"'!

_DTIC_-"
1P
%L WQ r'L
Im
60-WA-265
JUNf 2 2 19941 I
F
i
Cavitation Along Slain
of Separation
APR 17 1
C DAVID W. APPEL
G Professor of Hydraulics,
, Department of Engineering
Mechanics, The University This document has been apprcved
of Kansas, Lawrence, Kan. for public release a.d sale; its
distribution
is unli-ited.
By means of a moving picture accompanying
this presentation,
the
author demonstrates
the development
of cavitation
in flow along a
two-dimensional
surface of separation.
This film has been chosen as
the best means of presenting
some of the qualitative
characteristics
of
this type of cavitation
which are important
to engineers concerned
with
cavitation
in diverse practical
situations.
A description
of the experi-
ments, some views from the movie, and a discussion
of the observations
of cavitation
in a zone of separation
are given in the paper.
The majority
of enginccrs
coniccrned
with various aspects of cavi-
tation are especially
interested
in well-streamlined
flows such as in the
passages
of pumps and turbines,
through tunnels,
over spillways
and
around the propellers
of ships and submerged
bodies.
Therefore,
it is
natural and best that primary attention
has been given to cavitation
as-
sociated
with these various boundary
forms. In many studies the pres-
ence of separation
and the possibility
of cavitation
originating
along
surfaces
of separation
has been recognized.
However,
mot of the
studies reported
in the engineering
journals
have not been concerned
primarily
with this ocurrence.
On the other hand, there has been much
interest in the basic structure
of flow in a zone of separation
in the ab-
sence of cavitation,
with much of the work in this area being of a theo-
reticalnature.
The exploratory
study described
here was undertaken
in
an attempt
to bridge the gap between
these two areas of endeavor.
Ini-
tially qualitative
information
both on cavitation
and on the develop-
ment of vortices
and/or turbulence
along surfaces of separation
was
sought by direct photographic
observation.
Even the initial results were
found to be very revealing.
To make them immzdiately
available
to
other investigators
as well as to engineers
in practice,
the films have
been prepared
for presentation.
Contributed
by the Hydraulic
Division
tr presentation
at the Winer Annual Meet.
Ing. New Yolk, N. Y., November
t7.December
2. 1960, ol The American
Society
of
Mechanical
Engineers.
Manuscript
received at ASME Headquarters,
August 30, 1960.
Written discussion
on this paper will be accepted up to January 10, 1"1
Copies will be available until October I, 1961. -
0 06
r4!
f.7 '
QULT~
~ 7~
-
DTTC
.i.
.9-
T 61-
X ~
4 ra A D- p., L L ~B ER 60-WA-265
fffB 189 UIHIII 2U 22 199
V ~~Cavitation Along SfaCe
of Separation
A~PR t.
DAVID W. APPEL
Professor of Hydraulics,
Department of Engineering _________________
Mechanics, The University This docLunen* has bee prv~
ansa, lwrene, an. for Pubic' release and sale; its
distribution is Lmnhiited.
By means of a moving picture accompanying this presentation, thc
author demonstrates the development of cavitation in flow along a
two-dimensional surface of separation. This film has beer, chosen as
the best means of presenting some of the qualitative choracterisics of
this type of cavitation which are important to engineers concerned with
cavitation in diverse practical situations. A description of the experi-
ments, some views from the movie, and a discussion of the observations
of cavitation in a zone of eparation are given in the paper.
Ci'l iiajorit,- ot engineers coixcrnd w itli v.arious aspccts of cavi-
ration are especially interested in well-streamiliitcd flows such as ini the
)aS3ages Of puisIP3 and turbines, through tunnels, over spillways and
:iround the propellers of ships and submerged hodies. Therefore, it is
natural and best that primary attention has becn given to cavitation as-
cociated with these various boundary forms. In many studies the pres-
enrce of separation and the possibility of cavitation originating along
surfaces of separation has been recognized. Hlowever, most of the
-tudie5 reported in the engineering journals have not heen concerned
rsrimarilv with this ocurrence. Onl the other hand, there has been much
interest in the basic structure of flow in a zone of separation in the ab-
sence of cavitation, with much of the work in this area being of a theo-
retical nature. The exploratory study described here was undertaken in
an attempt to bridge the gap betw.een these two areas of endeavor. Ii[-
rially qualitative information both on cavitation anJ on the develop-
ment of vortices and/or turbulence along surfaces of separation w~as
sought by direct photographic observation. Even the initial results were
* . - .found to be very revealing. To make them immadiately available to
othier investigators as well as to engineers in practice, the filmns have
been prepared for presentationi.
Contributed by the Hydraulic Division tor presentation at1 the Winter AnhudI Meel-
ing, New York. N. Y., November 27-December 2, 1960, of The American Society of
Mechanical Engineers. Manuicript received at ASMdE Headqsurters, August 30, 2960
Writter, discussion an !his paper wilt be accepted up to January to, 1961.
S ~ ~ ~Copies witl be available until October 2. 2962. -
- 7Y
- t - i~MIT
f
Fig. 1 Water tunnel used in experiment
EXPLORATORY EXPERIMENT In Figs . and 4 enlargements of individual
frames of the film are reproduced, the first
An existing small recirculating water tunnel showing the first 9 in. of the flow starting at
in the Fluid Mechanics Laboratory of the Univer- the abrupt expansion (visible in the upper left-
sity of Kansas was adapted for a first close look hand corner of the picture) and the second ex-
at flow along a surface of separation. The main tending from 3 to 12 in. from the expansion.
ducts of the tunnel, Fig.l, are 10 in. diam with These reveal a certain regularity in cavity for-
vaned elbows. Flow produced by a Peerless axial- mation which must be associated with reduced
flow hydrofoil pump passes through a honeycomb to pressure within individual vortices. Growth of
insure against rotation of fluid in the tunnel the vortices is evident. The small filament of
and thence through a well-streamlined contraction bubbles linking adjacent main cavities occurs due
into a rectangular test section 3 in. wide by 6 to the boundary layers on the side walls. The
in. high and 4b In. long made of clear acrylic process can be described in this way. Initially,
lastic. Further contraction was provided by an an essentially two-dimensional vortex is gen-
insert in the test section forming a nozzle 5 in. erated near the beginning of the surface of sep-
square followed by an abrupt expansion to 6 in., aration. The ends of the vortex lie in the
as shown in Fig.2. Thus, a surface of separation boundary layers on the side walls and consequent-
was formed along a horizontal central plane in ly do not have as large a translational velocity
the test section. With a width of section of on- as the main part of the vortex. As the ends of
ly ; in., absolutely two-dimensional flow was not the vortex begin to trail behind, their motion
assured, but the effects of side-wall boundary becomes influenced by flow in the boundary lay-
layers were easily distinguishable. In these in- ers, which itself follows the counter-clockwise
itial qualitative observations, no adverse ef- rotation of fluid in each main vortex. The
fects were encountered because of the side wails. traling vortices thus are defiected downward by
ML. '- of' the flow along the surface of sop- circulation in one main vortex and upward by the
aration were taken with a 16-mm Cine-Kodak Spe- following vortex. In this process, the ends of
,lal camera provided wth a contactor to trigger the vortices are stretched which results In their
a microflash lamp each time the shutter opened. becoming smaller in diameter and faster in rota-
The -amera was mounted at tri,_ sam level as the tional speed As a result, cavitation initially
,'I'r l.n, c ti,, t. st ;,.:cti on ankd in front so occurs within these trallink vortices, as shown
*.hat :'low i3 frm l,:t to right in the photo- In Fic.:,.
e'raphs which follow. The flasn lamp, was 1,laced
directly b#,low the test section so that refle'- COMPARISON WITH JET CAVITATION
iouns wrt obs,.rve,.d from, It',5 and tas nucleI
'r th, f'low. A :'1I s, I d of fram,-s jer c S of avltatlon in the mixint. zone of
13.; , 'or il !nI tial I ! t ,I . s imerk,_,d circular ,. ts at he Iowa Institute
-- ==
Fig. 2 Dimensions of abrupt expansion
(1,2)1 revealed very little if any regularity in
the formation of vortices along the surface of
separation. From the beginning, the fluctuations
in flow appeared to take on a randomness charac- Fig. 3 Cavitation at abrupt expansion. X = 0 to
teristic of turbulence. In contrast, vortices X = 9 in.; K = 0.4; Uo = 34 FPS
developed with regularity in the bounded expan-
sion. This raises the question as to whether the
eddy between the zone of separation and the walls
of the expansion in some way controls the forma- When separation occurs, it is no longer nec-
tion of vortices in a periodic fashion. The only essary that the mean pressure in the flow ap-
other evidence found to indicate that this might proach the vapor pressure of the fluid. Within
be the case is a photograph of cavitation around transient vortices, the local pressure may be
a circular disk taken at Caltech (3). In Fig.l0 lower than the mean pressure by 50 per cent or
of this reference, two successive areas of cavi- more of the dynamic pressure of the flow, pVo
2
/2.
tation are revealed, though only one is labeled. The extent to which vortices locally reduce pres-
Although no definite conclusion can be drawn from sure in the flow, the relationship between vor-
this limited evidence, it is apparent that regu- ticity and minimum pressure, and the spatial de-
lar vortices are created along surfaces of sep- velopment of vortices along surfaces of separa-
aration under some conditions and perhaps these tion are all important characteristics of this
are associated with flows having distinct and phenomenon which need to be investigated. Once
limited areas of separation. Additional informa- this information has been obtained, scaling re-
tion on this subject is needed to better under- lationships for this type of cavitation can be
stand the development of flow in the diverse sit- established.
uations where separation occurs.
DEVELOPMENT OF SCALING PARAMETERS
PRACTICAL IMPORTANCE OF SEPARATION-CAVITATION
The factors affecting onset of cavitation
In general the value of the cavitation param- along surfaces of separation are essentially the
eter same as those for boundary layers as described
by Daily and Johnson (4). Their studies indicate
K = (po.Pv)/(PVo
2
/2) that the incipient cavitation index may be ex-
pressed by a relationship of the form
for incipient cavitation is higher where separa- p*
tion exists than it is for streamlined flows. K
i
= Cp + K
t
= PVo2/2
Thus, there is always a possibility that where
cavitation is unexpectedly encountered, separa- To make the relationship applicable to flow along
tion also exists even though it may not have been a surface of separation, slightly different def-
expected. One example of this was recently en- initions of the first two terms on the right are
countered where noisy operation of a pump arose introduced here, as follows:
due to cavitation. In this case, maximum noise
occurred at shutoff when the static pressure in C = Po " P , pressure coefficient
the pump was a maximum and net flow was zero. P PV2 /2
In this case separation, apparently in the scroll
case, was the cause of cavitation. po = reference pressure
1 Numbers in parentheses designate References p = average pressure at point of inception
at the end of the paper. of cavitation
Fig. 4 Cavitation at abrupt expansion. X 3 in. Fig. 5 Incipient cavitation at abrupt expansion show-
to X = 12 in.; K = 0. 4; Uo = 34 FPS ing cavities in trailing vortices
- for it separates the effects dependent upon vis-
K
t
p-o-2-, a pressure-difference param- cous and surface forces.
pVo/2 eter representing reduction
of pressure in vortices along AckowlcdJ/eflt
surface of separation
The observations reported here were made as
part of a basic study of flow through abrapt ex-
p* critical-pressure parameter indicat- pansions sponsored by the Kimberly-Clark Corpora-
PVo
2
1
2 1
Ing pressure at which gas nuclei tion and carried out in the Fluid Mechanics Lab-
will become unstable and tend to in- oratory of the University of Kansas. The assist-
crease in size without limit ance of Charles L. Sanford with the experiments
and Harry Wright of the University Photographic
Daily and Johnson have shown that in flows Department with the moving pictures is gratefully
at high velocity (30 fps or more) with macro- acknowledged.
scopic nuclei present (.001 in. or more), the ef-
Pe fereric e
feet of the last parameter is negligible. Fur- 1 H. Rouse, Cavitation in the Mixing Zone
thermore, data on the pressure coefficient Cp of a Submerged Jet," La Kou14e Blanche, January
are already available for abrupt expansions (5),
so that observations of the inception of cavita- 1 p. 9.
2 D. W. Appel, "An Experimental Study of the
tlon from nuclei will permit calculation of Kt. Cavitation of Submerged Jets," Report to Office
The magnitude of this factor will certainly de-
pond upon the Reynolds number of the flow and to oNv Research frne Iow I te Hr
some extent on the geometry of the approach to i Rac June 5, 1.
5 R.W. Kermeen, J.T. McGraw and B.R. Parkin,
the abrupt expansion. However, there is a pros- "Mechanism of Cavitation Inception and Related
pect of correlating Kt with Reynolds number and
thereby completing the last step needed to formu- Sa-fc Pbm Tas AEvo '1
late a scaling law for separation-cavitation. 195 , PP. D l nV-5.I

J h o , T r l n
Certainly the law would have to be confirmed for 4 J.W. Daily and V.E. Johnson, 'Turbulence
flows in which the parameter representing the ef- and Boundary-Layer Effects on Cavitation Incep-
flows in wfhich the o parametrthtion
from Gas Nuclei," Trans. ASME, vol. 78,
November 1956, pp. 1695-1706.
P* 5 N.K.D. Sharma, "A Study of Flow Through
(pV
0
2
/2) Abrupt Two-Dimensional Expansions," Report No.2,
Studies in Engineering Mechanics, Center for Re-
is not negligible. However, this approach, de- search in Engineering Science, University of*
veloped by Daily and Johnson, holds real promise Kansas, February 196u.
AvaI y I Of
Dist SpCI
46
Compton Press Inc.,
Morristown, N4. J.