Sie sind auf Seite 1von 20

# 1

Modeling and numerical simulation of the forces acting on a sphere during early-water
entry

J ohn Abraham
a
J ohn Gorman
b
Franco Reseghetti
c
Ephraim Sparrow
b
J ohn Stark
b

a
University of St. Thomas
School of Engineering
2115 summit Ave
St. Paul, MN 55105-1079, USA
jpabraham@stthomas.edu
phone: 651-962-5766
fax: 651-962-6419

b
University of Minnesota
Department of Mechanical Engineering
111 Church St. SE
Minneapolis, MN 55455-0111

c
ENEA, UTMAR-OSS,
Forte S. Teresa,
19032 Pozzuolo di Lerici, Italy

Abstract:

Mathematical modeling, absent simplifying assumptions and coupled with numerical simulation
has been implemented to determine the motions and forces experienced by a sphere penetrating a
water surface from an air space above the surface. The model and simulation are validated by
comparisons with extensive experimental data and with trends from approximate analyses.
Although the present work adds to the understanding and quantification of the sphere as an entry
object, its major contribution is model development and validation to enable investigation of
water entry of objects of practical utility such as the expendable bathythermograph (XBT). The
XBT device is widely used in the determination of temperature distributions in large water
bodies such as oceans. The measured temperature distributions are, in turn, used to determine
the thermal energy content of oceans. During the course of the numerical simulations,
parametric variations were made of the sphere velocity, surface tension, flow regime (laminar or
turbulent), and Reynolds number. The drag-coefficient results were found to be independent of
these quantities. This outcome indicates that momentum transfer from the sphere to the adjacent
liquid is responsible for the drag force and that friction is
a secondary issue.

1. Introduction

Studies of the impact of solid objects onto horizontal liquid surfaces have led to a detailed
understanding of the forces that act on the object in the very early stages of impact and on the
shape and size of an air cavity formed adjacent to the object. This information is helpful for
predictions of the motion of objects that breach the surface as they pass into the liquid. There are
many applications for which such predictions are important, such as launching of torpedoes,
motion of watercraft, and the deployment of oceanographic measuring devices. It is this last
application which has motivated the present study.
Each year, many thousands of oceanographic measurement devices are employed throughout
the world. Those devices have a variety of shapes and sizes, and their launch conditions differ
significantly. For example, one of the most commonly used devices to measure ocean
temperatures for subsequent calculation of heat content is the expendable bathythermograph
(XBT). It can be launched from heights that vary between 2-30 meters above the surface of the
ocean. This large variation of launch heights has led to significant differences in the impact
velocity and the angle of entry.
As the XBT device travels through increasing ocean depths, it collects temperature
information which is conveyed to an on-ship computer by means of a trailing copper wire. The
wire unspools as the device falls, and canted fins engender a rotating motion that aids in the
unspooling process.
For climate studies, it is important to accurately measure the ocean temperature distribution
throughout a specified depth. XBT timewise-temperature records, combined with the
instantaneous depth of the device, enable the determination of the thermal energy content of the
ocean which, in turn, facilitates closure of the Earths energy balance.
One shortcoming in the current use of the XBT device is that the depth is not measured
directly, but is inferred from a fall rate equation (derived under idealized experimental
conditions). Its coefficients were originally proposed by the manufacturers in the 1960's and
have been recalculated in experiments where the XBT depth-temperature profiles were compared
with the results of more sophisticated and accurate devices. Those controlled experimental
conditions are not always encountered in the field, so that the global accuracy of XBT
measurements in the historical archive is a hard problem to solve in the ocean heat content
calculations.
The present authors have created a methodology for predicting the rate at which XBT devices
fall through the water column (Stark et al., 2011; Abraham et al., 2011; 2012a; 2012b). That
model is able to deal with motion of fully submerged devices; however, it does not account for
the impact of water entry on the device motion.
As recognized in Abraham et al., (2012a), surface impact could affect the fall rate of the
device through water and its inferred depth. This realization has motivated the present study. A
recent publication which focuses on the XBT device reinforces the importance of the entry forces
(Xiao and Zhang, 2012). Those investigators included an entry calculation for their device and
then extended the solution to fully submerged motion. They concluded that the probe velocity at
entry had little impact on the subsequent rate of fall. On the other hand, that paper did not deal
with the motion of the object prior to impact with the water. Also, the verification of those
results with literature-based information on spheres only extended partway through the entry
process (normalized depth of 0.4 times the sphere radius), thereby avoiding issues related to the
trailing separation region. Furthermore, the calculations presented in that paper utilized a

regularly shaped grid of elements and did not, therefore, use special boundary layer elements that
are typically employed to resolve flow in the boundary layer. Additionally, the type of XBT was
unspecified in the paper, it does not appear to be one of the standard XBT devices employed in
oceanographic temperature measurements. Consequently, while that paper was a significant
step forward, further advancements in this area are required.
The goal of the present work is to numerically explore the entry forces on a sphere that passes
from air to water. The simulation will carried through to the situation in which the device is fully
submerged in the water. Drag forces will be extracted at all instances and compared with
experiments from the literature. The sphere is chosen as the shape of interest here because it has
been more extensively studied than any other shape. Consequently, it can serve as an accepted
baseline for the validation of the present physical model and simulation method. The successful
validation of the simulation model will justify its use for actual oceanographic devices.
The literature on the entry problem is rich and extends back more than a century. The first
significant study used novel photographic methods to illuminate the dynamics of the fluid flow
in the cavity following sphere entry (Worthington and Cole, 1897). This work provided the seed
for a treatise (Worthington, 1908). A related issue was addressed by Von Karman (1929) in
connection with seaplane floats. Wagner (1932) investigated phenomena related to the impact of
objects on liquid surfaces. Later, pioneering studies were able to extract drag forces on objects
passing into water (Watanabe, 1934; Gilbarg and Anderson, 1947; May and Woodhull, 1948;
May, 1951; 1952) and pressure distributions on the surfaces of objects during impact
(Richardson, 1948).
Simultaneous with these early experiments, analytical methods and models were developed
that allowed predictions of entry forces, particularly in the very early stages of entry (Courant et
al., 1945; Trilling, 1950; Shiffmann and Spencer 1945a; 1945b; 1947). During the following
decades, a number of analytical studies extended the available information to other shapes and to
oblique entry situations (McGehee et al., 1959; Nisewanter, 1961; Waugh and Stubstad, 1966;
Verhagen, 1967).
A revitalization of this research occurred in the 1970s with a focus on the dynamics of the air
cavity formed upon the entry of the object into the fluid. Pressure measurements within the
cavity have been performed (Abelson, 1970). Also, new modeling strategies were devised to
deal with non-spherical shapes (Hughes, 1972). Much of the research was sponsored by the
United States Government in order to understand the motion of naval weapons which undergo
liquid-surface entry (Baldwin, 1975a; 1975b; Baldwin and Steves, 1975; May, 1975; Koehler
and Kettleborough, 1977; Moghisi and Squire, 1981). Those studies used analytical and
experimental methods and investigated a variety of shapes including spheres, cylinders, wedges,
and Ogives.
A fourth generation of studies has emerged more recently that extend the available
information to high-speed entry problems (Shi and Takami, 2001; Gekle et al., 2009; Guo et al.,
2012). New analytical or numerical methods have been developed (Korobokin and Pukhnachov,
1988; Park et al., 2003; Bergmann et al., 2009; Do-Quang and Amberg, 2010). Further
contributions to cavity dynamics are reported (Duclaux et al., 2007; Grumstrip et al., 2007).
Finally, experiments continue to be performed to further investigate the issues that affect entry
forces (Truscott, et al., 2012)
Despite this extensive history, it appears that the details of the distribution of drag on the
surface of a sphere passing into water have not been thoroughly investigated nor have flow
patterns in the neighborhood of the sphere been presented. The goal of this study is to develop a

model and a concomitant numerical approach that can be used to definitively provide such
information. The devised methodology will utilize commercially available software and will
explore the impact of both laminar and turbulent flow conditions. The results will be compared
with the best available experimental data to validate the approach. If this methodology is
supported by comparisons with experimental data, it can be used in conjunction with already
existing information for fully submerged objects in order to improve the identification of the
instantaneous depths of sensory devices.

2. Mathematical Model

The mathematical model was formulated to take account of three-dimensional unsteady fluid
motions in air and water. The volume-of-fluid method (VOF) is used to separate the two fluid
regions. Zones completely filled by air are represented by VOF
air
=1 (VOF
water
=0), while water
regions are defined by VOF
water
=1 (VOF
air
=0). The VOF values of air and water add to 1
throughout the entire solution domain. The air-water interface is identified by VOF =0.5.
Coupling between the descending sphere and the respective fluids is due to fluid-solid
friction. The sphere is initially positioned in the air above the water-air interface. To initiate the
action, the sphere is given a vertical downward velocity. Gravity acts on both the sphere and on
the respective fluids; buoyancy effects are also accounted for.
In each fluid, the multi-dimensional, unsteady equations of fluid motion are solved. Those
equations include conservation of mass and momentum. Separate numerical solutions were
performed for the respective regimes of laminar and turbulent flow. Subsequently, comparisons
will be made between the results for the respective flow regime models to elucidate whether
turbulent fluctuations in the fluid motion have a significant impact on the results.
A specific motivation for considering both laminar and turbulent flows is that in
practice, it is not known a priori what is the state of the fluid flow. In particular, under
controlled laboratory conditions, it is possible that the air and water are both slowly moving and
are laminar. On the other hand, for in-field deployments of devices, wind-driven airflow and
water currents may lead to turbulence. A more fundamental issue is that for flows over blunt
objects, the leading-edge boundary layer may be laminar, whereas turbulence may prevail
downstream.
For laminar flow, the governing equations for fluid flow are expressed by Eq. (1) and by
Eq. (2) without the quantity
turb
These equations respectively represent conservation of mass
and momentum. When turbulent flow is being analyzed,
turb
, the turbulent viscosity, must be
included.

0 =
c
c
i
i
x
u
(1)

( ) 3 , 2 , 1 = +
|
|
.
|

\
|
c
c
+
c
c
+
c
c
=
|
|
.
|

\
|
c
c
+
c
c
j f
x
u
x x
p
x
u
u
t
u
j
i
j
turb
i j i
j
i
j
(2)

The symbol u
i
denotes the fluid velocity in one of the three coordinate directions, is the
density, p is the static pressure, and the molecular viscosity. The symbol f
j
is a body-force

term and allows inclusion of buoyancy forces. Note that the quantities and that appear in Eq.
(2) must be specific to air and water in the respective fluid regions.
For those situations where turbulent motions occur, additional equations are necessary to
quantify the turbulent viscosity
turb
. To this end, a number of phenomenological models have
been proposed, starting in 1972. The model chosen here, the Shear Stress Transport (SST) model
(Menter et al., 1994), has been shown to provide acceptable results for complex fluid flow
situations. Two supplementary quantities, k and e, respectively the turbulence kinetic energy
and the specific rate of turbulence dissipation, were introduced by Menter from which
turb

follows as

( )
2
, max SF a
a
turb
e
k
= (3)

The quantities k and e are found from two additional transport equations which are presented as
Eqs. (4) and (5).

( ) ( )
(

c
c
|
|
.
|

\
|
+
c
c
+ =
c
c
+
c
c
i
turb
i i
i
x x
P
x
u
t
k
o

ke |
k k
k
k 1
(4)
and
( ) ( )
( )
i i i
urb t
i i
i
x x
F
x x
S A
x
u
t
w
c
c
c
c
+
(

c
c
|
|
.
|

\
|
+
c
c
+ =
c
c
+
c
c e k
e o

e
o

e |
e
e e 2
1
2
2
2
1
1 2
(5)

The term P
k
represents the production of turbulence kinetic energy, while the o terms are
Prandtl-number-like coefficients for the respective transported variables. The quantity S is the
magnitude of the shear strain rate, and the F terms are blending functions which transition
between the ke model near the wall and the kc model away from the wall. Equations (1)-(5)
complete the definition of the SST methodology.
The solution domain encompassed the two fluids (air and water) and the sphere which
begins its descent from a position slightly above the water surface. A schematic diagram of the
solution domain, Figure 1, also shows the initial position of the sphere. The dimensions were
selected to assure that the solutions were independent of the size of the domain. The mass of the
sphere was 0.0335 kg, with a radius of 0.02 m. Values of the sphere initial velocity were varied
parametrically from 2 to 20 m/s, while the starting point of the center of the sphere was held
fixed at 0.03 m above the water surface. Properties for air and water are specified in Table 1.

Fig. 1 The solution domain with annotations showing the physical size and boundary surfaces

Table 1 Properties of air and water
Air properties Water properties
Density (kg/m
3
) 1.19 997
Kinematic Viscosity (m
2
/s) 1.54e-5 8.93e-7
Dynamic Viscosity (N-s/m
2
) 1.83e-5 8.9e-4

Calculations were performed both with and without consideration of surface tension. For the
cases where surface tension was included, the value of the surface tension was 0.07 N/m.
The motion of the sphere was determined by a force-momentum balance as embodied in
Newtons Second Law. The forces exerted on the sphere are due to fluid-solid interactions
(drag), gravity, and buoyancy. If the vertical component of Newtons second law is considered,
there results

m
dv
dt
= F
dug
+F
buounc
+F
wcght
(6)

Both the weight and the buoyancy force are strictly vertical, and the vertical component of the
drag force is readily evaluated. In Eq. (6), V denotes the time-dependent vertical velocity, and m
is the mass of the sphere.
To complete the description of the problem, it is necessary to specify boundary conditions at
each of the surfaces which bound the solution domain. With reference to Fig. 1, the lower
boundary was designated as an opening which permits water to pass into and out of the domain.
There, entrainment (dragging of fluid by shear forces) was allowed. A similar condition was
imposed at the top of the domain. The nature of the opening boundary condition requires that the
static pressure be specified at the boundary in question. In that regard, the pressure of the
ambient air was set equal to zero, and hydrostatic pressure was imposed at the upper and lower
boundaries when appropriate. At the vertical boundary of the solution domain, the static
pressure was specified, either zero (ambient air) or hydrostatic.
Advantage was taken of symmetry so that the solutions were only carried out only within a
wedge that subtended 10 degrees. The two symmetry planes that bound the wedge prevent any
fluid flow across them and also require that the normal derivatives of the dependent variables be
zero. The imposition of symmetry is supported by visual observations reported in numerous
experimental studies.

3. Solution Implementation

The numerical solutions were performed by means of ANSYS CFX software, which is a
finite-volume code. That code provides access to several turbulence models. The one utilized
here is the Shear-Stress Transport model (SST) which has proven to be highly successful in
predicting complex fluid flows.
Two independent sets of numerical solutions were performed with the boundary conditions
imposed at different distances from the sphere. Comparisons between the two sets of solutions
demonstrated total insensitivity to the different locations at which the boundary conditions were
imposed. The final configuration of the solution domain used for the numerical solutions is that
displayed in Fig. 1. The initial condition for the solutions was no motion in either fluid.
Numerical solutions were carried out until the sphere was fully immersed in the water. The
temporal duration of the solution was dictated by the magnitude of the initial velocity. The
timewise step size was varied until the solution was independent of it .
The mesh which was used to discretize the solution domain was also a major focus issue,
particularly in the near vicinity of the sphere. Figure 2 has been prepared to quantify the fineness
of the mesh. The figure displays a sequence of magnifications to illustrate the mesh deployment
adjacent to the sphere surface. One zone which requires careful meshing is the sphere-fluid
interface. In addition to careful mesh refinement, special elements were employed at the sphere
surface itself. These special elements, called boundary elements or inflation elements, are
aligned with the solid surface to resolve the large gradients there. In order to accurately capture
the boundary layer flow, it was necessary to tailor the mesh to satisfy the refinement criterion y+
~1, where y+is a dimensionless distance measured from the fluid-wall interface. A more
detailed presentation of the impact of mesh and time step will be presented later.

Fig. 2 Magnifications showing refined elements in the vicinity of the sphere-fluid interface.

4. Results and Discussion

The key result which is sought from these simulations is the drag coefficient in the early
stages of entry. Specifically, focus is directed to the period between the instant of initial contact
of the sphere with the water surface to the moment when the sphere became fully submerged.
The depth of penetration is commonly referenced to the size of the entry object (sphere radius),
so that the dimensionless penetration depth is defined as

b =
pcnctuton dcpth
sphcc udus
(7)

The fully submerged condition corresponds to b >2.

4.1. Drag Results
Values of the drag coefficient are obtained at any instant of time by nondimensionalization of
the instantaneous vertical force F exerted on the sphere by the fluid. Mathematically, the drag
coefficient is

C
d
=
F
1
2
pAI
2
(8)

where V is the instantaneous velocity of the sphere. The density used in the denominator of
Eq. (8) is that of water, and the symbol A represents the frontal area of the sphere. The thus-
determined drag coefficient is that required for use in dynamic models such as those that enable
prediction of the depth of oceanographic devices (Stark et al., 2011; Abraham et al., 2011;
2012a; 2012b).
A physical phenomenon that leads to very high drag coefficients at the beginning of the entry
is the acceleration of the water contiguous to the sphere. Figure 3 is a schematic diagram
prepared to show the water region which experiences that acceleration. This acceleration, which
requires a transfer of momentum from the sphere to the fluid, results in a brief elevation of the
drag coefficient. The thus-accelerated water is typically referred to as a virtual mass.

Fig. 3 Illustration showing the liquid adjacent to the sphere which is accelerated.

With this discussion as background, entry-drag results are presented as shown in Fig. 4. This
set of results is an outcome of a study of spatial mesh- and time-step independence. Three sets
of results are shown, labeled respectively as, finer mesh, small time step; fine mesh, small time
step; and coarse mesh, large time step. The small and large time steps are 2e-5 and 2e-4
seconds, respectively. The coarse-mesh case encompassed 49,000 elements with the elements
nearest the sphere being1.5 mm thick. In contrast, the fine mesh case consisted of 664,000
elements, with the elements nearest the sphere of 0.002 mm thickness. Despite these large
differences in the spatial mesh and the time step, the results are seen to be nearly the same over
the range of dimensionless depths set forth in the figure. A further refinement was carried out
with a mesh of 7,150,000 elements and a nearest wall element thickness of 0.001 mm. This
third-level of refinement was performed to verify the mesh-independence of the calculations.
Calculations were performed using a quad-core 3.1 GHz processer with 8GB RAM. The
coarse-mesh, large-time-step solutions required 12 clock hours for the calculations, whereas the
fine-mesh, small time-step simulations required 210 hours for the laminar solution and 240 hours
for the turbulent calculation. The third level of refinement required 2500 hours (~105 days) for a

10

full solution. Furthermore, the volume of data required to store all simulation variables at each
time step was approximately 4TB. As a consequence, the finer mesh, small step solutions were
only carried out for a portion of the entry cycle (up to b ~0.5). It can be seen that at least for this
portion of the entry problem, the results are quite similar, despite the more than 100-fold
difference in mesh and solution time.

Fig. 4 Dependence of the predicted drag coefficient on mesh and time step. The small and large
time steps are 2e-5 and 2e-4 seconds, respectively. The coarse, fine, and finer meshes consisted
of 49,000, 664,000, and 7,150,000 elements, respectively.

The next set of results to be shown, which are presented in Fig. 5, compares the drag results
for the laminar and turbulent simulations. Also shown in the figure are experimental data
(Baldwin and Steves, 1975). These experiments were utilized because they cover the largest
range of dimensionless depths among those found in the literature. It should be noted that for the
simulation results displayed in this figure, the initial velocity of the sphere was 2 m/s (Reynolds
number ~80,000), and the specific gravity of the sphere was one. The experiments of Baldwin
and Steves utilized a 3-inch-diameter (0.0762 m) and a 5-inch-diameter (0.127 m) sphere with
specific gravity values of 2.42 and 0.795, respectively. The initial velocities used in the Baldwin
and Steves study ranged from 15-23 ft/sec (4.6-7.0 m/s).
The aforementioned differences in the operating conditions between the present simulations
and the Baldwin-Steves experiments is readily rationalized by the use of dimensionless
parameters, as demonstated by the excellent agreement between the simulation and experimental
results displayed in Fig. 5. The present results also agree very well with other experiments, such

11

as those of Moghisi and Squire, 1981. Also significant is the fact that the spread among
experimental results from numerous investigations is greater than the differences between the
present results and those of Baldwin-Steves. The drag coefficient results of Figs. 4 and 5 (and
the results that follow thereafter) agree to within 5% with those predicted from a theory based on
simplifying assumptions due to Shiffmann and Spencer, (1945a; 1945b; 1947).
Another novel finding seen in Fig. 5 is the seeming insignificance of turbulence. This
important outcome can be attributed to the dominance of momentum transfer between the sphere
and the contiguous water as the source of drag, with friction being a secondary issue. The
simulations, which extend far beyond the experimental results, continue to show excellent
agreement between the laminar and turbulent solutions.
When taken together, the excellent accord between the present simulations, literature-based
experiments, and theoretical predictions serve to validate the model and the simulation
methodology developed here.

Fig. 5 Drag coefficients in the entry zone, comparison of experimental results with laminar and
turbulent predictions.

In order to explore the possible dependence of the drag coefficient on velocity, simulations
were performed for initial velocities that varied from 2 20 m/s. This issue is addressed in Fig.
6, where results corresponding to initial velocities of 2 and 20 m/s are displayed. It is seen that
from the onset of impact to a dimensionless depth of one, the drag coefficients are nearly
identical, indicating that over the investigated range of velocities, the drag coefficient is nearly
independent of velocity. This outcome is attributable to the fact that the momentum transfer
between the sphere and the contiguous water is proportional to V
2
. This conclusion is reinforced
both by theory (Shiffmann and Spencer, 1945a; 1945b; 1947) and by experiments (Moghisi and
Squire, 1981).

12

## The Reynolds numbers corresponding to the range of investigated velocities is approximately

80,000 800,000. It should be recognized that the displayed universal nature of the drag
coefficient is not expected to continue for deeper penetrations, particularly for cases which lead
to a trailing cavity. On the other hand, the recognition that the drag coefficients are universal
over the time period when its values are highest is important for subsequent calculations of the
downward trajectory of the object. The results shown in Fig. 6 were implemented using the
coarse mesh and large time step model, as discussed in connection with Fig. 4.

Fig. 6 Comparison of drag coefficients for initial sphere velocities of 2 and 20 m/s.

Another relevant issue is whether surface tension has a significant effect on the drag
coefficient. In order to explore this issue, particularly during the duration of object entry when
the drag coefficient is largest, two calculations have been performed. The results, presented in
Fig. 7, correspond to laminar flow with and without surface tension included. It is not surprising
that surface tension has little impact on the results because the drag is primarily due to
acceleration of adjacent liquid (virtual mass).

13

Fig. 7 Impact of surface tension on drag coefficient.

Of primary utility for the prediction of the trajectory of an object as it descends beneath the
surface of the water is a convenient representation of the drag coefficient during entry. Based on
the foregoingdetermined independence of the drag coefficient from the entry velocity and
surface tension, algebraic representations have been developed and are displayed in Fig. 8 One
fitted polynomial pertains to dimensionless depths 0 <b <1, and another polynomial is
applicable for 1 <b <2. The complex behavior of the drag coefficient made it impossible to
develop a single algebraic representation which would be suitable over the entire range of
depths.

14

Fig. 8 Functional relationships for drag coefficients in the entry region.

4.2. Flow Patterns
In addition to the presented quantitative information for the drag coefficient, the numerical
simulations also provide visualizations of the fluid patterns in both the air and the water regions.
Such visualizations are presented in Figs. 9 (a) (e). Those figures respectively show the sphere
at different levels of penetration into the water region. The respective air and water regions are
identified by different shades of gray. The impact and initial penetration of the sphere into the
water are shown in the (a) and (b) parts of the figure. In particular, the latter shows the rise of
water along the periphery of the sphere and the subsequent formation of a cavity region at the
upper surface of the sphere. The initiation of separation and cavity formation is seen in the (c)
(e) parts of the figure.

15

(a)

(b)

(c)

Figs. 9 (a) (e) Images of the water-air
regions during the entry process

16

Another useful way to view the results, particularly with respect to impact forces, is to show
regions where water is moving in the upward or downward directions. The transference of
momentum to the adjacent water region, illustrated in Fig. 3, is responsible for a significant
portion of the impact forces. To provide the requisite information, Fig. 10 has been prepared.
That figure shows four instances in time during the entry. For each image, two contours are
shown. Beneath each of the spheres, a large region of downward-flowing fluid is indicated and
is connected with the legend at the right side of the image. To the side of the sphere, there is an
upward flowing region of water which is linked to the legend at the left of the image. Both
legends are expressed in meters/second. Velocity magnitudes less than 0.01 m/s are not shown.
It is seen that both regions of flow grow during the entry process. It is also seen that the
downward fluid occupies a significantly larger space than does the upward flow. The time
values listed in the figure are referenced from the onset of sphere-water collision.

Figs. 10 (a) (d) Regions of upward and downward flow of water

5. Concluding Remarks

This investigation was motivated by the need to formulate and definitively validate a
physical model and its numerical implementation that can be used to accurately describe the entry
of a solid object into water from an air space above the air-water interface. Although the work
significantly advances the modeling of a sphere as the entry object, its greater significance is to
provide a validated model for the entry of other objects of practical utility such as oceanographic
temperature measuring devices. The reason for the present focus on the sphere is that, among all
other object geometries, there is considerably more experimental data available for sphere entry.
Such data is a prerequisite for validation of numerical simulation models. The result of special
focus here is the drag coefficient.
The drag coefficient presented here is the ratio of the drag force to the momentum flow rate.

17

When plotted as a function of the dimensionless penetration depth, it was found that the drag
coefficient results are independent of the sphere velocity, surface tension, flow regime (either
laminar or turbulent), and Reynolds number. These remarkable findings can be attributed to the
fact that the drag force is due to momentum transfer between the moving sphere and the water
into which sphere is intruding. Friction-based drag is negligible in the situation under study.
In a graphical presentation of the results, it was shown that in the very early stages of sphere
penetration, very high values of the drag coefficients are encountered which lead to high
retardation forces on the object. These high forces tend to slow the fall rate of objects that
impact the water surface. Correlation equation in algebraic form were developed to quantify the
relationship between the drag coefficient and the dimensionless depth of immersion of the sphere
The model and simulation methodology developed in this study can be extended to other
geometries such as expendable bathythermograph (XBT) devices which are used to measure
ocean temperature distributions. XBT devices are released from a wide range of heights above
the water surface with a corresponding extensive range of impact velocities on the surface. It is
expected that the model and methology set forth in this paper will facilitate the determination of
the impact and subsequent penetration of various objects of practical utility. The full penetration
results from the entry model will serve as the starting condition for a fully immersed model of
the trajectory of the object.

References
Abelson, H.I., 1970. Pressure measurements in the water-entry cavity. J . Fluid Mech. 44, 129-
144.

Abraham, J .P., Gorman, J .M., Reseghetti, F., Trenberth, K., Minkowycz, W.J ., 2011. A new
method of calculating ocean temperatures using expendable bathythermographs, Energy and
Environment Research, 1, 2-11, 2011.

Abraham, J .P., Gorman, J .M., Reseghetti, F., Sparrow, E.M., Minkowycz, W.J ., 2012a. Drag
coefficients for rotating expendable bathythermographs and the impact of launch parameters on
depth predictions, Num. Heat Trans. A. 62, 25-43.

Abraham, J .P., Gorman, J .M., Reseghetti, F., Minkowycz, W.J ., Sparrow, E.M., 2012b.
Turbulent and transitional modeling of drag on oceanographic measurement devices,
Computational Fluid Dynamics and its Applications 2012, article ID 567864,
doi:10.1155/2012/567864.

Baldwin, J .L., 1975a. Vertical water entry of some ogives, cones, and cusps. National Technical
Information Services, U.S. Department of Commerce, AD-A009 300.

Baldwin, J .L., 1975b. Prediction of vertical water-entry forces on ogives from cone data, NSWC
Technical Report, Naval Surface Weapons Center, Silver Spring, MD.

Baldwin, J .L., Steves, H.K., 1975. Vertical water entry of spheres, NSWC Technical Report,
Naval Surface Weapons Center, Silver Spring, MD.

18

Bergmann, R., Van Der Meer, D., Gekle, S., Van Der Bos, A., Lohse, D., 2009. Controlled
impact of a disk on a water surface: cavity dynamics. J . Fluid Mech. 633, 381-409.

Courant, R., et al., 1945. The force of impact on a sphere striking a water surface, approximation
by the flow about a lens. AMP Report 42.1R, prepared for: Applied Mathematics Panel, National
Defense Research Committee, prepared by: Applied Mathematics Group, New York University.

Do-Quang, M., Amberg, G., 2010. Numerical simulation of the coupling problems of a solid
sphere impacting on a liquid free surface. Math. and Comp. in Sim. 80, 1664-1673.

Duclaux, V., Caille, F., Duez, C., Ybert, C., Bocquet, L., Clanet, C., 2007. Dynamics of transient
cavities. J . Fluid Mech. 591, 1-19.

Gekle, S., Gordillo, J .M., van der Meer, D., Lohse, D., 2009. High-speed jet formation after solid
object impact. Phys. Rev. Lett. 103, article 034502.

Gilbarg, D., Anderson, R.A., 1947. Influence of atmospheric pressure on the phenomena
accompanying the entry of spheres into water. J . Appl. Phys. 19, 127-139.

Guo, Z., Zhang, W., Wei, G., Ren, P., 2012. Numerical study on the high-speed water entry of
hemispherical and ogival projectiles. AIP Conf. Proc. 1426, 64-67.

Grumstrup, T., Keller, J .B., Belmonte, A., 2007. Cavity ripples observed during the impact of
solid objects into liquids. Phys. Rev. Lett. 99, article number 114502.

Hughes, O.F., 1972. Solution of the wedge entry problem by numerical conformal mapping, J .
Fluid Mech. 56, 173-192.

Koehler, B.R., Kettleborough, C.F., 1977. Hydrodynamic impact of a falling body upon a
viscous incompressible fluid. J . Ship Res. 21, 165-181.

Korobokin, A.A., Pukhnachov, V.V., 1988. Initial stage of water impact. Ann. Rev. Fluid Mech.
20, 159-185.

May, A., Woodhull, J .C., 1948. Drag coefficients of steel spheres entering water vertically, J .
Appl. Phys. 19, 1109-1121.

May, A., 1951. Effect of surface condition of a sphere on its water-entry cavity. J . Appl. Phys.
22, 1219-1222.

May, A., 1975. Water entry and the cavity-running behavior of missiles. National Technical

May, A., 1952. Vertical entry of missiles into water. J . Appl. Phys. 23, 1362-1372.

19

## McGehee, J .R., Hathaway, M.E., Vaughan, V.L., 1959. Water-landing characteristics of a

reentry vehicle. NASA Memorandum 5-23-59L, National Aeronautics and Space

Menter, F., 1994. Two-equation eddy-viscosity turbulence models for engineering applications,
AIAA J. 32, 1598-1605.

Moghisi, M., Squire, P.T., 1981. An experimental investigation of the initial force of impact on a
sphere striking a liquid surface. J . Fluid Mech. 108, 133-146.

Nisewanger, C.R., 1961. Experimental determination of pressure distribution on a sphere during
water entry. NAVWEPS Report 7808, U.S. Naval Ordinance Test Station, China Lake, CA.

Park, M.S., J ung, Y.R., Park, W.G., 2003. Numerical study of impact force and ricochet behavior
of high-speed water-entry bodies. Comp. and Fluids. 32, 939-951.

Richardson, E.G., 1948. The impact of a solid on a liquid surface. Proc. Phys. Soc. 61, 352-367.

Shi, H.H., and Takami, T., 2001. Some progress in the study of the water entry phenomenon.
Exp. In Fluids. 30, 475-477.

Shiffmann, M., Spencer, D.C., 1945a. The force of impact ton a sphere striking a water surface.
AMP Report 42, 1R. AMG-NYU, No. 105.

Shiffmann, M., Spencer, D.C., 1945b. The force of impact ton a sphere striking a water surface.
AMP Report 42, 2R. AMG-NYU, No. 133.

Shiffmann, M., Spencer, D.C., 1947. The flow of an ideal incompressible fluid about a lens.
Quart. Appl. Math. 5, 270-288.

Stark, J ., Gorman, J ., Hennessey, M., Reseghetti, F., Willis, J ., Lyman, J ., Abraham J ., Borghini,
M., 2011. A computational method for determining XBT depths. Ocean Sci. 7, 733-743.

Toscott, T.T., Epps, B.P., Techet, A.H., 2012. Unsteady forces on spheres during free-surface
water entry. J . Fluid Mech. 704, 173-210.

Trilling, L., 1950. The impact of a body on a water surface at an arbitrary angle. J . Applied Phys.
21, 161-170.

Verhagen, J .H.G., 1967. The impact of a flat plate on a water surface. J . Ship Res. 11, 211-223.
Watanabe, S., 1934. Resistance of impact on water surface. Part V Sphere, Inst. Phys. Chem
Res., 23, 202-210.

Von Karman, T., 1929 The impact on seaplane floats during landing. NACA Technical Report
321, Washington, DC.

20

Wagner, H., 1932. Phenomena associated with impacts and sliding on liquid surfaces. Z. Angew.
Math. Mech. 12, 193.

Waugh, J .G., Stubstad, G.W., 1966. Water-entry cavity modeling, Part 1. Vertical cavities.
NAVORD Report 5365, Armed Services Technical Information Agency, Dayton, OH.

Worthington, A.M., Cole, R.S., 1897. Impact with a liquid surface, studied by the aid of
instantaneous photography. Phil. Trans. Roy. Soc. Lond. 189, 137-148.

Worthington, A.M., 1908. A study of splashes. Longmans Green and Co., London.

Xiao, H., Zhang, X., 2012. Numerical investigations of the fall rate of a sea-monitoring probe.
Ocean Eng. 56, 20-27.