You are on page 1of 11

Partial cavitation as drag reduction technique and problem

of active flow control

Eduard Amromin , Mechmath LLC, Edmond, OK 73034, and Igor Mizine, member, SAIC,
McLean, VA

ABSTRACT

Partial cavitation can substantially reduce the ship total drag in a wide range of her speed.
Vented partial cavitation manifested certain advantages in comparison with other drag reduction
techniques. Its successful employment, however, requires both a special hull design and
development of an active flow control system. Such a system will stabilize the cavity under
perturbations of incoming flow in seas. The paper contents an analysis of achievements in drag
reduction, description of design fundamentals of ships with cavitating hulls/elements and
suggestions on application of ship drag reduction by cavitation.

NOMENCLATURE INTRODUCTION
CR = coefficients of residuary drag Ship drag reduction is an eternal problem, because
CF = coefficients of friction drag customers would like to move faster and shipowners
CL =lift coefficient would like to reduce fuel consumption. Drag reduction
Cp= pressure coefficient in water has been a goal of many efforts. Any
D=ships displacement successful effort really leads to reduction of one drag
D =caliber of a body of revolution component with a small increase of another. Since
Fn=roude number Froude, the ships drag is divided into two components.
H= submergence of hydrofoil/body The first one is friction drag. The second one is
L= ship total length residuary drag generated by pressure applied to the
l=cavity length hull:
N={Nx,Ny,Nz} normal to a boundary
PC=pressure in cavity U 2
Rx = (C R + C F ) SW (1)
P=unperturbed water pressure 2
Q=intensity of sources
QA= air consumption An efficiency of ship motions is defined as the ratio:
Rx = total ship drag
gD 2 D / L3 1 1
RC= cavitating drag = = (2)
Rn= Reynolds Number SW C X U / 2 U / gL C X SW / L2
2 2

S* = the cavity surface There is no solely ultimate drag reduction technique for
SW = wetted surface area all ships and ranges of speed, but a best solution
U =ships speed actually depends on the kind of designed vehicle and its
U-water velocity prescribed speed. Nevertheless, the friction drag
=velocity potential reduction is usually the most important aspect for whole
= cavitation number drag reduction and the goal of numerous attempts of
A=air density researchers (though the friction reduction sometimes
= water density leads to an increase of residuary drag).
According to Eqs.(1) and (2), there are two
general paths in friction reduction at a fixed ship speed.

1
Marine Technology 2

The first path is reduction of wetted surface. The second thickness, micro-bubbles unavoidably increase the form
path is reduction of the friction coefficient. The most resistance. Riblets must suffer from biofouling. The
impressive recent results in drag reduction are positive effect of polymer ejection critically depends on
summarized in Tab.1. These results were obtained with distribution of ejection points. For a high effect, any
the use of both paths. streamline over the hull must have several points of
ejection and these points must be tuned to both the drift
Table 1: Achievements in drag reduction angle and the actual value of Rn. Such tuning was
Methods Positive effect successfully carried out in quoted experiments with 2-D
Partial Theoretically partial cavitation allows up to and axissymmetric steady flows. However, it is a
cavities 80% of friction reduction in 2-D flows. challenge to make a similar tuning for 3-D ship hulls
Vented cavities can be adjusted to a wide under unsteady incoming flows in real seas.
range of speeds. It gave from 15% of total Although drag reduction by vented cavitation has
drag reduction for low-speed barges to 30- been successfully applied to shipbuilding for several
40% of total drag reduction for high-speed decades, tuning to perturbations in the real seas is an
patrol boats (Ivanov et al, 1996). issue for vented cavitation, too. Generally speaking, this
Riblets Riblets reduce up to13% of friction in issue limits both an eventual band of application and the
model tests. This method was also attainable effect of drag reduction by cavitation. It is,
combined with other methods (Mizunuma however, possible to extend this band and increase the
et al, 1999). drag reduction by introducing an active flow control
Polymer Theoretically this allows 62% friction system. Such system is the entirely novel topic for
ejection reduction in 2-D flow (Amromin et al, cavitating flow, and this paper explains its
1999). Measured maximum drag reduction fundamentals. This explanation also requires emphasize
was 55% for a flat plate (Fruman & Tulin, the key points of drag reduction by cavitation.
1976) and 35% in axissymmetric motion of
a body of revolution (Dronov et al, 1996). PRINCIPLES AND LIMITS FOR DRAG
Micro- Friction reduction of up to 80% was REDUCTION BY CAVITATION
bubble measured for a flat plate (Dimotakis, 1999) A refined classification exists for cavitating flows,
injection and up to 20% for a ship model (Watanabe, but there are two principal types in studies on drag
1998), but total drag reduction was 5%. reduction: Partial cavitation and supercavitation. A
supercavity covers the ship/body stern and deflects
The cavitation reduces SW, as well as employment water jets from the major part of the submerged
of underwater wings (Amromin et al, 1999). It is easy to body/ship surface. There is a large gaseous bubble over
show the wing efficiency rises as Fn2. There are, such parts and there is no noticeable friction on the
however, limitations for hydrofoil ships displacement surfaces covert by cavities. The jets, however, merge
because of existence of the structural restriction behind the supercavity. A merging region is highly
(Besnard et al, 1998) and, as a result, payload of unstable. The reverse jets appear there and cause a
hydrofoil ships falls behind the displacement to drag significant drag that can be described by a coefficient
ratio E=D/Rx. Drag reduction by cavitation has no CD. A possible friction reduction occurs when the
restriction related to D increase. An important cavitation- caused friction decrease is higher than the
advantage of ship drag reduction by cavitation consists drag increase caused by cavitation. It is easy to make a
of several decades of its employment to full-scale ships comparison for a body of revolution with the
of diverse displacements and at diverse speed. head/cavitator of radius r. Thus, it should be
Polymers, micro-bubbles and riblets reduce the U2r2CD/4<U2SWCF, but there is
friction coefficient at high Reynolds numbers. These SWlR*=r2CF(2R*/r)2l/R*. Let us employ the
drag reduction techniques really affect boundary layer dependencies (Knapp et al, 1970) of the cavity length l
turbulence. There are also other techniques to control and its maximum radius R* on cavitation number. As a
turbulence, but the mentioned techniques have shown result, the following condition appears:
the positive effects in numerous model tests and several (-0.133/2) (0.066+1.72)/ (+0.008) < 4CF (3)
full-scale experiments with underwater missiles
The noted advantages are, however, coupled with One can see with Eq.(3) that for the value CF ~ 0.002
disadvantages. The micro-bubbles shown an excellent (the usual value for ships), drag reduction by
efficiency on the bottom side of horizontal plates, but supercavitation is profitable for <0.05. What is this
their efficiency highly depends on orientation of the limitation in terms of ships speed and displacement?
wetted surface. Besides, increasing the boundary layer The answers can be done with regards to the definition
of cavitation number:

2
Marine Technology 3

P PC in a wider range of speeds (Partial cavity length is


=2 (4) smaller than the hull/body length). First of all, it is
U 2 possible to generate a partial cavity of zero cavitating
drag, because cavitating drag is proportional to intensity
The unperturbed pressure P depends on the
of the reverse jet (Knapp et al, 1970) and it is possible
body/hull submergence only, whereas the pressure in
to design a body without reverse jet behind a partial
cavity PC can be either controlled by gas ejection, or
equal to the vapor pressure. For the last case, PC is cavity at the selected value of . The friction reduction
will be then proportional to the ratio of cavity surface to
much smaller than P and can be neglected in
the entire body surface. One can consider examples for
estimations. Submerged supercavitating hulls were
studied even for the surface ship (Varghese, 1999), and Fr= in Fig.1 as the examples of such cavities and
for a 5-meter hull submergence, one can calculate with estimate a more than 80% drag reduction for =0.06
Eq.(4) that a vapor supercavity at <0.05 exists for a and 70% for =0.08, because there is no cavitating drag
ship speed over 150 knots. A vented supercavity can with such ideal cavity. The shedad part of side surface
exist at 75 knots with a significant gas ejection in cavity and disks compose the wetted surface in these
only (to keep PC0.75 P), but a high air consumption examples. For a finit Fr, however, the flotation of the
(and attendant energy losses for compressors) is not the cavity tail is also an issue. Two schemes of cavity
main trouble here. The key trouble is gravity force generation shown in Fig.2 help to avoid this issue.
effect on the cavity shape. Principally, semi-submerged keels can give a greater
drag reduction.

Figure 2: Two concepts of cavity generation. A ship


section is shaded. The curve is the section of cavity
that covers a flat bottom only. The hull sides are also
Figure 1: Sections of the cavities behind a disk. Sb is covered by the cavity with a surface section , but the
the disk section. Solid lines are sections of keels 1 and 2 give some increase of the wetted surface.
axissymmetric cavities for FrU/(gr)1/2= and =0.08
Historically, drag reduction by partial cavities has
(smaller cavity) and 0.06 (larger cavity). The length of a
been first developed for the flat bottom (Ivanov et al,
fictitious cone behind the axissymetric cavities equals to
1996). There were a lot of failed attempts to cover the
1. The dashed curve is cavity section for Fr=1 and
entire bottom by cavities past a single
=0.08.
appendage/cavitator at small U and small (by
augmentation of PC) before a key role of Fn (in the
This effect is shown in Fig.1 for cavities behind a
range of its values typical for ships) was discovered
disk. The presented computation (Amromin et al, 1996)
(Butuzov, 1966). Roughly generalizing this discovery,
relates to Fr=1. It is easy to see that approximately 60%
one may say that such a cavity must be shorter than
of side surface out of cavity there. It occurs due to the
U2/g at any . A sketch (from Ivanov, 1980) of a
gravity force effect. For this example, rL/13 and for a
large bottom with cavities is plotted in Fig.3.
60-knot speed, for example, L80m corresponds to
The cavity system plotted in Fig.3 was first
Fr=1. Thus, the leading third of hull can be covered by
installed at a Soviet barge and successfully tested (with
a supercavity for this Fr (the vertical deflection of
15-17% of fuel saving) in the Volga river more than 30
cavity is generally proportional to Fr-2~r~B). Therefore,
years ago. Nevertheless, the real customer was later
supercavitation is a promising drag reduction technique
found for smaller and faster ships that were supplied by
for very high U and small B inherent to missiles, but
a solely partial cavity. For a set of gliders (patrol boats),
this technique meets great difficulties for ship
the obtained drag reduction has increased up to 30%.
applications.
Partial cavities can be a drag reduction technique

3
Marine Technology 4

analysis of cavitation over this body. Looking for the


large cavities past some appendages, one can employ
ideal fluid theory for this analysis. The key aspects are
the same in 2D and 3D cavitating flows of ideal fluid,
but it is easier to describe (and understand) these
aspects in 2D approach. Thus, the core description is
made for 2D problems. It is usual to make an
assumption that the flow can be determined by using a
Figure 3: Sketch of a beam section of bottom with dimensionless velocity potential (related to the body
multiplies partial cavities. Cavitator-wedges (dark size and free-stream speed). The governing equation
triangles) must be completely wetted to generate and the boundary conditions are then following:
cavities and a wetted gap between cavity tails and
wedges must also exist. =0 (5)

It is evident that the drag reduction by partial =0 (6)
cavitation can rise together with the cavity length only, N S
but there is no monotonous dependency. Vice versa, for grad x = {1,0} (7)
any shape of appendages down to flat bottom and pair
{, Fn}, there is a very narrow range of cavity length
with negligible cavitating drag (a curve computed by ( grad )2
= 1+ (8)
S*
Ivanov (1980) is presented in Fig.4).
The surface S includes the cavity surface S* (that
1 is a surface of the constant pressure) and the really
0.8 wetted surfaces, but a fictitious boundary in the tail of
Drag coefficient

0.6
cavity is also included in S. The same ideas were
employed for cavities over a hydrofoil and down the
0.4
hull. Thus, the fictitious body (contour) is a common
0.2
feature of cavity schemes in Fig. 3 (where this body is
0 the wedge of length b1 on the bottom; there is a closed
2 3 4 5
scheme) and in Fig.5 (there is an open scheme for a
Cavity length / wedge length
hydrofoil). For a drag-free cavity, either the fictitious
Figure 4: Effect of cavity length l on the drag contour disappears (for closed schemes), or merges with
coefficient CC caused by vented cavitation behind the wetted surface (for open schemes).
wedge down the bottom. The value of CC is related to its 0 .1 2

value for cavity length equal to 2.02 length of the 0 .1

wedge. 0 .0 8

0 .0 6
y

One may fear that even a limited perturbation of 0 .0 4

0 .0 2
incoming flow could overturn this drag reduction. Thus,
0
there are two coupled hydrodynamic problems in ship 0 0 .2 0 .4 0 .6 0 .8 1
-0 .0 2
drag reduction by partial cavitation. The first one is to x
fit the hull shape (and appendages/cavitators) for drag
reduction at the given values of and Fn. Such problem Figure 5: Open scheme of partial cavity. Streamline
is basically the hull sections design problem, but it can does not attach at hydrofoil (solid line) downstream the
be also named as the problem of passive flow control. cavity (dashed line), and drag rises with the distance
The second one is to keep this reduction under between fictitious body (market by triangles) and the
perturbations of real flows by some actuators, ejection, hydrofoil trailing edge.
suction, etc. Such problem can be also named as the
problem of active flow control The cavity shape and thickness are initially
unknown. Let us assume that the velocity distribution
PASSIVE CONTROL OF CVITATING FLOW: U(x) over cavitation-free body/hydrofoil is already
SHIPS ELEMENT SHAPE TUNNING FOR A computed and the thickness of the cavity h is much
SELECTED CAVITATION NUMBER smaller than its length. The boundary conditions (6), (8)
Design of a cavitating body with the ordered can then be simplified with the use of an auxiliary
hydrodynamic performances requires a numerical velocity potential of density q (deduction of simplified

4
Marine Technology 5

conditions is described by Amromin & Bushkovskii, For a hull or hydrofoil near the water surface, the
1994). The function h(x) is then the following function pressure constancy condition on this surface must be
added to Eqs. (5)-(7). However, if the cavity
of U(x) and U*(x)= 1 + + yg / U :
2
submergence H>>h, and lg<<U2, this condition can be
satisfied by distribution of sources q(x) along the line
1
x
y=H over this surface. An example of design for
h( x ) =
U ( x)
h( x0 )U ( x0 ) + q( )d
(9) moderate H value is shown in Fig.7.
x0 The corresponding computed pressure distribution
l + x0 at the design angle of attack is plotted in Fig.8 (actually,
R( x) U * ( ) U ( ) d
q ( x) =

x0 R( ) x
(10) this angle is 2 degrees). It is easy to see that a cavity
spreads over the suction side up to x=0.75C at the speed
that corresponds to =0.2. For example, for H=2m and
Here R ( x ) = (l + x0 x)( x x0 ) , x0 is the cavity =0.2, U=34m/s (the effect of gravity force is
negligible, because lg/ U2<0.019).
detachment point (location of cavitator). The formula
(10) is the solution of a singular integral equation. Such 6
solution exists under the following condition:
5
l + x0
U * ( ) U ( )

Cavity length
4

x0 R( )
d = 0 (11)
3

2
Finally, it is possible to represent h as a direct function
of U: 1

0
yg
l + x0 1 + + 2 U ( ) 0.2 0.24 0.28 0.32

R( )
x
U Cavitation number
h( x ) =
x
x0
U ( x) R( )( )
dd
0
(12) Figure 6: Cavity length for ellipsoid of revolution
z2/4+r2=1; 5- observation (Ivanov, 1980), g - our
However, the values of l and are coupled by the computation with cavity detachment located in observed
condition points; - our computation for cavity detachment
l + x0 determined with the classic Brilluin-Villa condition.
1 + + zg / U 2 U ( )

x0 R ( )
d = 0 (13)
0 .0 4 5

0 .0 3
The above-described procedure can be employed
once (as Tulin (1964) did) or in iterations (as Amromin 0 .0 1 5
y

et al (1999) did). For iterative approach, U(x) is 0


corrected by solving Eqs.(6)-(8) in changed boundaries, 0 0 .2 0 .4 0 .6 0 .8 1

and h becomes a correction to the cavity thickness). -0 .0 1 5


x
The ideal fluid approach satisfactory describes the
cavity shape and forces if the cavity detachment is Figure 7: Hydrofoil JENNY-2034 designed (Amromin
correctly fixed (see Fig. 6). For cavities behind et al, 1999) for H=0.5C and =0.2.
appendages, it is certainly takes place.
Returning to Eqs.(9) and (10), one can see that for
a body where U(x)/U*(x)-1<<1, the cavity is thin and
drag is low. Let us design a body surface that coincides
with such a cavity surface. According to the above
analysis, this body must have advanced cavitation
performance. Therefore, design of ships elements with
improved cavitation performances through solving
Eqs.(5)-(8) found many applications (starting from Shen
& Eppler, 1981).

5
Marine Technology 6

This 2D consideration can be generalized for 3D flows


1 around ships hulls by the path previously used for
cavitating 3D flows on the marine propeller blades
(Amromin et al, 1995): The velocity U must be
determined from 3D linear problem (5)-(7). Further, the
0.5 shapes of cavities will be found from 2D problems on
cavitation in non-uniform incoming flow.
Cp

However, the dependencies from Fig.9 (for the


0 sole cavity) manifest that the 2% fluctuations of the
speed may cancel this drag reduction. Further, a half-
0 0.25 0.5 0.75 1
degree increase of the angle of attack completely
cancels it. One can link this cancellation with the
-0.5 increasing of cavity length (see Fig. 11) and
x transformation of the partial cavity into supercavity.
(For CL =0.3 the minimum abscissa of cavity end is
X/C=0.73). However, for CL =0.26 the cavity end is the
Figure 8: Pressure distribution over hydrofoil JENNY- same, but the cavity detachment x0=0.6C and such short
2034 at the design angle of attack. cavity really gives no drag reduction.
Computation shows the significant drag reduction
for the design speed and angle of attack (see Fig.9). For
H=2m and U=34m/s, the cavity can be vapor-filled.
The drag reduction would be increased by generation of
a vented cavity on the pressure side. It is easy to
estimate with Fig.8 that such cavity will corresponds to
=-0.1. According to Eq.(4), this cavity must be
generated by air ejection with the pressure that exceeds
the unperturbed hydrostatic pressure at this H.
Particularly, for the considered example, PC is
approximately 1.8 of the atmosphere pressure. The gas
ejection is also necessary to generate a cavity with >0
at moderate ship speed. Let us return to Fig.1 and Figure 10: A modified hydrofoil GENNY-2034. Partial
analyze cavities at =0.08 and H=2m. For U=25m/s, cavities are showns as dark strips. The cavity
PC should then be 1.45 of the atmosphere pressure. detachments over suction side is spreaded in a cutting,
the cavity head is fixed by a step. The vented cavity on
4 the pressure side is generated by an appendage-wedge.

0.5
3
Rx / Rxo

cavitation number

0.4
2
0.3
1
0.2

0 0.1
0.95 0.97 0.99 1.01 1.03 1.05
0
Vs / Vsd 0 0.2 0.4 0.6 0.8 1
l /C
Figure 9: Drag coefficient of 2D hydrofoil JENNY-
2034 in the vicinity of design speed Vsd.
Figure 11: Cavity end location over hydrofoil JENNY-
Thus, two cavities on two sides of the hydrofoil 2034 as a function of the cavitation number for H/C=0.5
could give up to 75% of friction reduction with very (bottom curve for CL=0.345, top curve for CL=0.39).
moderate energy consumption on keeping the air The similar issues exist for cavities down the
pressure into the vented cavity (see a scheme in Fig.10). ships bottoms, and for locking the cavity edges, the

6
Marine Technology 7

niches were designed there. A niche, however, gives yg


significant additional drag at smaller speeds (where the I + x0 l + x0 + U ( )
drag reduction by cavitation is generally possible, but a R( ) U 2
high influence of Fn would require another design of
x
x0
U ( x0 ) R( )( )
dd = h0 (14)
appendages, etc). Thus, an optimum shape of hull 0
cannot be selected once for all, but a tuning to the
Here h0=h(x0), =1+. Besides, Eq.(13) must be
varying incoming flow is necessary.
satisfied (this condition geometrically means
dh/dx(x0)=dh/dx(l+x0)=0 [14]). Therefore, two hidden
ACTIVE CONTROL OF CAVITATING FLOW:
parameters must be introduced to satisfy Eqs. (13) and
CAVITY STABILIZATION BY VARYING GAS
(14). There is a freedom in their selection because of
PRESSURE AND/OR SHAPE OF APPENDAGES
possibility to vary by tuning PC, and vary U(x) by
Such a tuning is often named as an active flow
shapingf the cavitation-free body surface, as well as h0
control. The active flow control for low-drag cavities is
by advancing an appendage at the cavity detachment.
a novel research topic that was first described recently
There is also an effective method to affect the cavity
and very briefly (by Amromin, 1999, and Amromin et
length by varying the slope at the cavity detachment
al, 1999). The two-decade history of drag reduction by
point (see Fig.13).
partial cavitation in Soviet Union did not include such
Although the flow active control can be developed
topic because of the very skeptical view of Soviet
on the basis of ideal fluid theory and be considered as
admirals on the capability of any computer-driven
inverse hydrodynamic problem, such a problem is
devices.
principally different from the classic inverse problems
It is appropriate to start from the fundamentals that
(like Shen & Eppler (1881) problem). For this control,
can be first done for 2D flow of ideal fluid. The basic
request to the active control of cavitating flow is to keep besides of possible variations of (which effect is
h(x0+l)=0 that physically means the absence of the illustrated by Fig.10), it is necessary to affect the
reverse jet in the cavity tail. There are both theoretical velocity distribution on S* by variation of surface shape
arguments and experimental data to show the out of S*. Such a problem is an incorrect problem in
significance of this request, and a strong proportionality potential theory, and there are many solutions of this
between measured cavity length oscillations and problem.
computed h(x0+l) was found (by Amromin, Briancon- 0.4
Marjolet & Vacilyev, 1994) and shown in Fig.12.
0.3

0.6
l /C

0.2

0.1
0.4
0
0.04 0.06 0.08 0.1 0.12 0.14
0.2 K

Figure 13: Cavity length versus relative variation of


0
cavity number K=-d for different slopes of streamlines
1.5 1.7 1.9 2.1
at cavity detachment point on hydrofoil NACA-0010.
Cavitation number Solid line corresponds to 26o, dashed to 37 o.

Figure 12: Proportionality between measured It is evident that a pressure gradient may become a
oscillation of cavity length (rhombs) and computed lock for the cavity tail. Such gradient can be basically
cavity thickness h(x0+l) at the cavity end (solid curve induced either by a vortex, or by a source. Distributing
shows 50 h(x0+l); the computations were carried out such singularities out of S*, it is possible to take into
with a closed scheme, and h(x0+l) is equal to the account their contribution to U(S*) and consider their
fictitious body size). intensity as the decision variables in Eqs. (13) and (14).
According to Eq.(12), this request can be rewritten as: For example, it was suggested (Amromin et al, 1999) to
generate the pressure gradient by some kind of flap
located near the hydrofoil trailing edge and coupled
with a spoiler located at the leading edge. A flap can be

7
Marine Technology 8

modeled as a vortex distribution, its intensity can be cavitation).


determined with the use of Eqs. (13) and (14). Of A relatively small surface deformation can be
course, such a flap provides additional wetted surface mathematically described with the employment of
and friction (as shown in Fig. 14). source distributions in the same simplified equations.
Let us denote an unsteady perturbation of velocity on
40 the time-averaged shape of S* as U**. One can write
their following generalization of Eqs. (9) and (10):
wetted area ( % )

30
d
q* = 2 [U * ( x)h( x)] (15)
20 dx
gh( )
10 l + x0 d U * *( ) +
R ( x) U 2
0
q* =

x0 [x ]U * ( ) R( )
d (16)
0.16 0.18 0.2 0.22
cavitation number Here h* is the thickness that is explained by the sketch
in Fig.15. The thickness h** includes the variation of
the appendage thickness for x<x0, but it is a given
Figure 14: Additional wetted surface provided by flaps
function that smoothly drops from h**(x0) down to 0
that keep the partial cavity drag-free. Dashed curve
along a given interval (approximately equal to the
relates to l=0.73C, solid curve to l=0.82C.
appendage length). The value d is variation of
cavitation number, h/=h* - h**.
Nevertheless, comparing Figs.9 and 13, one can
The velocity perturbation U** appears due to two
find out that even a 15% variation of speed does not
circumstances. The first one is incoming flow
completely cancel the drag reduction by cavitation with
perturbation. Let us transform the right-hand side in
employment of the active flow control, whereas a 2%
Eq.(7) into {1-akek(y-y0)sin(kx-t), akek(y-y0)cos(kx-t)},
variation of speed without such control cancels the
for example. Such transformation leads to changes in
reduction.
solution of Eqs.(5)-(7). Thus, the velocity U*(x) must
obtain a variation u(x) on S*. The second circumstance
is caused by the displacement h** and can be calculated
as the following integral:
l + x0
1 U * ( )h * *( )
u * *( x) =
x0 b [x ]
d (17)

where b is the appendage length. Substituting the


quantity U**=(u+u**)/U* in the right-hand sides of
Figure 15: Scketch of cavity change by active control Eqs. (15), (16) and (17) and excluding q* from these
system. Line ABCDE is the hull section. Line BE is a equations, one will obtain an equation for h/(x) with two
neutral position of appendage. Line FD is the section of
undetermined parameters { h**(x0), d }. However, it is
unperturbed cavity. The dashed line shows deflected
easy to rewrite the condition h/(l+x0)=0 with the
position of appendage and perturbed cavity section. The
employment of these parameters and the following
shaded area thickness is h*(x) from Eqs.(15), (16).
existence condition for the solution (16):
Another possibility to affect U(S*) is deformation l + x0
d U * *( ) + gh( ) / U 2
of the surface by some embedded actuators, and it
should not increase the wetted surface. The employment

x0 U * ( ) R( )
d = 0 (18)

of such actuators from smart materials/structures is


broadly developed in aviation (Yu et al, 1997). Of Of course, 3D consideration for ship hull will
course, there is a substantial difference between involve 2D integrals. Besides, a precise estimation of
materials in aviation and shipbuilding, as well as scale effects on cavitation will require corrections to
between flow-induced loads in air and water. make it possible to take into account the Reynolds
Nevertheless, the similar actuators can be applicable to number influence on h, etc. Nevertheless, one can see
affect cavitating flows, because their location has to be that the problem of determination of hidden
in zones of small pressure (especially, for vapor parameters of the active flow control system is

8
Marine Technology 9

mathematically clear. assigned mainly to operation in shallow waters. A novel


Looking for energetically profitable design, one problem is to adjust drag reduction by vented cavitation
must keep in mind losses associated with the air to the wider diversity of hull shapes and to larger
consumption. For its estimation, let us represent the air displacements. In spite of a relative high fraction of
flux QA as a function of cavitation-caused drag RC and wave resistance in the total drag of the fast wide-deck
the air density A(PC): ships with displacement below 4000 tons, their friction
reduction is also important.
QA=CQRCA/[U] (19)
According to experimental data (Epshtein, 1970) for
disks, the coefficient CQ is between 0.6 and 1 for
>0.05 (however, this coefficient can rise over 10 for
<0.04). Taking a vented cavity with ~0.1 on a ship
moving at speed 15m/s as an example, one will receive
PC that is very close to Pa (the atmosphere pressure),
and the air density should be approximately of the
atmosphere density. Let us suppose that fluctuations of
cavitating drag could rise up to 50% of saved friction
Figure 16: View of the boat SERNA with the bottom
drag, and the ship displacement is about 2,000 tons.
adjusted to cavity existence.
Then the air consumption will be around 30 kg/s, and it
is not an issue.
The second type is large fast slender multi-hulls
Finally, it is necessary to emphasize two principal
(Mizine et al, 1999). There are different possible
aims of the considered active flow control system. The
combinations of hulls for such displacement ships (or
first aim is to make it possible to run the ship without
hulls and hydrofoils with hydrofoil-assisted ships), and
energy losses in a wide range of speed. It looks the
an effect of drag reduction by partial cavitation will be
removable flow control devices are good enough for
also very different. Let us get for an example recently
obtaining this aim. The second aim is to minimize the
(by Amromin, Mizine et al, 2002) considered (and
wavy sea negative influence on drag reduction by
undergone the model tests) variants of a 313-meter
cavitation (active control systems are already used to
trimaran assigned for speeds up to 70 knots. For a
improve fast ships seakeeping, see Arii et al, 1999 for
trimaran, there are effects of both the hull shapes and
example).
the hull mutual positions (see Fig.17).
Here the resistance coefficients are traditionally
SOME FUTURE APPLICATIONS OF SHIP DRAG
calculated with employment of the area of wetted
REDUCTION BY PARTIAL CAVITATION
surface at Fn=0). A small negative effect at small
The history of drag reduction by partial cavitation
speeds usually appears due to contribution of additional
let us emphasize on applications to fast ships. There are
devices (these devices are shown in Fig.2).
at least two types of displacement ships that are suitable
On the other hand, the drag reduction by cavitation
for this drag reduction technique.
can have also a cumulative effect because of a possible
The moderate speed and displacement are inherent
re-design of the entire ship. Particularly for this
to the first type of wide-deck ships (30-35 knots and
trimaran, a 25%-30% drag reduction at the highest
1000-4000 tons). One may note that this speed range is
speed allows installation of three waterjets instead of
already explored by Russian boats SERNA (see its view
four waterjets. The obtained profit is not limited by a
in Fig.16). One has to return to Tab.1 and emphasize:
payload augmentation. It also includes an opportunity to
30% is the total drag reduction of this low-draft ship,
decrease the stern fullness. Consequently, it would be
the friction reduction is higher; but there is no
possible to reduce the values of CR in the entire range of
significant attendant increase of the form resistance (in
ships speeds. As a result, a small rise of CR for Vs<40
contradistinction with friction reduction by micro-
knots (plotted in Fig.17) could be also compensated.
bubbles, where such increase cut 75% of friction
reduction of a ship model). SERNA is the ship of 100
ton displacement is able to run at 30 knots in calm seas
or at 27 knots in sea state 3 (for the sea state 5, the
cavity is not maintained, the bottom niche gives a high
additional drag, and the speed drops down to 8 knots).
Her operation range is from 100 miles with 45-ton
cargo to 600 miles with 25 ton cargo.
Nevertheless, such flat-bottom boats were

9
Marine Technology 10

3.6 ACKNOWLEDGMENT
The authors are grateful to Prof. Alexander Ivanov
3 for permanent discussions on cavitation during
approximately two decades (and we hope this paper
1000 (Cr+Cf)

2.4
would be a present to his close 80 anniversary). The
1.8 authors appreciate to late Dr. Anatoly Butuzov for his
1.2 decisive contribution to development of passive flow
control for drag reduction by partial cavitation. They are
0.6 also grateful to Dr. Theo Kooij for a possibility to start
in the development of active control system for
0
cavitating flows.
30 40 50 60 70
Speed(knots) REFERENCES:
1. Amromin, E. L., Khodorkovsky, Y. S. ,
Figure 17: Computed total drag coefficient for diverse Kovinskaya, S. I. 1999 Fast Ship Drag
variants of a fast large trimaran (KMM project). Dashed Reduction. Proceedings of FAST-99 Conference,
curve related for an initial design. The solid curve is the Seattle, 911-924
estimation for hulls with drag reduction by partial 2. Besnard, E. et al. 1998 Report AE-98-1, Aerospace
cavitation. Engineering Department< California State
University, Long Beach.
There is also a possibility to significantly increase 3. Ivanov, A.N.& Kalyzhny, V.G. 1996 The
the lift to drag ratio of hydrofoil ships by using vented Perspectives of Applications of Ventilated Cavities
cavities at their highest speed (because a fraction of the on Naval Combat Ships, Intl. Conf. on 300th
inductive drag decreases at such speed (Coopersmith & Ann.of Russian Navy, St. Petersburg, A2, 41-46
Harris, 1998) and significance of friction reduction 4. Mizunuma, H., Ueda, K., Yokouchi, Y. 1999
decreases). It is, however, important to keep in mind Synergetic Effects in Turbulent Drag Reduction
that the active flow control must play the key role for by Riblets and Polymer Additives. J. Fluids
implementation of vented cavities on all ships assigned Engineering, 121, 535-540
to operate in open seas. 5. Fruman, D. H. & Tulin, M. P. 1976 Drag
reduction by Polymer Diffusion at High Reynolds
CONCLUSIONS Numbers, J. Ship Research, 20,171-180
Partial cavitation is the highly effective drag 6. Dronov, B.F., & Barbanel,B.A. 1996
reduction technique that has been successfully applied Development of Large-Scale Surfacing Models
to shipbuilding during three decades. The partial Tuna for the Research of Boundary Layer Control
cavities, however, are sensitive to variations of ships Methods. CRF-96 Conference. St.-Petersburg,
speed and perturbations by the incoming flow. Because Russia, 2, 322-331 ,
of this, a successful design requires the profound 7. Dimotakis, P. E. 1999 S&T Issues Summary: Gas
hydrodynamic knowledge and preliminary mathematical injection ONR workshop on Gas Based Surface
analysis. Ship Drag Reduction, Newport
It is necessary to design some hydrodynamic locks 8. Watanabe, K. 1998 Drag Reduction in Japan,
to stabilize the cavity under perturbations. The Proceedings of Intl. Symposium on Seawater Drag
traditional path has been to put the cavities into Reduction, Rhode Island, 19-23
specially designed niches, but this passive flow control 9. Knapp, R.T., Daily, J.W. & Hammit, F.G.
technique does not make it possible to get a maximum Cavitation. McCraw-Hill, NY, 1970
drag reduction and cannot be optimal for any shape of 10. Varghese, A.N. 1999 Supercavitation drag
ship hulls. The active flow control can give more reduction technology for high speed ship, ONR
flexibility to the hydrodynamic locks. As a result, it will workshop on Gas Based Surface Ship Drag
be possible to obtain the higher drag reduction and have Reduction, Newport
no losses for cavitation-free motion of the ships. 11. Amromin, E.L., Bushkovskii, V.A & Yakovlev,
The mathematical fundamentals of active flow A.Y. 1996 On the maximum possible reduction in
control for cavitating bodies/hydrofoil is described in the resistance of water to the motion of bodies.
this paper. The detailed development of the numerical Technical Physics, 41,499-501
technique on the basis of these fundamentals must be 12. Butuzov, A.A. 1966 Extreme parameters of vented
fitted to the actual projects. cavity on the top surface of horizontal wall, Fluids

10
Marine Technology 11

Dynamics, 1, 167-170 Helicopter Blade -Vortex Interaction Noise by


13. Ivanov, A.H. 1980 Hydrodynamic of Developed Active Rotor Control Technology, Progress in
Cavitating Flow. Sydostroenie, Leningrad (in Aerospace Sciences, 33, 647-687
Russian) 21. Epshtein, L.A. 1970 Similarity methods in ship
14. Amromin, E.L. & Bushkovskii, V.A. 1994 hydrodynamics. Sydostroenie, Leningrad (in
Approximate Solution of Three- Dimensional Russian)
Inverse Problem for Non-lifting Bodies with 22. Arii, T., Hatta,K., Sugano,M. & Miyata, H. 1999
Optimum Cavitation Characteristics. Fluid Development and Operation of Hydrofoil
Dynamics, 29, 373-379 Catamaran SUPERJET, Proceedings of FAST-99
15. Tulin, M.P. 1964 Supercavitating flow small Conference, Seattle, 511-522
perturbation theory. J. Ship Research, 7, 16-37 23. Mizine, I. O. & Amromin, E. L. 1999 Large High-
16. Shen,Y.T. & Eppler,R. 1981 Wing section for Speed Trimaran: Optimization Concept.
hydrofoils, Journ. .Ship Res., 35, 191-200 Proceedings of FAST-99 Conference, Seattle
17. Amromin, E.L., Vasilyev, A.V. & Syrkin, E.N. 24. Amromin, E., Mizine, I, Crook, L., Day, W &
1995 Propeller Blade Cavitation Inception Korpus, R. 2002 High-Speed Trimaran Drag:
Prediction and Problems of Blade Geometry Numerical Analysis and Model Tests. Be
Optimization: Recent Investigation at the Krylov published in J. Ship Research
Shipbuilding Research Institute, J. Ship Research, 25. Coopersmith, R.M. & Harris, B.W. 1998 Final
39, 202-212 Report N00014-97-C-0351, Lockheed Martin
18. Amromin, E.L. 1999 Analysis of Actuator Aeronautical System
Distribution for Cavitating Body Drag Reduction.
ONR workshop Gas Based Surface Ship Drag
Reduction, Newport
19. Amromin, E.L., Briancon-Marjolet, L. & Vacilyev,
A.V. 2004 Sheet Cavitation: Comparison between
Measured and Calculated Length. Intern.
Shipbuilding. Conf., St.-Petersburg, vB, 59-66
20. Yu,Y.H., Gmelin,B., Splettstoesser, W.,Philippe,
J.J.,Preur,J. & Brooks,T.F. 1997 Reduction of

11

You might also like