JOURNAL OF SPACECRAFT AND ROCKETS

Vol. 42, No. 5, September–October 2005

Prediction of the Pitch-Damping Coefficients

Using Sacks’s Relations

Paul Weinacht∗ and James E. Danberg†

U.S. Army Research Laboratory, Aberdeen Proving Ground, Maryland 21005

Throughout its development, slender-body theory has been generalized to predict a large variety of aerodynamic
coefficients for a wide class of flight bodies. For most applications, slender-body theory provides only a qualitative
predictive capability. There is, however, a set of slender-body relationships that have been previously derived by
Sacks that allow the individual pitch-damping coefficients and the pitch-damping coefficient sums to be related
to each other. Until recently, it has been difficult to assess the accuracy of these relationships because of the lack
Downloaded by UNIVERSITY OF CALIFORNIA - DAVIS on February 13, 2015 | http://arc.aiaa.org | DOI: 10.2514/1.10718

of high-quality pitch-damping data or the lack of a higher-order theory. The current work applies a recently
developed computational fluid dynamics capability for predicting all three pitch-damping coefficients. From this
analysis, the accuracy of these relationships has been assessed and their engineering significance demonstrated.
One important result is that the pitch-damping relations developed by Sacks allow the individual pitch-damping
coefficients to be determined from the pitch-damping coefficient sum with a high degree of accuracy.

Nomenclature p = pressure
a = speed of sound q = transverse angular rate of body as used in flight
Cm = pitching-moment coefficient, M̄/ 12 ρ∞ V 2 Sref D mechanics equations
Cm q = pitch-damping-moment coefficient slope due to qT = transposed vector of dependent variables as used
body transverse angular rate, ∂Cm /∂(q D/V ) in Navier–Stokes equations
Cm q + Cm α̇ = pitch-damping-moment coefficient sum Re = Reynolds number, a∞ ρ∞ D/µ∞
Cm α = pitching-moment coefficient slope with respect R0 = helix radius
to angle of attack, ∂Cm /∂α r = radial coordinate
Cm α̇ = pitch-damping-moment coefficient slope due to Sref = reference area, π D 2 /4
angular rate associated with angle of attack, Ŝ = viscous flux vector
∂Cm /∂(α̇ D/V ) Ŝc = viscous terms due to cylindrical
C2m α = pitching-second-moment coefficient coordinate formulation
angle-of-attack slope, ∂C2m /∂α U, V, W = contravariant velocity components
CN = normal-force coefficient, F̄/ 12 ρ∞ V 2 Sref u, v, w = velocity components in the x, φ, r directions
C Nq = pitch-damping-force slope due to body transverse V = freestream velocity
angular rate ∂C N /∂(q D /V ) X e , Ye , Z e = Earth-fixed coordinates
C Nq + C Nα̇ = pitch-damping-force coefficient sum x = axial location along body from nose
C Nα = normal-force coefficient slope with respect xcg = axial location of center of gravity from nose
to angle of attack ∂C N /∂α x̄ = integration variable associated with axial
C Nα̇ = pitch-damping-force coefficient slope due to location along body from nose
angular rate associated with angle of attack α = angle of attack
∂C N /∂(α̇ D/V ) α̇ = angular rate associated with angle of attack
Cn = side-moment coefficient γ = cosine of total angle of attack
D = reference diameter δ = sine of total angle of attack
Ê, F̂, Ĝ = flux vectors in transformed coordinates  = deviation or error in Sacks’s relations for force
e = total energy per unit volume ¯ = deviation or error in Sacks’s relations for moment
F̄ = force µ = viscosity
Ĥ = source term in Navier–Stokes equations due to ξ, η, ζ = transformed coordinates in the
rotating coordinate frame Navier–Stokes equations
Ĥc = source term in Navier–Stokes equations due to ρ = density
cylindrical coordinates φ = circumferential coordinate
L = body length  = angular rate associated with coning
M̄ = moment and helical motions
ω = angular velocity about longitudinal axis
Presented as Paper 2003-5467 at the AIAA Atmospheric Flight Mechanics Subscript
Conference, Austin, TX, 11–14 August 2003; received 7 May 2004; revision
received 24 November 2004; accepted for publication 29 November 2004. ∞ = quantity evaluated at freestream conditions
This material is declared a work of the U.S. Government and is not sub-
ject to copyright protection in the United States. Copies of this paper may
be made for personal or internal use, on condition that the copier pay the Introduction
$10.00 per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood
Drive, Danvers, MA 01923; include the code 0022-4650/05 $10.00 in cor-
respondence with the CCC.
∗ Aerospace Engineer, Aerodynamics Branch, Weapons and Materials Re-
T HE pitch-damping-moment coefficients Cm q (due to body
transverse angular rate) and Cm α̇ (due to angular rate associated
with angle of attack) play an important role in the performance and
search Directorate. Associate Fellow AIAA. dynamic stability of flight bodies. The pitch-damping-moment co-
† ARL Guest Researcher; also Professor Emeritus, Department of Me- efficient sum Cm q + Cm α̇ is of most practical importance, although
chanical Engineering, University of Delaware, Newark, DE 19716. Asso- the individual damping coefficients are often required in aerody-
ciate Fellow AIAA. namic analyses. Throughout the past several decades, a variety of
865
 866 WEINACHT AND DANBERG

techniques and theories have been developed for predicting the pitch-damping-force (or moment) coefficients can be determined
pitch-damping coefficients.1−10 These techniques vary in their ease (C Nq , C Nα̇ or [C Nq + C Nα̇ ]), the other two damping coefficients can
of use as well as their ability to accurately predict the pitch-damping be obtained using simple closed-form expressions. This, of course,
coefficients. assumes that the pitching-moment coefficient Cm α (first moment of
During the course of its development, slender-body theory was the normal force) and the second moment of normal force C2m α can
generalized to predict a large variety of aerodynamic coefficients be obtained as well. Both of these coefficients can be obtained if the
including the pitch-damping coefficients.1,2 In general, direct ap- normal-force distribution is known as shown in Eqs. (10) and (11):
plication of these methods provides only qualitative results for the  x
aerodynamic coefficients. However, elements of slender-body the- (xcg − x̄) dC Nα
Cm α = dx̄ (10)
ory have been incorporated into current engineering methods. These 0 D dx̄
methods3−5 have evolved considerably, although their implementa-  x
tion is fairly complex. Apart from implementation issues, modern (xcg − x̄)2 dC Nα
engineering methods, once embodied into a computer code, are rel- C2m α = dx̄ (11)
0 D2 dx̄
atively easy to use and provide fast and reasonably accurate aero-
dynamic predictions for a large variety of flight geometries. Predictive methods for the static normal force and static pitching
From slender-body theory, some important relationships between moment are well established, even for fast-design methods.
the various aerodynamic coefficients can be derived, although these Because Sacks’s relations were derived using simple approximate
Downloaded by UNIVERSITY OF CALIFORNIA - DAVIS on February 13, 2015 | http://arc.aiaa.org | DOI: 10.2514/1.10718

relationships only hold rigorously within the context of theories theories, it remains to be shown whether the validity of the relations
from which they were obtained. Bryson2 derived the relatively well- shown in Eqs. (4–9) exists only within the context of the theo-
known slender-body result that relates the pitch-damping-moment ries from which they were derived or whether they are universally
coefficient sum to the normal-force coefficient, shown in Eq. (1): valid for slender bodies, or perhaps, more importantly, whether they
Cm q + Cm α̇ = −[(L − xcg )/D]2 C Nα (1) are of general engineering significance. The preceding relations in
Eqs. (4–9) differ somewhat from the results in Eqs. (1) and (2) be-
cause they relate the pitch-damping coefficients to each other rather
Sacks,1 using the Blasius method for calculating the forces and
than purely to the static normal-force coefficient. This suggests that
moments on slender bodies from the crossflow potential, found that
Sacks’s relations might more properly represent the physics com-
many of the aerodynamic coefficients were related to each other.
pared to their counterparts that directly relate the damping coeffi-
Sacks obtained expressions that directly related the individual pitch-
cients to the normal-force coefficient. Because of this, there is reason
damping coefficients to the normal-force coefficient, such as the
to investigate the validity and accuracy of Sacks’s relations.
pitch-damping-force coefficient shown in Eq. (2). These expressions
Recently, a computational approach for predicting all three of
have a form similar to Bryson’s result shown in Eq. (1):
the pitch-damping coefficients has been developed.8,9 The approach
C Nq = [(L − xcg )/D]C Nα (2) solves the three-dimensional thin-layer Navier–Stokes equations for
three different imposed motions that allow the three pitch-damping
In practice, these relations that directly relate the damping coef- coefficients to be predicted independently. The pitch-damping-force
ficients to the normal-force coefficient do not perform particularly and -moment coefficient sums are determined from the computation
well, even when the slender-body evaluation of the normal-force co- of a body undergoing an imposed coning motion.8 The individ-
efficient is replaced with a more accurate evaluation of the normal- ual pitch-damping coefficients are obtained from computations of
force coefficient from sources such as experimental measurement a body undergoing two specific types of imposed helical motions.9
or computational fluid dynamics (CFD).6 However, these relations Each of these motions is described in the following sections. One
can be combined with empirical corrections to yield more reliable of the key components of this method is that steady flow techniques
results.6,7 can be employed to predict aerodynamic derivatives normally asso-
Sacks also found expressions that related the individual damp- ciated with time-dependent motions.
ing coefficients to each other, including the following relationship Using the computational approach, the validity, accuracy, and ap-
between the pitch-damping-force coefficients: plicability of Sacks’s relations are assessed in the current paper. A
brief description of the helical and coning motions used to generate
C Nq = C Nα̇ − Cm α (3) the pitch-damping coefficients of interest are presented in the follow-
ing two sections, followed by a section discussing the computational
Sacks’s explicitly derived relation shown in Eq. (3) can be easily approach. Results are then presented examining the performance of
generalized using his theory to the individual pitch-damping-force Sacks’s relations for two axisymmetric body geometries.
and -moment coefficients and the pitch-damping-force and -moment
sums as shown in Eqs. (4–9). For the purposes of this paper, these Helical Motions and the Individual
relationships will be referred to as Sacks’s relations: Pitch-Damping Coefficients
C Nq = C Nα̇ − Cm α (4) Forces and moments related to the two individual rates q and α̇
can be excited independently using two types of motion in which
Cm q = Cm α̇ − C2m α (5) the center of gravity of the flight vehicle traverses a helical flight

 path. The first motion requires the vehicle’s longitudinal axis to be
C Nq + C Nα̇ = 2C Nα̇ − Cm α (6) oriented in the same direction as the center of rotation of the helix but
displaced by a constant distance. Figure 1 shows a three-dimensional

 view of the motion.
Cm q + Cm α̇ = 2Cm α̇ − C2m α (7)


C Nq + C Nα̇ = 2C Nq + Cm α (8)


Cm q + Cm α̇ = 2Cm q + C2m α (9)

Sacks’s contribution is the recognition that these relationships exist,

although the same relations are implicitly contained in other theories
such as that presented by Bryson.2
For slender bodies, these relations are independent of configu-
ration and are applicable to both winged and wingless bodies. The
importance of these relationships is that if just one of the three Fig. 1 Helical motion with nonzero α̇ and zero q.
 WEINACHT AND DANBERG 867

Fig. 2 Helical motion with zero α̇

and nonzero q.
Fig. 3 Schematic of coning motion.

freestream velocity vector. This is shown schematically in Fig. 3. In

the context of this paper, coning motion also requires the c.g. to tra-
verse a rectilinear path at constant velocity such that the freestream
velocity vector has a fixed orientation with the inertial frame.
This particular motion produces no rotation of the body-fixed Relative to a nonrolling coordinate frame, both the transverse an-
nonrolling coordinate frame relative to an Earth-fixed coordinate gular rate of the body and the angular rate associated with angle
frame, and hence the transverse angular velocity of the body is zero. of attack vary in a periodic manner, thereby exciting the aerody-
The angle of attack and its angular rate vary continuously, produc- namic forces and moments associated with both of the individual
ing moment components associated with the coefficients Cm α and pitch-damping coefficients. Here, a specific form of coning motion,
Downloaded by UNIVERSITY OF CALIFORNIA - DAVIS on February 13, 2015 | http://arc.aiaa.org | DOI: 10.2514/1.10718

Cm α̇ , respectively. This motion is referred to as q = 0 helical motion described as zero-spin coning motion, is employed. In zero-spin
because the angular rates associated with the damping coefficient coning motion, the total angular velocity of the body along the lon-
Cm q are zero. gitudinal axis (the spin rate) is zero. By imposing zero spin rate on
For the second motion, the longitudinal axis of the flight vehicle the body, the contributions from the Magnus forces and moments
remains tangent to the helical flight path at each point along the are eliminated.
trajectory. Figure 2 shows a three-dimensional view of this motion. The time dependency is removed by transforming the body-fixed
The angle of attack of the incident airstream is zero because both nonrolling coordinate frame to an orthogonal right-handed coordi-
the longitudinal axis of the body and the freestream velocity vector nate system that has its x axis aligned with the longitudinal axis
are tangent to the flight path. The resulting yawing rate is also zero of the body and its z axis in the pitch plane of the body. Within
because the angle of attack is constant. The angular orientation of this transformed coordinate frame, in-plane moment Cm and side-
the flight body changes continuously with respect to an Earth-fixed moment Cn coefficients have the following form.
reference frame, producing a nonzero transverse angular rate. As a Zero-spin coning motion:
result, moment components associated with the damping-moment

coefficient Cm q are produced. This motion is referred to as α̇ = 0 Cm + iCn = iδ(D/V ) Cm q + γ Cm α̇ + Cm α δ (14)
helical motion because the angular rates associated with the damping
coefficient Cm α̇ are zero. Here, the side moment is proportional to the pitch-damping-moment
For each of the helical motions, the transverse aerodynamic mo- coefficient sum and varies linearly with the coning rate  and sine of
ment in the nonrolling frame will be periodic in time, which also the total angle of attack δ. For small angles of attack, the cosine of the
indicates that the flowfield will be periodic in time when viewed total angle of attack γ can be assumed to be one. A more complete
from the nonrolling coordinate frame. The time dependency is re- discussion of the coning motion, the transformed coordinate system
moved by transforming to an orthogonal right-handed coordinate and the transverse force and moment equations is contained in Ref. 8.
system that has its x axis aligned with the longitudinal axis of the The pitch-damping coefficient sum can also be determined by
body and its z axis along a line between the body c.g. and the axis simply adding the individual damping coefficients. In practice, there
of rotation of the helix. is very little difference in the two results.9 However, for the cur-
For each of the helical motions already described, the spin rate rent study, directly predicting the pitch-damping coefficient sum
of the body has not been defined. To eliminate any contributions using coning motion provides an alternative prediction of the pitch-
to the aerodynamic forces and moments from the Magnus forces damping sum and additional confirmation of the predictions of the
and moments, the spin rate is fixed to zero (see Refs. 8 and 9 for individual coefficients.
details). The resulting in-plane moment Cm and side moment Cn
coefficients in the transformed coordinate system for both types of Computational Technique
helical motions are shown in Eqs. (12) and (13). In the preceding sections, several types of steady motion that
Zero-spin q = 0 helical motion: produce aerodynamic forces and moments from which the vari-
ous pitch-damping coefficients can be obtained were presented.
Cm + iCn = −Cm α̇ (D/V )(R0 /V ) + iCm α (R0 /V ) (12) One unique feature of these motions is that they are steady mo-
tions. The advantage of a steady motion over an unsteady motion
Zero-spin α̇ = 0 helical motion: is that a potentially time-independent flowfield can be produced by
a steady motion, permitting analysis using steady flow CFD tech-
Cm + iCn = Cm q (D/V )(R0 /V ) (13) niques. Such techniques can be computationally less expensive than
time-dependent CFD approaches. To fully exploit the steady char-
Here,  is the angular velocity of the body about the helix axis, R0 acter of the flow, special body-fixed coordinate systems have been
is the perpendicular distance between the helix axis and the body employed to capture the steady flowfield. One feature of these coor-
c.g., and V is the total linear velocity of the c.g. Similar expressions dinate frames is that they are rotating at a constant rate with respect
for the individual damping-force coefficients can be developed us- to an inertial frame. Because of this, the governing equations of
ing the same approach as applied for the moment coefficients. A fluid motion must be modified to take into account the centrifu-
more complete discussion of the helical motions, the transformed gal and Coriolis force terms associated with the noninertial rotating
coordinate system, and the transverse force and moment equations frame.
is contained in Ref. 9. The steady thin-layer Navier–Stokes equations are shown in
Eq. (15):
Coning Motion and the Pitch-Damping Sum  
To predict the pitch-damping coefficient sum, coning motion is ∂ Ê ∂ F̂ ∂ Ĝ 1 ∂ Ŝ
+ + + Ĥc + Ĥ = + Ŝc (15)
employed. In steady coning motion, the longitudinal axis of the ∂ξ ∂η ∂ζ Re ∂ζ
flight body performs a rotation at a constant angular velocity about
a line parallel to the freestream velocity vector and coincident with The inviscid flux vectors Ê, F̂, and Ĝ; the viscous term Ŝ; the in-
the body’s c.g., while oriented at a constant angle with respect to the viscid and viscous source terms due to the cylindrical coordinate
 868 WEINACHT AND DANBERG

formulation Ĥc and Ŝc ; and the source term containing the Coriolis
and centrifugal force terms that result from the rotating coordinate
frame Ĥ are functions of the dependent variables represented by
the vector qT = (ρ, ρu, ρv, ρw, e), where u, v, and w are the ve-
locity components in axial, circumferential, and normal directions.
The inviscid flux vectors and the source term due to the rotating
coordinate frame are shown in Eq. (16). Details of the thin-layer
viscous term and the source terms due to the cylindrical coordinate
formulation are available in Refs. 11 and 12, respectively:
   
ρU ρV
ρuU + ξx p  ρuV + ηx p 
1  1
 

Ê =  ρvU   F̂ = ρvV + ηφ p/r 
J  J  
ρwU   ρwV + ηr p 
(e + p)U (e + p)V Fig. 4 Schematic of Army–Navy Spinner Rocket configuration.
Downloaded by UNIVERSITY OF CALIFORNIA - DAVIS on February 13, 2015 | http://arc.aiaa.org | DOI: 10.2514/1.10718

   
ρW 0 afterbody. The accuracy of Sacks’s relations is examined for each
 ρuW + ζx p   ρ fx  of these body geometries in supersonic flight.
1  1 
Ĝ = ρvW + ζφ p/r 
 Ĥ =  ρ f φ

 Ogive-Cylinder Results
J  J 
ρwW + ζr p  ρ fr  The computational approach was applied to a secant-ogive cylin-
(e + p)W ρu f x + ρv f φ + ρw fr der body [designated as the Army–Navy Spinner Rocket (ANSR)]
shown in Fig. 4. Sample results for Mach 2.5, sea-level atmospheric
(16) conditions (Re D = 1.1 × 106 ) are shown as an example, although
other supersonic velocities, Reynolds numbers, and geometries were
The pressure p can be related to the dependent variables by ap- considered during the course of the study. The results presented here
plying the ideal-gas law: are representative of the other flight conditions examined.
For the force coefficients, the deviation or error in the application
p = (γ − 1)[e − (ρ/2)(u 2 + v 2 + w 2 )] (17) of Sacks’s relations is defined in Eq. (19). The deviation 1 repre-
sents the difference between the right-hand and left-hand sides of
The turbulent viscosity µt , which appears in the viscous matrices, Eq. (4):
was computed using the Baldwin–Lomax turbulence model.13
The Coriolis and centrifugal acceleration terms due to the rotating  = C Nα̇ − Cm α − C Nq ≡ 1 (19)
coordinate system, which are contained in the source term Ĥ, are
shown in Eq. (18): Note that by simple algebraic manipulations, the following relations
that are algebraically equivalent to Eq. (19) are found:
f = 2Ω × u + Ω × (Ω × R) (18)

 = C Nq + C Nα̇ − Cm α − 2C Nq ≡ 2 (20)
The Coriolis acceleration is a function of the angular velocity of


the coordinate frame with respect to the inertial frame Ω and the  = 2C Nα̇ − Cm α − C Nq + C Nα̇ ≡ 3 (21)
fluid velocity vector u, which can be represented by the velocity
components u, v, and w. The centripedal acceleration is a function Essentially, these deviations are a measure of the accuracy of the
of the angular velocity of the rotating frame Ω and the displacement Sacks’s relations shown in Eqs. (4), (6), and (8). It could be argued
vector R between the axis of rotation and the local position in the that the differences 1 , 2 , and 3 are numerical errors that are not
flowfield. The acceleration vector f can be written in terms of its representative of physical phenomena. However, the deviation 
components along the x, φ, and r axes, f x , f φ and fr . can be computed independently in three different ways. The devia-
The steady thin-layer equations are solved using the parabo- tion 1 is computed using the individual damping coefficients C Nq
lized Navier–Stokes technique of Schiff and Steger.11 This “space- and C Nα̇ obtained independently from two different types of helical
marching” approach integrates the governing equations from the motion, while 2 and 3 are computed from the pitch-damping
nose of the flight body to the tail. Following the approach of Schiff coefficient sum C Nq + C Nα̇ and one of the two individual damping
and Steger, the governing equations, which have been modified here coefficients C Nq and C Nα̇ , respectively. The importance of having
to include the Coriolis and centrifugal force terms, are solved using three independent methods of determining  is to demonstrate that
a conservative, approximately factored, implicit finite difference nu- the deviation is not due to an error in the prediction of any one of
merical algorithm as formulated by Beam and Warming.14 Details the three damping coefficients.
of the implementation of the source term that contains the Coriolis Similar expressions for the deviation of the moment coefficients
and centrifugal force terms are given in Refs. 8 and 9. ¯ can also be obtained and are shown in Eqs. (22–24):
The technique has been validated with available experimental
data where possible, and excellent agreement is found.8,15 Grid- ¯ = Cm α̇ − C2m α − Cm q ≡ 
 ¯1 (22)
resolution studies were also performed in the original studies to

ensure grid-independent solutions.8,9  ¯2
¯ = Cm q + Cm α̇ − C2m α − 2Cm q ≡  (23)

Results

¯ = 2Cm α̇ − C2m α − Cm q + Cm α̇ ≡ 
 ¯3 (24)
In the current context, this computational procedure allows the
general applicability of Sacks’s pitch-damping relations to be ex- Figure 5 shows the computed deviation for the force coefficient
amined. Arguably, up to now, it has not been possible to assess the obtained from Eqs. (19–21) as a function of longitudinal distance
validity or accuracy of these relationships because of the uncertainty from the nose. Here, the body has been lengthened to 20 calibers
associated with experimentally derived pitch-damping data and the to more clearly illustrate the variation of the deviation along the
lack of a higher-order theory. In the current research effort, the tech- body length. Results are obtained from a single computation of a
nique has been applied to two axisymmetric body geometries: an 20-caliber-long body. Values of the force coefficients at intermediate
ogive-cylinder configuration and a cone-cylinder body with a flared body axial locations are obtained by integrating the force distribution
 WEINACHT AND DANBERG 869

Fig. 5 Comparison of the three computed deviations ∆1 , ∆2 , and ∆3

as a function of longitudinal distance from the nose. Fig. 7 Comparison of the longitudinal distribution of the three com-
Downloaded by UNIVERSITY OF CALIFORNIA - DAVIS on February 13, 2015 | http://arc.aiaa.org | DOI: 10.2514/1.10718

puted deviations ∆ ¯ 1, ∆
¯ 2 , and ∆
¯ 3 relative to the damping-moment co-
efficients; xcg /L = 0.5598, L/D = 9.

Fig. 6 Comparison of the longitudinal distribution of the three com-

puted deviations ∆1 , ∆2 , and ∆3 relative to the damping-force coeffi-
cients; xcg /L = 0.5598, L/D = 9. Fig. 8 Comparison of the damping-moment coefficients distributions
predicted from Sacks’s relations with CFD predictions; xcg /L = 0.5598,
L/D = 9.
from the nose to the corresponding axial location. The results shown
in Fig. 5 can be interpreted either as the distribution of the force Figure 7 shows the distribution of the computed deviation for the
deviation as a function of axial position for a fixed length body or as moment relative to the three damping-moment coefficients and the
the force deviation as a function of body length. Both interpretations pitching second-moment coefficient along the length of the body.
are equivalent under the assumptions associated with the space- Again, there is very good correlation between the three methods
marching approach used here. for computing the deviation. The computed deviation is small com-
There is very good correlation between the three different meth- pared to individual damping-moment coefficients and to the pitch-
ods of computing , suggesting that the deviation is representative damping moment sum. Sacks’s relation for the moment [Eq. (5)]
of a physical effect rather than simply numerical error. The deviation shows that the difference between Cm α̇ and Cm q is the pitching
is very small over the nose and begins to grow in a linear fashion second-moment coefficient C2m α . This also can be seen in Fig. 7. In
a couple of body diameters aft of the nose. The deviation (for the a sense, Figs. 6 and 7 provide some measure of the expected error in
force) can be shown to be independent of c.g. position using the applying Sacks’s relations to obtain the various damping-moment
c.g. translation relations for the individual force coefficients. This coefficients given that normal-force distribution (or Cm α and C2m α )
was also confirmed by varying the c.g. position in the computations and one of the pitch-damping coefficients is known.
as well. The computed deviation was also found to be somewhat The most likely application of Sacks’s relations in practical sit-
dependent on nose length and Mach number, although the results uations is to compute the individual damping coefficients from the
shown are representative of the trends observed for the other flight pitch-damping coefficient sum because the pitch-damping moment
conditions. sum is much easier to measure. Through simple algebraic manipu-
Figure 6 shows the distribution of the computed deviations for lations of Eqs. (7) and (9), the following form of the Sacks relations
the force 1 , 2 , and 3 compared with the distribution of the indi- can be obtained:
vidual pitch-damping coefficients and the pitch-damping sum along 
 
the length of the body. Relative to C Nα̇ and the pitch-damping force Cm α̇ = Cm q + Cm α̇ + C2m α 2 (25)
sum, the deviation is quite small. The distribution of the deviation is 
 
also small compared with the pitch-damping force coefficient C Nq Cm q = Cm q + Cm α̇ − C2m α 2 (26)
except near the end of the body where C Nq itself is nearly zero.
Equation (4) shows that, according to Sacks’s relation, the differ- Figure 8 shows the longitudinal distribution of the pitch-damping-
ence between C Nα̇ and C Nq is the pitching-moment coefficient Cm α . moment coefficients Cm q and Cm α̇ obtained by applying Sacks’s
This is also graphically shown in Fig. 6. Near the nose of the body, relationships using the pitch-damping-moment coefficient sum
the difference between the two damping coefficients C Nα̇ and C Nq [Cm q + Cm α̇ ] and the second-moment coefficient of the normal force
increases at nearly the same rate as the pitching moment coefficient. C2m α . Here, CFD has been used to compute both [Cm q + Cm α̇ ]
On the aft portion of the body, the pitching-moment coefficient is and C2m α . The results from the application of these relations are
nearly constant, and the difference between C Nα̇ and C Nq becomes compared with the CFD predictions of the individual damping
relatively constant. coefficients. The coefficients Cm q and Cm α̇ are overpredicted and
 870 WEINACHT AND DANBERG

Fig. 11 Comparison of the damping-moment coefficients derived from

experimental data or CFD using Sacks’s relations with direct CFD pre-
Fig. 9 Comparison of the damping-force coefficient distributions pre-
Downloaded by UNIVERSITY OF CALIFORNIA - DAVIS on February 13, 2015 | http://arc.aiaa.org | DOI: 10.2514/1.10718

diction; L/D = 9.
dicted from Sacks’s relations with CFD predictions; xcg /L = 0.5598,
L/D = 9.

Fig. 12 Schematic of the flared projectile geometry; all dimensions in

calibers (1 caliber = 8.28 mm).

available. However, Sacks’s relations do require the normal-force

distribution to determine C2m α because generally only the normal
force and pitching moment are available as global coefficients from
many sources. In lieu of more sophisticated computational methods
for determining the normal-force distribution, fast-design aeropre-
diction codes3 should provide acceptable accuracy for a variety of
Fig. 10 Comparison of the predicted damping-moment coefficients flight vehicle geometries.
with CFD and slender-body theory results; L/D = 9.
To demonstrate the use of Sacks’s relations when only global
forces and moments are available, pitch-damping-moment coeffi-
underpredicted by about 5 and 12%, respectively, although the ab- cient sum data from range firings of the ANSR have been used
solute error is similar for both coefficients. The distribution of the to estimate the individual pitch-damping-moment coefficients us-
damping coefficients over the body is also very well predicted. ing Sacks’s relations. The normal-force distribution from the fast-
Similar analysis can be performed for the force coefficients. design aeroprediction code AP023 has been used to predict C2m α .
Figure 9 shows the comparison of the longitudinal distributions Figure 11 shows the individual pitch-damping coefficients predicted
of the individual pitch-damping-moment coefficients C Nq and C Nα̇ from Sacks’s relations [Eqs. (25) and (26)] using experimental mea-
predicted using the form of Sacks’s relations in Eqs. (27) and (28) surements of the pitch-damping-coefficient sum and C2m α obtained
with direct CFD predictions. The distribution of the force coeffi- from AP02. The results are compared with direct CFD predictions
cients along the body is very well predicted using Sacks’s relations: of the individual damping coefficients as well as the results obtained

  from Sacks’s relation using CFD predictions of the pitch-damping-
C Nα̇ = C Nq + C Nα̇ + Cm α 2 (27) coefficient sum and C2m α shown in Fig. 10. The experimentally de-
rived values of the individual pitch-damping coefficients compare

  well with the predicted results. The results also indicate that the
C Nq = C Nq + C Nα̇ − Cm α 2 (28)
biggest source of error is produced by the uncertainty in the experi-
Figure 10 shows the results for the individual pitch-damping co- mentally derived pitch-damping-moment coefficient rather than the
efficients for various CG positions for the body with the length-to- error in Sacks’s relations. (The pitch-damping data used here are
diameter ratio (L/D) of 9 at Mach 2.5 obtained by applying Eqs. (25) from a highly regarded data set and are representative of the ex-
and (26). The results are compared with CFD results and with direct pected accuracy for the pitch-damping coefficient sum from range
slender-body theory results. There is very good correlation of the firings.)
results obtained with Sacks’s relations and the CFD results, and the
results are a significant improvement over the slender-body theory. Flared Projectile Results
For the results shown in Figs. 8 and 9, comparisons have been The performance of Sacks’s relations was also examined for a
made between the force and moment distributions along the body. flared projectile geometry shown in Fig. 12. Direct CFD predictions
This was done because these distributions are easily extracted from of the pitch-damping-coefficient sum have been validated previously
the computational results, and they allow detailed examination of for this configuration using the same method described previously.15
the performance of Sacks’s relations. It must be emphasized that it is For the current paper, additional CFD predictions of the individual
not necessary to obtain the pitch-damping force and moment distri- pitch-damping coefficients were also performed for flight velocities
butions in order to apply Sacks’s relations because they work for the between Mach 2 and Mach 5 for sea-level atmospheric conditions
global force and moment coefficients as well. This is particularly im- (Re D = 4.5 × 105 and 1.125 × 106 , respectively).
portant to consider when the source of the pitch-damping-coefficient Figure 13 shows the longitudinal distribution of the computed
data is an experiment in which only global coefficients are typically deviations of the moment determined from Eqs. (22–24) compared
 WEINACHT AND DANBERG 871

Fig. 13 Comparison of the longitudinal distribution of the three com- Fig. 16 Comparison of the damping-moment-coefficient distribution
Downloaded by UNIVERSITY OF CALIFORNIA - DAVIS on February 13, 2015 | http://arc.aiaa.org | DOI: 10.2514/1.10718

puted deviations ∆ ¯ 2 , and ∆

¯ 1, ∆ ¯ 3 relative to the damping-moment co- predicted from Sacks’s relations with CFD predictions, 15-deg flared
efficients, flared projectile, Mach 2. projectile, Mach 5.

second-moment coefficient for the flared projectile geometry. Com-

parisons were made with direct CFD predictions. Figures 14 and 15
show comparisons of the longitudinal distribution of the individual
pitch-damping-moment coefficients at Mach 2 and 5, respectively,
for the flared projectile geometry. Very good agreement between
the results obtained by applying Sacks’s relations and direct CFD
predictions is found. Additional results for the same flared projectile
with the 6-deg flare replaced by a 15-deg flare were also obtained
and are shown in Fig. 16. Despite a nearly doubling of the pitch-
damping coefficient due to the larger flare, the Sacks’s relations
results are in very good agreement with the direct CFD predictions.

Conclusions
In conclusion, the results presented here indicate that Sacks’s
pitch-damping relations are only strictly valid under the context of
Fig. 14 Comparison of the damping-moment-coefficient distribution the theory from which they were originally developed. They do,
predicted from Sacks’s relations with CFD predictions, flared projectile, however, provide a reasonably good means of estimating the pitch-
Mach 2. damping coefficients when one of the three pitch-damping coeffi-
cients can be determined. The most likely practical use of these rela-
tions might be to provide estimates of the individual pitch-damping
coefficients using values of the pitch-damping-coefficient sum de-
termined from some other source, such as experimental data and
engineering estimation approaches or when the additional expense
of the separate computational fluid dynamics computations of the in-
dividual pitch-damping coefficients is not justified. In some cases,
it appears that the error in applying these relationships is smaller
than the error associated with generating the initial pitch-damping
coefficient (such as with engineering estimation approaches) from
which the other two damping coefficients are derived using Sacks’s
relations.
Applying Sacks’s relations to determine the individual pitch-
damping coefficients from the pitch-damping sum represents only
one possible application of Sacks’s relations. These relations could
also benefit theoretical developments because theories for predicting
the pitch-damping coefficients need only focus on a single damping
coefficient. The other damping coefficients could then be obtained
Fig. 15 Comparison of the damping-moment-coefficient distribution from Sacks’s relations. Such an approach has been already used as an
predicted from Sacks’s relations with CFD predictions, flared projectile, estimation procedure for the damping coefficients.6 In this work, the
Mach 5. distribution of C Nα̇ along the body is predicted using slender-body
theory with empirically based corrections. The damping coefficient
with longitudinal distribution of the individual pitch-damping coef- C Nq can then be obtained from Sacks’s relations. Once the damping
ficients predicted using CFD at Mach 2. Again, the deviation over force distributions are known, the damping moments can be easily
the body is small compared to the pitch-damping coefficients, even obtained by integration of the force loadings. Improvements in the
on the afterbody where both the geometry and moment coefficients estimates of C Nq from Sacks’s relations can also be obtained by
are changing significantly. Similar results were found at the other correlating the error  as well, an approach used in Ref. 6 to further
flight velocities. improve the estimates of C Nq .
Similar to the results for the secant ogive/cylinder configura- Finally, in the current paper only axisymmetric configurations
tion, the form of Sacks’s relations embodied in Eqs. (25) and (26) in supersonic flight have been considered. The theory from which
was applied to determine the individual pitch-damping coefficients Sacks derived the relations considered here is applicable to both
from the pitch-damping-moment-coefficient sum and the pitching- wingless and winged vehicles. Further research is still required to
 872 WEINACHT AND DANBERG

assess the performance of Sacks’s relations for winged vehicles and 8 Weinacht, P., Sturek, W. B., and Schiff, L. B., “Navier–Stokes Predictions

for other flight velocity regimes. of Pitch-Damping for Axisymmetric Projectiles,” Journal of Spacecraft and
Rockets, Vol. 34, No. 6, 1997, pp. 753–761.
9 Weinacht, P., “Navier–Stokes Predictions of the Individual Components
References of the Pitch-Damping Sum,” Journal of Spacecraft and Rockets, Vol. 35,
1 Sacks, A. H., “Aerodynamic Forces, Moments, and Stability Derivatives No. 5, 1998, pp. 598–605.
10 Park, S. H., Kim, Y., and Kwon, J. H., “Prediction of Damping Co-
for Slender Bodies of General Cross Section,” NACA TN 3283, Nov. 1954,
p. 27. efficients Using the Unsteady Euler Equations,” Journal of Spacecraft and
2 Bryson, A. E., Jr., “Stability Derivatives for a Slender Missile with Ap- Rockets, Vol. 40, No. 3, 2003, pp. 356–362.
11 Schiff, L. B., and Steger, J. L., “Numerical Simulation of Steady Super-
plication to a Wing-Body-Vertical-Tail Configuration,” Journal of the Aero-
nautical Sciences, Vol. 20, No. 5, 1953, pp. 297–308. sonic Viscous Flow,” AIAA Journal, Vol. 18, No. 12, 1980, pp. 1421–1430.
3 Moore, F. G., and Hymer, T. C., “The 2002 Version of the Aeropredic- 12 Weinacht, P., and Sturek, W. B., “Computation of the Roll Character-

tion Code: Part I—Summary of New Theoretical Methodology,” U.S. Naval istics of a Finned Projectile,” Journal of Spacecraft and Rockets, Vol. 33,
Surface Warfare Center, Rept. NSWCDD/TR-01/108, Dahlgren, VA, March No. 6, 1996, pp. 769–775.
13 Baldwin, B. S., and Lomax, H., “Thin Layer Approximation and Al-
2002.
4 Vukelich, S. R., and Jenkins, J. E., “Missile DATCOM: Aerodynamic gebraic Model for Separated Turbulent Flows,” AIAA Paper 78-257, Jan.
Prediction on Conventional Missiles Using Component Build-up Tech- 1978.
14 Beam, R., and Warming, R. F., “An Implicit Factored Scheme for the
niques,” AIAA Paper 84-0388, Jan. 1984.
5 Whyte, R. E., “Spinner—A Computer Program for Predicting the Aero- Compressible Navier–Stokes Equations,” AIAA Journal, Vol. 16, No. 4,
Downloaded by UNIVERSITY OF CALIFORNIA - DAVIS on February 13, 2015 | http://arc.aiaa.org | DOI: 10.2514/1.10718

dynamic Coefficients of Spin Stabilized Projectiles,” General Electric Co., 1978, pp. 85–129.
15 Weinacht, P., “Navier–Stokes Predictions of Pitch-Damping for a Fam-
Class 2 Rept. 69APB3, Burlington, VA, Aug. 1969.
6 Danberg, J. E., and Weinacht, P., “Approximate Computation of Pitch- ily of Flared Projectiles,” AIAA Paper 91-3339, Sept. 1991.
Damping Coefficients,” AIAA Paper 2002-5048, Aug. 2002.
7 Sigal, A., “Correlation of the Damping in Pitch Stability Derivatives for R. Cummings
Body-Tail Configurations,” AIAA Paper 94-3482, Aug. 1994. Associate Editor