence for open-loop yaw
flight test experience with these kinds of airplanes is frequencyresponse
very limited or nonexistent. The safe and efficient
flight testing may be further complicated by a multi- U control vector
plicity of control effectors that may be present on this
class of airplanes. This paper describes some novel
flight test and analysis techniques in the flight dynam-
ics and handling qualities area. These techniques were The X-29A airplane, which began the initial
utilized during the initial flight envelope clearance of flight tests in late 1984, is a fascinating example of a
the X-29A airplane and were largely responsible for the statically unstable, highly augmented, multisurface
completion of the flight controls clearance program airplane. Its evolution, design, development, and
without any incidents or significant delays. initial flight test results are documented in Refs. 1 to
11. This paper describes those flight test techniques
Nomenclature that were used for the fmt time at NASA Ames Re-
search Center, Dryden Flight Research Facility (Ames-
A, B, C,D matrices defining the linearized Dryden), in the flight dynamics and flying qualities
mathematical model of the test area and contributed considerably to the safe and effi-
airplSme cient flight testing of the airplane.
ARI aih-to-ncdda intacoMect gain The most significant of these techniques, from
the point of view of ensuring that adequate levels of
FFT fast Fourier transform dynamic stability existed at each test point, is the near-
real-time computation of the open-loop frequency re-
RAI rudder-to-aileron interconnect gain sponse of the X-29A in the pitch axis from pilot-gen-
erated frequency sweeps. The frequency sweeps had to
X state vector be performed with all the feedback loops left intact be-
cause of the e x m m e static instability of the unaug-
X time derivative of state vector m n t e d airframe. The open-loop frequency response
yielded the actual gain and phase margins immediately,
XR input sequence for open-loop roll allowing a quick comparison with precomputed
frequency response stability margins stored for each test point. Any un-
expected nonlinearities, such as position or rate
XY input sequence for open-loop yaw saturation of actuators or time delays associated with
I
frequency response data conversion between the analog and digital ele-
ments of the airplane, were reflected in reduced stability
margins. Gain scheduling errors due to unmodeled air
m
data characteristics also became obvious long before
*Aerospace Engineer. Member A I M . any indications were given by conventional flight test
techniques. The utilization of the open-loop frequency
The abstract of this paper was submitted by the authors response turned out to be a highly successful endeavor
to the Society of Flight Test Engineers and was accepted in other respects as well: It resulted in a roughly 30-
by the Society for presentation at its 18th Annual Sym- percent reduction in the time allotted for initial enve-
posium, Amsterdam, Sept. 28 - Oct. 2, 1987. lope clearance.
 Although not related directly to safety-of-flight gain and phase margins, respectively. As the final
issues, the closed-loop frequency responses of the pitch control system design evolved, however, it was found
and roll axes were also computed by the same trans- that these requirements were not met in the longitudi-
form algorithms. This computation, followed by nal axis when the analysis included higher order dy-
finding the lower order equivalent system dynamics, namics. Since the requirements are usually applied to
allowed a rigorous application of current military han- airplanes in series production, they were relaxed for rhe
dling qualities specifications to the X-29A aircraft dur- X-29A airplane, an experimental aircraft built for
ing the early phases of the flight test program. flight research. The relaxed requirements are shown on
a typical open-loop frequency response plot for the
Recent improvements in the computational ca- pitch axis in Fig. 2. This plot is obtained from the
pabilities at Ames-Dryden allow many routine data linear transfer function model of the augmented air-
processing tasks to be performed in real time because plane.
of the utilization of extremely fast parallel processing
of the data. One of the first utilizations of this capa- After encouraging results from simulation and
bility is the comparison of X-29A flight data with the postflight analysis of the data from the initial flights,
output of the linearized simulation in real time. the procedure for measuring frequency responses,
shown in Fig. 3, evolved and was used with consider-
Test Obctives able success throughout the envelope clearance pro-
gram. The procedure involves the computation of the
The main objective in the flight controls disci- open-loop frequency response while all feedback loops
pline was the demonstration of the design dynamic remain intact. The computation relies on a fast Fouri-
stability levels during 1-g flight in each of the control er transform (FFT) algorithm, which is executed in a
system modes that could be selected easily by the pi- high-speed parallel processor on the ground using
lot. In addition to the normal digital mode, the pilot- telemetered data from the test aircraft as input. As
selectable modes include a digital and an analog rever- shown in Fig. 3, the input and output sequences to
sion mode. The verification of the design margins the FFT, designated as X and Y,respectively, are the
during I-g trimmed flight was felt to be sufficient to error and feedback signals in the pitch axis control
extrapolate to higher load factors by using either ana- loop. Excitation of the loop is provided either by the
lytical or simulator results. pilot through the command shaping path or by an up-
linked signal that is summed with the pilot command.
Since the determination of stability levels re- Although the latter approach results in more precise
veals little about how the airplane flies as far as the control of the excitation signal, it was found in prac-
pilot is concerned, a parallel test objective was to es- tice that the pilot-generated frequency sweeps were en-
tabiish the handling qualities of the X-29A airplane tirely satisfactory. The use of the uplinked signal be-
with the assumption that the requirements for high- came necessary for computing the frequency response
maneuverability airplanes apply. of the roll and yaw axes whenever roll-to-yaw or yaw-
to-roll interconnects are employed in the control sys-
Flight Test A m tem. Figure 4 is a schematic of such a system, which
is similar to the X-29A lateral-directional stability
Flight testing of the X-29A airplane was differ- augmentation system. It can be seen from the figure
ent from routine evaluation of the airplane in flight that the total error and feedback signal cannot be used
since many new technologies were incorporated into in either axis to define the open-loop frequency re-
the design. From the flight controls point of view the sponse through a Fourier transform since the pilot-
most significant of these new technologies were the generated frequency sweep excites the control system
approximately 35-percent static instability of the air- through more than a single location. Although several
frame, the digital fly-by-wire primary flight control possibilities exist for obtaining the open-loop roll and
system, and three kinds of pitch control surfaces: ca- yaw axis frequency responses, the approach used for the
nards, symmetric flaps, and strake flaps. The general X-29A testing involved an uplinked frequency sweep
arrangement of these surfaces is shown in Fig. 1. The signal directly summed into the aileron and rudder
approach to the initial flight tests and envelope clear- actuator commands. This procedure amounted to
ance was influenced to a significant extent by the new mathematically breaking either the roll or yaw axis
technologies incorporated in the flight control system. feedback loop at the actuator while keeping the other
loop closed. It is important to keep in mind, however,
Originally, the design criterion for dynamic sta- that no feedback loop was opened physically, thus
bility waq the usual requirement of 6 dB and 45" of flight safety was not compromised by the frequency
response tests.

2
 In addition to monitoring dynamic stability in quiet in flight. Transitions among the flight control
the frequency domain in near real time, the availability system modes and between the ground and air were
of the linearized mathematical model of the test air- smooth.
plane at each test condition and the high-speed data
processing capability on the ground ma& it possible to
compare the response of the airplane with that of the The computation of the open-loop frequency re-
!icexized simu!atien !c Identical pi!ot inputs in real sponses turned out io be a surprisingly uoubie-free
time. This procedure is shown conceptually in Fig. 5. operation. For the longitudinal axis, the computation
In practice the procedure involved the computation of was performed as soon as enough data were accumulat-
the linear equatons of motion at each test condition ed for the FFT algorithm. For the particular FFT al-
immediately before flight from an all-FORTRAN, gorithm in question, this was the case after accumulat-
nonlinear simulation of the airplane. To facilitate the ing 2048 data points. This required 52 sec of trimmed
real-time solution of the linear differential equations, flight during which the pilot performed not only a
the state transition matrix was also computed for the pitch stick frequency sweep but also a series of longi-
sampling interval of the flight control computers, so tudinal pulses and doublets. The execution of the
that during flight only the solution of the difference FFT algorithm required a negligible amount of com-
equations was required. The comparison is made be- puter time, and a video display of the frequency re-
tween the linear and measured time histories of the sponse and the associated stability margins was pro-
principal motion variables, such as pitch rate, angle of duced in less than 3 sec. Figure 6 shows a typical
attack, and normal acceleration. The initial or aim pitch axis open-loop frequency response plot obtained
values are subtracted from subsequent flight measure- during flight. Also shown is the frequency response
ments in order to make the flight data directly compa- predicted by linear analysis at the same flight condi-
rable to the linear solutions. tion. A remarkably close fit between the flight data
and the prediction may be noted in the rigid-body fre-
The open-loop frequency domain analysis was quency range. Generally, this was the case below and
also extended to include the computation of the closed- above transonic Mach numbers, indicating where the
loop frequency response of the airplane between mathematical modeling of the airplane was most suc-
longitudinal stick and pitch rate, and between lateral cessful. It should be noted that the flight data were
stick and roll rate. In contrast with the open-loop fre- consistent and repeatable everywhere in the flight
quency response and the time history comparison, the envelope. In fact, the frequency response determined
closed-loop frequency responses were computed post- from flight data was of sufficiently high quality that it
flight. The objective of this work was to obtain a was possible to make changes in the pitch axis control
quantitative measure of handling qualities. The closed- system loop gain based solely on the frequency re-
loop frequency responses were approximated in the fre- sponse results. An example of this is shown in Fig.
quency range of 0.3 to 10 rad/sec by a lower order 7. The initial determination of the open-loop frequen-
system that also contained a pure time delay or trans- cy response clearly shows that the value of the loop
port lag term. The approximation yielded the equiva- gain is too high by approximately 2.5 dB,resulting in
lent modal response characteristics and the associated inadequate stability margins. Reducing the loop gain
time delays. The lower order system characteristics by this amount by simply changing the flight control
could then be compared with the requirements for high- system software restored the stability margins to
maneuverability airplanes. Data from this comparison nearly optimal values. This gain change was the only
and the pilot ratings and associated comments during major control law change that affected stability, and it
standard handling qualities tasks revealed whether the was accomplished during scheduled airplane mainte-
requirements were applicable to airplanes with the un- nance without any delay in the envelope clearance
usual characteristics of the X-29A. program. With this change in place, the longitudinal
dynamic stability exceeds the minimum margin
requirements throughout the flight envelope. This
example demonstrates the utility of the on-line
The envelope clearance of the X-29A airplane frequency domain analysis of flight data; namely, the
was accomplished in all selectable flight control sys- direct information on the overall system stability and
tem modes without any unusual occurrences such as the ability to make control system adjustments with-
control surface oscillations, limit cycles, or unfore- out the precise knowledge of conventional stability and
seen interactions between the flight control system and control derivatives. In previous flight test programs,
the structure. The airplane appeared to the pilot to be similar adjustments required a considerable amount of
well damped in all axes, and the control surfaces were time for postflight data reduction, analysis, and gain
correction.

3
 Since it is the pitch axis in which the X-29A that the amplitudes, frequencies, and damping levels m
airplane exhibits highly relaxed static stability charac- close enough to predictions that flight safety is not
teristics, efforts to obtain stability margins during compromised.
flight tests concentrated on the pitch axis. Attempts
are being made to perform similar computations for the
roll and yaw axes. Although no flight results have Closed-loop frequency response characteristics
been obtained to date, simulator results are encourag- were obtained for both the longitudinal and the lateral-
ing. Figures 8 and 9 show the comparison of linear directional axes. For the longitudinal axis, the fre-
predictions with nonlinear simulator results. As can quency response was computed between the pitch stick
be seen from these figures, the augmented X-29A air- displacement and pitch rate; for the lateral-directional
plane exhibits very generous stability margins in the axes, frequency response was computed between the
lateral-directional axes; these margins are reasonably roll stick displacement and roll rate. The pitch and roll
close to the predicted values obtained from linear frequency responses were then fitted in the mean square
analysis. sense with a pure time delay term in conjunction with
second- and fmt-order transfer functions, respectively.
Figures 12 and 13 are examples of closed-loop
frequency responses along with the results of the fit-
To enhance flight safety during the initial enve- ting procedure. Similar data were obtained throughout
lope clearance, monitoring the stability margins was the flight envelope of the test airplane. According to
augmented by comparing time histories obtained dur- the data, the airplane appeared to the pilot as a well-
ing stick and rudder pulses and doublets in flight with damped system in all axes.
time histories generated by the linearized simulation of
the test airplane in response to identical pilot inputs. In the pitch axis at some flight conditions the
These inputs were obtained from the telemetry data and equivalent short-period damping ratios were above the
were used in real time to generate time histones that maximum recommended values for high-maneuver-
were paired with the corresponding flight data and dis- ability airplanes. At the highest dynamic pressures at
played in the identical coordinate systems. A typical which the airplane has been tested to date, the equiva-
comparison plot is shown in Fig. 10 for pitch axis lent short-period frequencies were below the recom-
variables. For the pitch axis, due to the extreme in- mended values. The equivalent time delay in the pitch
stability of the unaugmented airframe, the linear math- axis at all flight conditions was found to be slightly
ematical model included not only the rigid airframe above the recommended values, being mostly in the
equations of motion but also the entire flight control 110- to 140-msec range.
system including such details as sensor dynamics,
transport lag, notch filters, and antialiasing filters. For For the roll axis, the equivalent system results
the longitudinal normal mode, this resulted in a 48th- indicate that the roll mode time constant is in the
order linear system. The processing of this sytem in neighborhood of 0.3 sec throughout the flight enve-
real time was well within the capability of the array lope, well within the range of recommended values.
processor that was utilized for the ground computation. As in the pitch axis, the equivalent time delay is
The use of linear simulation for comparison with slightly longer than the recommended 100 msec or
flight data is not without certain advantages over a less, being in the 120- to 150-msec range.
complete nonlinear simulation. These advantages in-
clude the increased speed of computation necessary for A limited amount of testing was devoted to
real-time comparison and the ease of detecting unex- evaluating the handling qualities of the baseline flight
pected nonlinearities by the test personnel. control system during air-to-air tracking. In this con-
text the term baseline means that no modifications or
For the lateral-directional axes, the comparison tuning of the control system had been made for the
required considerably less computation. Since the test specific purpose of improving the handling qualities of
airplane does not have highly unstable modes in its the test airplane. The tasks used for the evaluation are
lateral-directional characteristics, there was no need to shown in Fig. 14. In each task except the simulated
model the entire flight control system. It was suffi- terrain following, the lead T-38 airplane was perform-
cient to consider the rigid airplane lateral-directional ing turn reversals either at the request of the X-29A
equations of motion. excited directly by the aerody- pilot or randomly while gradually increasing the load
namic surface positions, which were available from factor up to 3 g. The simulated terrain following was
telemetry data. A typical comparison plot is shown in a pure pitch axis task in which the lead airplane per-
Fig. 11. Although in the fine details of the motion formed mild pushover-pullup sequences at load factors
the Comparison is not as close as might be expected, it varying between 0.5 and 2.5 g. Since this task proved
yields adequate information to ascertain in real time to be the most difficult for the X-29A airplane, the
task was also performed by each pilot in another T-38

4
airplane. Figure 15 summarizes the pilot ratings. The the data allowed the verification of the linear
results show that the airplane has satisfactory charac- mathematical model of the test airplane and were used
teristics for the tasks, with some minor deficiencies for redesigning the pitch axis loop gain at transonic,
that warrant some flight control system improvements. low-altitude flight conditions.
Before discussing the deficiencies, it should be noted
that the airplane handled remarkably well considering The availability of a linear mathematical mod-
the number of new technologies incorporated in its el of the test airplane d b w e d the real-time computa-
design and the fact that no control law changes had tion of predicted time histories using the pilot inputs
been made for improving the handling characteristics. from telemetry. These time histories were compared
It is worth noting that one pilot found the air-to-air with flight data, also in real time. The comparison not
tracking characteristics to be excellent. only allowed an immediate assessment of frequencies
and damping levels at new test points but also gave a
Pilot comments indicated that they would have clear indication of any nonlinear behavior of the air-
preferred plane, which could result from rate or position satura-
- slightly faster initial pitch response, tion of any component of the flight control system.
- better stick geometry since pitch stick travel

is too large in comparison with lateral stick Postflight closed-loop frequency response data
travel, and were obtained from pilot-generated frequency sweeps.
- lower lateral sensitivity for small stick in- These were fitted numerically with a lower order
puts, faster roll rate for large stick inputs. equivalent system, which yielded the equivalent time
&lay and modal response characteristics. The results
In general, it was found that the current handling indicate that the current handling qualities requirements
qualities requirements are applicable to the X-29A air- for high-maneuverability airplanes are generally
plane, although more testing will be required to show applicable to the X-29A airplane.
whether the equivalent time delays indicated by the
analysis can be better reconciled with pilot comments. References
lKrone, N.J., Jr., "Forward Swept Wing Flight
Demonstrator," AIAA-80- 1882, Aug. 1980.
The initial envelope clearance of a statically
unstable, highly augmented airplane required a signifi- 2Spacht, G., "The Forward Swept Wing: a Unique
cant amount of on-line data processing. Examples Design Challenge," AIAA-80-1885, Aug. 1980.
from the X-29A flight tests illustrate that the open-
loop frequency response of an airplane with highly re-
laxed static stability can be successfully computed on 3Whitaker, A., and Chin, J., "X-29 Digital Flight
the ground from telemetry data. The data were obtained Control System Design," AGARD-CP-384, Active
while all feedback loops remained intact, so the process Control Systems - Review, Evaluation, and Projec-
did not compromise flight safety. The required compu- tions, Oct. 1984.
tation and graphical display of the results, which in-
cluded the gain and phase margins, were performed in 41shmael, Stephen D., and Wierzbanowski, Ted, "X-
less than 3 sec. In the pitch axis where the flight 29 Initial Flight Test Results," SETP 29th Sympo-
control system is essentially of a single-input, single- sium Proceedings, 1985, pp. 95-113.
output type, the frequency sweep required to excite the
system was performed manually by the pilot. In the 5Zislin, A., Laurie, E., Wilkinson, K., and Gold-
lateral-directional axes where the stability augmenta- stein, R., "X-29A Aeroservoelastic Analysis and
tion is accomplished by a simple example of a multi- Ground Test Validation Procedures," AIAA-85-3091,
input, multioutput system, the required frequency Oct. 1985.
sweep will be uplinked from the ground directly to the
aileron and rudder actuation system. %efic, Walter J., and Cutler, William, "X-29A
Advanced Technology Demonstrator Program
The on-line procedure that utilizes a fast Fourier Overview," AIAA-86-9727, Apt. 1986.
transform algorithm yielded open-loop frequency re-
sponse data that were consistent and repeatable
throughout the flight envelope of the X-29A test air- 7Smith, Rogers E., and Schroeder, Kurt C., "Flight
plane. The data were used as the principal means of Testing the X-29," SETP 30th Symposium Proceed-
monitoring the level of pitch axis dynamic stability ings, 1986, pp. 116-134.
throughout the envelope clearance flights. In addition,

5
kiera, Joseph. "Dynamics and Controls Flight Test- parison of X-29A Flight Data and Simulation Data,"
ing the X-29A Airplane," NASA TM-86803. 1986. AIAA-87-0344, Jan. 1987.

oBosworth, J.T., and West, J.C., "Real-Time Open- l2 Military Specification - Flight Control Systems -
Loop Frequency Response Analysis of Flight Test Design, Installation and Test of Piloted Aircraft, Gen-
Data," AIAA-86-9738, Apr. 1986. eral Specification for - MIL-F-9490D. June 1975.
(Supersedes MIL-F-9490C. Mar. 1964.)
losmith, R.E.. and Sarrafian, S.K., "Effects of
Time Delay on Flying Qualities: An Update," AIAA- l 3 Military Specification - Flying Qualities of Piloted
86-2202, Aug. 1986. Airplanes. MIL-F-8785C, Nov. 1980. (Supersedes
MIL-F-8785B, Aug. 1969.)
Bauer, Jeffrey E., Crawford, David B., Gera,
Joseph, and Andrisani, Dominick, "Real-Time Com-

27 11 2.44 in

I I I------ - a f t 1 in
I-

6247b

fig. I X-29Aairplane.

6
 Amp1itude
ratio, 0 V
d0

~onrorr
Pharo
.. ...
-120
.. ..
.. ..
. -1 40
. -140
10 -- - 1
-Fiioht\ ./ -160
-1 60 Ampiitudo Ph.80
Phase ratio, 0 .. angio,
-180
angle, -180 dB d.g
deg
-200
-10 - 200

-220 -20 -
I I

-240 -30 I IIII I 1 1111111 I I If h - 2 4 0

.1 1 10 100 .1 1 10 100
Frequoncy, radlnc Froquoncy, ndlaoc

Fig. 2 Stability margin requirements for the Fig. 3 Real-time frequency response analysis
X-29A envelope expansion program. procedure.

-.. -.-..
I umrionm
of
*q---"---
notion I
comp.nr8tlon

Up1inkod
ruddor

. FQ. 4 Representative lateraldirectional control system.

7
 -Fllghl data
. Llnoar dala

Downlink

J
/B
Telemetry
decommutatlon
statlon
T h e , sec

\ 3
Pilot
commands
stick ik: Aircraft
response
parametera
-lo 5 10 -
Tlme, sec

X-29A
aerod ynamlc i=AX+Bu
and control y 5 c x + Du
system models
Linear difference
equations

Fig. 5 Real-time comparison of linear simulation and flight.

30
20
c'..',
-Flight data
- - - Llnear data
20
Old galn
-.-margln
- \New
-New galna
galn
-- Old gains

Amplitude ratio, 0
low galn
d0
-10 -
-10 - 7

-20 - ... ... ...

.. ..: I :..
I j l I ' i I I I :.
... ..
111
..
... ... ....
-120 - -120 ...
..
...
..
...
..
-140 - -140

-160 - -160
Phase Phase
angle, -180 - angle, -1 80
deg de!J
-200 - -200

-220 - -220

-240 -240
.1 1 10 100 .1 1 10 100
Frequency, radlsec Frequency, radlmc

Fig. 6 Pitch axis open-loop frequency response Fig. 7 Effect of gain change on flight-measuredBode
Night data compared with linear analysis. plot.

8
 O
'r \ -- Nonlinear rimulatlon data
- - Linear data -Nonlinear rimulatlon data
lot \ 20
. - - Linear data
4
11

ArnDlitude
- O 10
ratio, - l o t Amplitude
dR ratio, 0
d0
-1 0

-30 -20

-40

-180 -

-120 -
I t
6
,
60 -
Phase I

angle, 0. Phare
deg I angle,
I
deg
-60 - I
t
,
a

Frequency, rrdiaec Frequency, radlaec

Fig. 8 Roll axis open-loop frequency response Fig. 9 Yaw axis open-loop frequency response
real-time simulation data compared with linear real-time simulation data compared with linear
analysis. analysis.

9
 -Flight
--- Llnear predictions
1-

Normal
acceleration, 0
g

-1

2-
Pitch
attitude, 0
deg

Pitch
4 r
.
rate, 0
deglsec

I I I I

Angle of
attack, 0
deg
-2 w

input,
in

-2

-
Strake
flap
0
deflection,
deg
-2

Symmetric
flap
deflection,
deg
-1

Canard
deflection,
deg

4 8 12 16
Time, sec 7123

Fig. IO Response of the X-29A to a pitch stick input

compared with linear simulation results.

10
 Roll - Flight
attitude, 0 --- Linear predictions
deg 3 r
1 I J
6 r- Rudder
I\
deflection, i -
deg
3 -
0
Yaw , -
t"i.;-.
1:
rate,
deglsec
0
----
'J
-1

-2 I I I
-3 - I1
b
-6 L I I 1 I "r
Differential
flap
deflection,
deg
Roll

t 1 I 1 I r
- 50

3 r Lateral
acceleration, 0
Angle of g
sideslip, 0
deg
-3 -.4o
0 4 8 12 16 0 4 8 12 16
Time, see Time, sec 7124

Fig. 1 1 Response of the X-29A to a differentialflap and

rudder input conpared with linear simulation results.

11
 t r
-Flight data - Flight data
- - Lower order fit
- - Lower order fit 40
35

Amplitude
ratio, Amplitude

J ratio, 25
dB
4 dB

,15
Ot

2
/
0

0.

50
loo[
-50

Phase Phase
angle, angle, - 100 '

dog dog

-200
- .1 1 5 1 0 10
L - 200.1 1 10

Frequency, radlsec Frequency, radlaec

Fig. 12 Longitudinal lower order system match of the FQ. 13 Lateral lower order system match of the roll
pitch rate due to longitudinal stick transfer function. rate due to lateral stick transfer function.

Finger tip Close trail

formation

-+
43
formation

Simulated terrain Air-to-air

following

+ tracking
b
Fig. 14 Pilot evaluation tasks.

12
 Level Ill
l o t
9

7 -
Level II
6 - A 00 0
Pilot
-0 O A 0 O O A
rating
4 -A 00 000 000 :0 000
Level I
3 - 0 A
2 - A A b

1
t
I I I I I I I

Fig. 15 Pilot rating summary.

13
1 Report No. 2. Government Accession No. 3. Recipient's Catalog No.
N A S A TM-88297

4 Title and Subtitle 5. R q m i Date

D Y N A M I C STABILITY AND HANDLING QUALITIES TESTS August 1987
O N A H I G H L Y AUGMENTED, STATICALLY UNSTABLE 6. Performing Organization Code
R [RPLANE

7 Authorb) 8. Performing Organization Report No.

.Joseph G e r a and John T. Bosworth H-1422

10. Work Unit No.

9 Performing Organization Name and Address RTOP 533-02-81
N A S A Ames R e s e a r c h C e n t e r
Dryden F l i g h t R e s e a r c h F a c i l i t y 11. Contract or Grant No.
P.O. Box 273
E d w a r d s , CA 93523-5000
13. Type of Report and Period Covered
2. Sponsoring Agency Name and Address T e c h n i c a l Memorandum
N a t i o n a l Aeronautics and Space Administration 14. Sponsoring Agmcy Coda
W a s h i n g t o n , DC 20546
I
5 Supplementary Notes
P r e p a r e d a s AIAA-87-2258-CP f o r p r e s e n t a t i o n a t t h e A I M G u i d a n c e , N a v i g a t i o n a n d C o n t r o l
C o n f e r e n c e , Monterey, C a l i f o r n i a , Aug. 1 7 -
1 9 , 1 9 8 7 , a n d f o r p r e s e n t a t i o n a t t h e SFTE 1 8 t h
Annual Symposium, S e p t . 28 - O c t . 2, 1 9 8 7 , Amsterdam.
-
6 Abstract

I n i t i a l envelope c l e a r a n c e and subsequent f l i g h t t e s t i n g o f a

new, f u l l y augmented a i r p l a n e w i t h a n e x t r e m e l y high d e g r e e o f s t a t i c
i n s t a b i l i t y c a n p l a c e u n u s u a l demands o n the f l i g h t t e s t a p p r o a c h .
Previous f l i g h t test experience with t h e s e k i n d s of a i r p l a n e s is
v e r y l i m i t e d or n o n e x i s t e n t . The s a f e a n d e f f i c i e n t f l i g h t t e s t i n g
may b e f u r t h e r c o m p l i c a t e d by a m u l t i p l i c i t y o f c o n t r o l e f f e c t o r s
t h a t may be p r e s e n t o n t h i s class o f a i r p l a n e s . This paper d e s c r i b e s
some n o v e l f l i g h t t e s t a n d a n a l y s i s t e c h n i q u e s i n t h e f l i g h t d y n a m i c s
a n d h a n d l i n g q u a l i t i e s area. These t e c h n i q u e s w e r e u t i l i z e d d u r i n g
t h e i n i t i a l f l i g h t e n v e l o p e c l e a r a n c e o f the X-29A a i r p l a n e a n d were
l a r g e l y r e s p o n s i b l e f o r t h e c o m p l e t i o n of the f l i g h t c o n t r o l s clear-
a n c e program w i t h o u t a n y i n c i d e n t s or s i g n i f i c a n t d e l a y s .

7 K e y Words (Suggested by Authorls)) 18. Distribution Statement

Handling q u a l i t i e s Unclassified - Unlimited
Lower o r d e r e q u i v a l e n t s y s t e m s
Relaxed s t a t i c s t a b i l i t y
S t a b i l i t y margins
X-29A S u b j e c t c a t e g o r y 08

19. Security Classif. (of this report) 20. Security Classif. (of this page) 21. NO. of pages 22. Rice'
Unclassified Unclassified 14 A02

*For s a l e by t h e Nationdl Technical I n f o r m a t i o n S e r v i c e , S p r i n g f i e l d , V i r g i n i d 22161.