Академический Документы
Профессиональный Документы
Культура Документы
Reprinted from
Journal
ofSpacecraft
and
Rockets
Volume31, Number4, Pages550-556
GI__ 4.
A publication of the
American Institute of Aeronautics and Astronautics, Inc.
370 L'Enfant Promenade,SW
Washington, DC20024-25t8
1994
Rarefied-Flow
Shuttle
Aerodynamics
Robert
NASA
Langl_
Research
Flight
Model
C. Blanchard*
Center,
Hampton,
Virginia
23681
and
Kevin
Lockheed
Engineering
T. Larman
* and Christina
and Sciences
Company,
D. Moats
Hampton,
_;
Virginia
23666
A model of the Shuttle Orbiter rarefied-flow aerodynamic force coefficients has been derived from the ratio
of flight acceleration measurements.
The in-situ, low-frequency (<1 lq[z), low-level (,-_1 10 -6 g) acceleration
measurements are made during atmospheric re-entry. The experiment equipment designed and used for this task
is the High Resolution Accelerometer Package (HiRAP), one of the sensor packages in the Orbiter Experiments
Program. To date, 12 HiRAP re-entry mission data sets spanning a period of about 10 years have been processed.
The HiRAP-derived aerodynamics model is described in detail. The model includes normal and axial hypersonic
continuum coefficient equations as functions of angle of attack, body-flap deflection, and devon deflection. Normal
and axial free molecule flow coefficient equations as a function of angle of attack are also presented, along with
flight-derived rarefied-flow transition bridging formulae. Comparisons are made between the aerodynamics model,
data from the latest Orbiter Operational Aerodynamic Design Data Book, applicable computer simulations, and
wind-tunnel data.
Nomenclature
a
C
= acceleration
g
Kn
M
= acceleration
= mass, kg
= 1 x 10-rg
= reference area, 249.9 m 2
micro-g
S
V
X,Y,Z
6
ol
P
Introduction
measurements
HE development of a reusable space vehicle that delivers payloads to orbit in the vertical rocket configuration, and returns
from orbit in a horizontal
aircraft configuration, provides unprecedented opportunities to the flight research community. One unique
flight research area that is assessed regularly
is the aerodynamic
performance of a winged re-entry vehicle traversing the rarefiedflow transition
regime from free-molecule
to hypersonic
continuum, roughly 160-60 km in altitude. Computational fluid dynamics
(CFD) research I was conducted during the early Orbiter development flight tests, but was not mature enough for aerodynamic design
predictions. Extensive wind-tunnel
testing went into the design of
the Shuttle Orbiter, 2 but not under conditions for the rarefied-flow
= Knudsen number;
=Mach
number
=
=
=
=
=
=
9.81 m/s z
i.e., mean-free
path/I 2.058 m
velocity, m/s
viscous parameter
Orbiter body axis system coordinates
deflection angle, deg
angle of attack, deg
atmospheric
density, kg/m 3
regime. There is, however, one set of data used to examine viscous
effects that is shown in this report. These wind-tunnel
data are taken
at conditions closest to those of rarefied flow. However, prior to
the initial Orbiter development
flights, no applicable rarefied-flow
re-entry aerodynamics test data were available.
The solution to this limitation was to adapt empirical expressions
resulting from earlier flight tests of Apollo-like
blunt bodies to a
winged re-entry vehicle)
Figure 1 illustrates the status of knowledge of the rarefied-flow
aerodynamics
at the starl of the Orbiter
Subscripts
A
= axial force
bf
=
=
=
=
=
c
EL
f
i
N
O[
body flap
continuum
elevon
free molecular
flow
the ith direction; i.e., X, Y, Z axes or normal
axis
= normal force
or axial
= angle of attack
Received June 14, 1993; presented as Paper 93-3441 at the AIAA 1 lth
Applied Aerodynamic, Conference, Monterey, CA, Aug. 9-1 I, 1993, revision received Jan. 6, 1994; accepted for publication Jan. 10, 1994. Copyright
(_ 1994 by the American Institute of Aeronautics and Astronautics, Inc. No
copyright is asserted in the United States under Title 17, U.S. Code. The
U.S. Government has a royalty-free license to exercise all rights under the
copyright claimed herein for Governmental purposes. All other rights are
reserved by the copyright owner.
*Senior Research Engineer, Aerothermody namics Branch, Space Systems
Division, Mail Stop 366. Member AIAA.
*Senior Engineer, Space and Systems Engineering Department, 144 Research Drive. Member AIAA.
_:Engineer, Space and Systems Engineering Department, 144 Research
Drive.
55O
BLANCHARD,
1 2_-
LARMAN,
AND
MOATS:
RAREFIED-FLOW
SttUI'rLE
AF.ROI)"/NAMI('S
551
Sources
_//_-_
W indt
,/////'Z_
fl,gh_
_es_s
....
Da_a obtained
-Emp
ricaF formula
computations
'Y_K(_".
Computa;,on
dJrmg
flighls
":_,,
08_::
D04
10-5
10-3
10-1
Knudsen
60
Fig. 1
Orbiter
STS 06
s
101
100
Altilude,
km
aI
.--
STS
STS
51B
STS35
ST S-09
STS
51_
STS-40
61C
CA
__)
_i(M)
. Mass
Pressure
spectromeler
Iransducer
Atmosphere
model
Accelerometer
experiment concepts,
Accelerometer
Package (HiRAP) was designed and installed on the
Orbiter and has been making linear acceleration
measurements
on
the Orbiter missions since STS-6 in April 1983. 5 This capability
provided flight data to construct predictive models. A rarefied-flow
aerodynamic
model has been formulated
by examining
the force
coefficient ratio on multiple flights. The force coefficient ratio was
used because it correlates directly to HiRAP acceleration
measurement ratios.
This simple model derived from HiRAP and Inertial Measurement Unit (IMU) flight data can be used to predict both present-e.g., the Shuttle--and
possibly future Shuttle-like
winged re-entry
vehicle rarefied-flow
aerodynamics;
although
separate research is
now available that addresses generalized
bridging formulae (e.g.,
Ref. 6). The model incorporates
aerodynamic
coefficient derivative
data from the OADDB 3 at the rarefied-flow
transition boundaries,
that is in the hypersonic
continuum
and in the free-molecule
flow
regime. In the hypersonic
continuum
regime, the absolute value of
the individual force coefficients in the model have been adjusted to
values determined
on Mission STS-61C,
when hypersonic
continuum pressure measurements
first became available. These measured
pressures, converted to freestream dynamic pressures, and the measured aerodynamic
accelerations
were used to calculate absolute
orbiter aerodynamic
performance
coefficients.
The Orbiter control
surface aerodynamic
contributions,
i.e., the body-flap
and elevon
coefficients,
are generated from curve fits to the OADDB. However,
to better match the average measured flight data, the elevon axial
force contribution
was reduced by one-third of these curve fit values
for positive deflection
angles. At the other end of the rarefied-flow
transition boundary, the free-molecule
flow force coefficient values
are also generated from curve fits to the OADDB.
The purpose of this report is to present the flight measurements
of the Orbiter CN/CA in the rarefied-flow
regime for 12 missions,
to provide a complete description
of the flight aerodynamics
model
derived from these measurements,
and to show the model capabilities and limitations
to predict the aerodynamic
force coefficients of
the Orbiter in the rarefied-flow
regime. To demonstrate
the details,
the model is used to predict the force coefficient ratios of two recent
flights, STS-35 and STS-40. In addition, the model is compared with
available wind-tunnel
data and recent CFD calculations.
and
STS-41C
STS-08
Equipment
STS07
140180
aA
Fig. 2
STS-61A
number
rarefied-flow aerodynamics
STS-41B
Experiment
Description
HiRAP
contains
three
orthogonal,
pendulous,
gas-damped,
micro-g resolution
accelerometers
mounted
on the Orbiter in the
wing box. The input axes of the accelerometer
sensors are co-aligned
with the Orbiter body axes. The HiRAP is designed to measure the
Orbiter aerodynamic
acceleration
signal during re-entry. In particular, HiRAP measures the low-frequency
(< 1 Hz), low-level acceleration (i.e., micro-g sensitivity)
signals during the re-entry from the
free-molecule
flow region to the hypersonic
continuum
transition.
During the unpowered
gliding re-entry of the Orbiter, the normalto-axial aerodynamic
acceleration
signal ratio is directly related to
the normal-to-axial
force coefficient
ratio, as shown in Fig. 2.
The total acceleration
signals
measured
by the HiRAP are
converted
into calibrated
aerodynamic
acceleration
data sets and
merged with mission-specific
trajectory
state vector data. These
merged aerodynamic
data sets are then further processed to determine Orbiter aerodynamic
coefficients.
The equation for this process, shown in Fig. 2, is a particular arrangement
of the classical
aerodynamic
force equation:
ai
Ci --
(1)
(I/2pV2)S/M
where Ci is the aerodynamic
force coefficient in the ith direction,
ai is the corresponding
acceleration,
and S/M is the vehicle area
to mass ratio. In general, the dynamic pressure
(l/2pV 2) can be
obtained from air data measurements,
such as pressure transducers
or mass spectrometer
measurements,
7 or from atmospheric
models.
For this report, pressure data taken in the hypersonic
region of flight
from STS-61C were included in the model development,
s
To calculate the coefficient
Ci in Eq. (1), all of the terms on the
right of the equation must be known. The HiRAP instrument
measures acceleration
ai along each Orbiter axis. The mass M and the
reference area S (249.9 m 2) are known and velocity V is determined
as a function of flight time from a best estimated trajectory
(BET)
process.9- i i Because no measurements
of the density were available
for most of the earlier flights, a statistical approach was developed
to determine the aerodynamic
coefficient ratio using the ratio of accelerometer
data. This technique
is described in the literature along
with the results for all of the earlier flights. 9'12-17
Flight
Data
Results
The fundamental principle behind making aerodynamic acceleration measurements during re-entry is to remove the nonaerodynamic
acceleration
inputs. In unpowered gliding flight, the predominant
nonaerodynamic
forces on the Orbiter are thrust firings, auxiliary
power units (APUs), and rotationally
induced linear accelerations
(the sensors are not at the center of gravity). The thrust firings are
measured acceleration impulses induced when the reaction control
system (RCS) is used to control the attitude of the Orbiter Vehicle during reentry. The three APU exhaust ports are located just in
front of the vertical tail and induce a pulsing acceleration signal in
the Z-axis direction. The flight data are transformed into aerodynamic acceleration data using a rigorous, detailed process reported
earlier. Is.19 After this process, the ratio of aerodynamic
acceleration is formed, which is also the ratio of the aerodynamic
force
coefficients.
A measure of aerodynamic
performance
is the ratio of the normal
to axial acceleration
measurement.
This ratio corresponds
to force
coefficient
ratios, CN/Ca.
Figure 3 presents the force coefficient
ratio data during re-entry for all 12 currently
available
missions.
The re-entry data set consists of about 11.6 hours of flight measurements.
It is from this compendium
of data that the aerodynamic
model has been determined.
For the purposes of this analysis, the
acceleration
data sets consist of time and X- and Z-axis acceleration
552
BLANCHARD,
LARMAN,
AND
MOATS:
RAREFIED-FLOW
SHUTTLE
AERODYNAMICS
',:[
_A
wind
i
_Z
5el
Aerodynamic
coefficients
Fig. 4
measurements.
Y-axis
The HiRAP
sensor
output,
6hi
"I_ 8 /
Aerodynamic
control surfaces
=L..,...,...,...,...,
Sign conventions.
instrument
..,
the
0.060
Rarefied-Flow
Aerodynamics
C,
Cat.= = 5.86689
x 10-2or - 0.839782
6.72027
10-sa 2 + 3.32044
10-70t
x 10-3or --0.0086314
3 -
(Iolal)
o_
36
38
40
Angk_of-Am,,_,
OADOB (basK;)
42
44
46
continuum.
X 10 -4 4 1.92931
X 10-3t_bf 43.6029
X 10-5_;_f
(4)
Cat.b/ = 2.39178
+ 3.03937
x 10 -4 44.81862
x 10-58_/
10-43by
-- 3.500869
x 10-TS2f
(5)
Elevons
deg < 8by < 15 deg:
1.00333
X 10 -3 + 5.76021
x 10-38EL
and axial
x 10-sa 2 + 5.23808
x
I
34
CN,..EL =
normal
0.045
0040
For -15
CNc.= = -9.25704
_OADDB
Model
+ 1.18538
(2)
CAc._. =
1.3981
+ 8.6081
(6)
X 10-45_L
x 10 -3 4 1.55864
10-38EL
(7)
x 10-3_2 L
(3)
if Cat.B_ >_ 0, then Ca,._.
= I/3CAc,_
(8)
functions
compared with the
flap and elevons, respectively.
The flight model varies from the OADDB for positive elevon angles
in Fig. 7 because the HiRAP flight data indicated that the elevon
effectiveness
was less than the OADDB predicted.
Then,
the total hypersonic
continuum
coefficients
are the
following:
CNc
CNc,_
CNc,
bf
CNc,
(9)
EL
Free-Molecule Flow
The formulae for the free-molecule
cients as a function of ,y are:
Ct:._ = -7.16528
+9.18422
40.0164864a
x 10-6c_ 3 4 9.66197
x 10-3ct + 1.58739
CA/._ = --1.17117
flow normal
x 10-4a 2
x 10 -3
x 10-%e3 4 5.92205
4 0.751105
(10)
(11)
x 10-4t_ 2
(12)
BLANCHARD,
LARMAN,
AND
MOATS:
RAREFIED-FLOW
SHUlq'LE
AERODYNAMICS
553
0.08
Equations ( 11 ) and (12) are for the Shuttle (payload doors closed)
at 0 < _t < 60 deg. A comparison
of these functions with the data
from the OADDB is shown in Fig. 8. The body-flap
and elevons
004
contributions;
that is, CN/ b:, CAf,
bf,
CNf. EL, and
CAr. EL,in the freemolecule flow regime are undetermined
and are currently
set to
Cn=
0 02
zero,
0.06
000
_.02
................
, _ ....
J ....
00'0
Q025I
0 020
0015
CA _
FLIGHT
flow coefficients
arc:
MODEL
0 O05
CNf
CNf,
CA/
CAf.a
Bridging Formulae
The method
to
free-molecule
flow
formulaeS.a.12A3.16:
"Jr- CNf,
bf
-]-
CNf,
EL
(13)
"-_
bf
-_
CAr,
EL
(14)
Caf,
bridge
the
regime
includes
hypersonic
the
continuum
following
to
the
"bridging"
0000
-0005
....
' ....
-10
-t5
t ....
-5
, ....
0
, ....
5
, ....
10
, ....
15
, ....
20
CN = exp [-0.29981(1.3849
=
25
- log,0 r,)
if
otherwise
Cu
L712s]
(15)
= 1.0
oc = 40 deg.
Ca = exp [-0.2262(1.2042
0T5
0 05
CNeu
FLIGHT
MODE
0 10
0CO
__D
_.05
where
4:1.10
....
J ....
, ....
' ....
' ....
I ....
, ....
, ....
0.04
the normalized
normal
(16)
1.0
use Knudsen
CN = CNc + (Cur
- Cuc)CN
(17)
CA
__
(18)
O
=
CA
Dr_ (CAf
CAc)CA
O.OZ
0.06 I
O.OO
-0.02
_OOEL
..............
-20
-1S
* .........................
-5
0
-10
10
15
Fig. 7
2.SOl
2.000
FLIGHT
ot = 40 deg.
1.ooo
Trajectory
Data
Re-entry trajectory data from each mission are required for the
development and the calculation of the rarefied-flow aerodynamics
model. These data are available after each flight, and they represent
the "as flown" trajectory. 9-H Other information is also available,
such as the as flown Orbiter configuration. The angle-of-attack
and
control surface data for all missions under consideration are obtained
from this source and presented in Figs. 9a-9c. The Orbiter average
angle of attack is seen to be 41 deg until an altitude of about 110
km is reached, when it decreases to about 40 deg. During the early
Orbiter flights, the initial body-flap and elevon settings were varied,
until an apparent optimum setting was realized. Currently, all flights
use -4.5 and -2.7 deg as the settings for the body flap and elevons,
respectively. The mass of the Orbiter at entry interface (121 krn)
MODEL
Re-Entry
Z0
CN
OAOOfl "''_o
C,._
if
otherwise
OADDB
0.500
_.500
the
1.800
FLIGHT
MODEL
1.40_
1.600
CA
1000
1 200
0.800
0.600
....
_ ....
10
i ....
20
, ....
30
Angle-of.Attack,
, ....
40
, ....
50
f
60
dog.
molecular flow.
Model
Comparisons
554
Data
!.,
,o
ob_inedduringflights:
3E
STS-06
STS-41B
STS
STS-07
STS-41C
STS-61C
STS-08
STS-51B
STS-35
STS*09
STS-51F
STS-40
,I
Aeroctynam=c
model
STS-07
ST$-41C
results
for flights:
STS-08
STS-51B
STS-35
$TS-09
STS-StF
STS40
STS-06
$TS-41B
61A
20
jr,
/<
C_r'CA
25t
1_15
$TSq_tC
STS-61A
.,,,.
<
15
35
60
80
100
120
140
Altilude,
160
180
200
: :': ::
km
, ..-..':._
10 _
104
10")
10-_
KnuCisen
Data
er_velo_e
dunng
Fig. 10
isons.
missions:
$TS.06
$TS-07
STS-41B
STS-41C
STS-Oe
STS-Oe
STS-51B
C_/CA
10 _
Nom0er
10 o
10 _
fief-MAC}
25
p_o
eo
10o
120
140
16J
leo
20
Aerodynamics
o-=40.
lhnvolOpe Ounng
D_a
STS-51F
STS-61A
_=0
Model
, _=0
m_sslons
ST$-35
STS-40
STS-S_C
i:o!
Wln_3
C
"156_
80
100
120
140
Allauchl,
160
,'C_,
5
1105
T_nnel
Fla_
10 4
10 .4
10
Knudse
Fig. II
STS-06
e nvldop_l
STS41B
= -
40
180
l,a'n
Dill
P_a_e
__s'0
10 z
t04
Number
10 _
10'
(reL=MAC)
data.
ot mi=_tons:
STS-07
$TSJ_lC
10 _
I0*
...........
12 _
0ata
cotame0
$TS.61A
4
during
m4ss=ons:
8000
STS-4_
$TS-61C
--$TS-40
>_
-4
8 _
,ooo
I/
6000
----5TS-35
.......
STS-51F
STS.35
s%o'''io"
%o'
,io'"
AJ_tude,
-12
6O
8O
100
120
A_itu_,
140
160
180
Angle-of-attack
Fig. 12
Orbiter
flight conditions.
km
_m
t;o4o''',_o'''_;o
b) Velocity profiles
of Kn.
The HiRAP
aerodynamics
BLANCHARD,
LARMAN,
AND
MOATS:
RAREFIED-FLOW
SHUTTLE
555
AERODYNAMICS
44
1.5
..---
----.
38
36
_
<
32
30
28
a) STS-35
a) STS-35
44
1.2
4o
38
0.8
t_
36
C,
_, 32
60
60
80
100
120
AItilU0a,
140
160
1g(]
B0
100
120
200
Altitude,
140
160
180
200
km
km
b) STS-40
b) STS-40
Fig. 13
Orbiter angle-of-attack
Fig. 16
profiles.
Flight aerodynamic
25
20
CjCA
_5
8_,deg, 2
-i
/,
a) Body flap
25
, j_
STS-40
Data
10
i "
F]i_l-lt
120
Altitude,
8EL. 0cO.
-i
60
M_OI
140
160
km
100
120
AJtitude,
14(]
km
160
180
200
b) Elevon
Fig. 14
8w. deg
a) Body flap
2
0
8EL, dog
-2
-4
-6
-8
60
80
100
120
Altil_do,
140
160
180
200
km
b) Elevon
Fig. 15
12 shows the Orbiter velocity and correof altitude for STS-35 and STS-40. The
556
BLANCHARD,
below
90
in order
km, when
they
to establish
are activated,
longitudinal
mostly
section
cient
results
attack
effects
angle-of-attack
altitudes
produces
variations
(see
Fig.
STS-35
and
ences
small,
performance
in CN at both
Because
coeffi-
low
of scale
seen
and high
selection,
at low
altitudes,
altitudes.
two graphs,
of acceleration
STS-40
between
tess than
seen
comparisons).
at high
17 shows
the ratio
data
in the
variations
are discernible
Figure
ments
of
aerodynamic
(the
IMU
each of
from
data
are
which
contains
the IMU
and
below
97
measureHiRAP
on
km).
the model
predictions
and the flight
1% throughout
the entire
regime.
The
data
differare
very
Summary
The
OEX
program
provides
a unique
peated
Shuttle Orbiter
re-entries
and aerodynamic
measurements.
the
HiRAP
experiment,
low-acceleration
rarefied-flow
transition
flight
a flight
sions
ation
region
biter
it may
make
for
regime.
during
the
to simulate.
These
an
aerodynamic
be used
performance
Orbiter
the
has
been
in the
re-entry
twelve
engineering
as an initial
engineering
other
similar
This
model
and recent
diaerodynamic
of the
re-entry.
prediction
winged
mis-
accelera concise
prediction
during
are in
HiRAP
developed.
performance
of any
re-
in the
measurements
Shuttle
With
model
providing
the
low-frequency,
measurements
is useful
dynamic
to
acceleration
aerodynamics
rarefied-flow
dition,
developed
about
10 years,
a unique
flight aerodynamic
has been accumulated.
From this database,
compares
well
rect simulation
model
regime
difficult
spanning
database
rarefied-flow
was
transition
to use
aerodynamic
flight
opportunity
Or-
In ad-
of the aero-
re-entry
vehicle.
References
i Moss,
Simulation
of Transitional
Flow for
Aerodynamic
Design
RAREFIED-FLOW
SHUTI'LE
5Blanchard,
direction,
STS-40
flight trajectory
flight model
described
are easily
13 for
the CA angle-of-attack
but
the
shown
in Fig.
that influence
in the pitch
control.
An application
of the STS-35
and
with the rarefied-flow
aerodynamics
preceding
international,
Space Division, Downy, CA, STS 85-0118 CHG 3, Sept. 199 I, pp. 5.2.1.1.19, 48, 49, 109, 148, 149, 168, 169, 188, 189, 5.2.1.1.2-9,
48, 49, 107, 146,
147, 166, 167, 186, 187.
4Schaaf, S. A., and Chambre, P. L., "'Flow of Rarefied
Gases," Fundamentals of Gas Dynamics,
edited by Howard W. Emmons,
Princeton Univ.
Press, Princeton, NL 1958, pp. 687-738.
AERODYNAMICS
Accelerometer
Package Experiment:
Spacecraft
and Rockets, Vol. 22, No.
6potter,
J. L., "Procedure
for
Dimensional
Bodies in Transitional
J. E,"Shuttle
Orbiter
High
Resolution
Preliminary
Flight Results," Journal of
4, 1985, pp. 474--480.
Estimating
Aerodynamics
of ThreeFlow," Rarefied
Gas Dynamics:
The-
Measurements
on STS-6
Aerodynamics
and Upper
Atmosphere
Density
from Multiple
Orbiter Flight Measurements"
AIAA
Paper 86-0269, Jan. 1986.
14Blanchard,
R C., "Rarefied
Flow Lift-to-Drag
Measurements
of the
Shuttle Orbiter," International
Council of Aeronautical
Sciences
Paper 862.10.2, Sept. 1986.
IS B lanchard, R. C., and Larman,
K. %, "Rarefied
Aerodynamics
and Up-
per Atmosphere
Flight Results from the Orbiter High Resolution Accelerometer Package Experiment,"
AIAA Paper 87-2366, Aug. 1987.
16Blanchard,
R. C., and Buck, G. M.. "Rarefied-Flow
Aerodynamics
and Thermosphere
Structure
from Shuttle Flight Measurements,"
Journal
of Spacecraft
and Rockets, Vol. 23, No. 1, 1986, pp. 18-24.
17,Blanchard,
R. C., and Hinson, E. W., "Hypersonic
Rarefied-Flow
Aerodynamics
Inferred from Shuttle Orbiter Acceleration
Measurements,"
International
Congress
on Hypersonic
Aerodynamics,
Paper No. 7, Sept.
1989.
ISThompson,
J. M., Russell, J. W., and Blanchard,
R. C., "Methods
for
Extracting
Aerodynamic
Accelerations
from Orbiter High Resolution
Accelerometer
Package
Flight Data;' AIAA Paper 87-2365, Aug. 1987.
19Blanchard,
R. C., Larman,
K. T., and Moats, C. D., "Flight Calibration Assessment
Jan. 1993.
2U.S. Standard
Oct. 1976.
of HiRAP
Atmosphere,
Accelerometer
1976,
NOAA,
Data,"
AIAA
NASA,
Paper
93-0836,
J. N., "Hypersonic
Rarefied Flow About
Vol. 29, No. 8, 1991, pp. 1250-1258.
Plates