A
MITSUBISH I TECHNICALBULLETIN MTB 010045
MITsuffisHI TECHNICAL BULLETIN No. 45
March 1967
Investigation into the Propeller
Cavitation in Oblique Flow
Investigation into the Propeller Cavitation in Oblique Flow
Introduction
There have been carried out many cavitation tests in order to provide propeller design data and to serve performance
prediction of high-speed vessels. It should be noted, however. that most of the propellers of high-speed vessels are operated
in oblique flow conditions. In the case of non-cavitating propellers, the effect of oblique flow on the performance of propeller has been studied by several investigators treating the problem theoretica1ly' and experimentally23, while in the case of cavitating propellers there are at present no published data or available informations on this effect. It is
highly desirable, therefore, to carry out propeller cavitation tests in oblique flow condition and to study the effect of
oblique flow on the performance of propeller, from which we
should be able to get a more reliable model-ship correlation of high-speed vessels and to predict better the cavitation pattern relating to the cavitation erosion on the propeller blades.
In 1963 Mitsubishi Experimental Tank (Nagasaki) started
the study on the propeller cavitation in oblique flow condition
under the sponsorship of the Office of Naval Research'3. The principal part of the study is to be described in this
report.
Firstly an attempt was made to develop a method to
calculate the performance of propeller in oblique flowcondi-tion from the propeller characteristics in axial flow condicondi-tion by the use of quasi-steady approach.
Then, in order to confirm the validity of the method of calculation, a series of supercavitating propellers including the variation of pitch ratio and expanded area ratio were tested in the cavitation tunnel in oblique flow condition. The test results were found to be in fairly good agreement with the propeller characteristics estimated by the quasi-steady method mentioned above.
Trial data of some high-speed boats were analyzed based on these investigations. The model-ship correlations seem to be improved to some extent by the use of the propeller
characteristics in oblique flow conditions.
Quasi-steady Calculation of Propeller
Charac-teristics
The most rigorous theoretical approach to the effect of oblique flow on the propeller would be the application of a
Dr. Eng., Manager. Nagasaki Technical Institute, Technical Headcicarters Experimental Tank, Nagasaki Technical Institute. Technical lleadquarters
Kaname Taniguchi5
Hidetake Tanibayashi55
Noritane Chiba**
Abstract
In order to serve the improvement of performance prediction of high-speed vessels, theoretical and experimental investigations into the propeller cavitation in oblique flow condition have been carried out.
As a first step, a simple quasi-steady method of calculation was developed to estimate the propeller cha racter-istics in oblique flow condition from those in axial flow condition.
Titen a series of supercavitating propellers were tested in cavitating oblique flow condition to confirm the
validity of the quasi-steady method of calculation. The test results were found to be in fairly good agreement with the propeller characteristics estimated by the quasi-steady method.
non-stationary lifting surface theory of propeller. This
method, however, requires a complicated calculation,
neces-sitating the use of a big electronic computer. Furthermore the non-linearity of propeller characteristics K, K J curves
due to the cavitation increase the difficulty in the formulation
and the numerical calculation of the problem. For practical application, therefore, we might as well attempt to develop a simpler method without so much sacrifice of the accuracy of
the calculation. To this purpose a quasi-steady method was tried as the most suitable one for the first step of this study. In the case of cavitating propeller operating in non-uniform velocity field, the results of quasi-steady method have been compared with those of non-stationary method, showing good agreement of the time-average thrust and
torque'.
The fluctuating terms, on the other hand, are affected by the frequency of oscillation of flow field around the propeller. Taking for example the reduced frequency at 0.7R, which may be taken as a representative factor for the non-stationary terms(eC07 (dR)07
k07
2V (1)k77 is about 0.5 for a propeller with narrow blades and is
about 1.0 for a propeller with wide blades. Certain difference
between a quasi-steady and an unsteady method, therefore, might have to be expected in the fluctuating terms.
In the present study, however, the prediction of the time-average thrust and torque were more accentuated than the fluctuating forces and the quasi-steady method was tried. (As mentioned later in 4.2, it is to be noted that the compari-son of the cavitation pattern between a quasi-steady and an unsteady approach showed pretty good agreement:
The quasi-steady method is based on the assumption that the elemental thrust and torque of a blade element at a certain angular position can be equated with those of axial flow condition, at the same resultant inflow velocity. The
quasi-steady advance ratio of a blade element, at the radial
distance xR and the angular position 'p (Fig. 1), can be represented as
cos O
J=xtanß=J0
. . (2)1+(J0Jirx)sin çc.sln O
and the local circumferential velocity is (neglecting the
induced velocities),
UsinO sine
Fig. i
Velocity diagramwhere u0= rnD (4)
Jo= v/nD (5)
and suffix o refers to the axial flow condition
Therefore, the elemental thrust and torque developed by this blade element are
dT'/z= (pn2D4/z) . (u/u)2 .K0(J, c5)0(x)dx (6)
dQ/z = (pn2D5/z) . (u/u)2 .KQ0(J, a0) ?Jf(x)dx (7)
where 0(x) and /J(x) are the thrust and torque distribution
function respectively.
From (3)-...(5), (u/u0)2is
(u/uo)2 = 1 +2(J0/x) sin ÇD.siflO + (Jt/1rx)2Sifl2ç.sin2 ¿I 8)
and
o= o,(a/u)2
19) Hence, using the equations (6)-.(9), we can calculate the thrust and torque developed by a blade (at the angular posi-tion ç) of a propeller operating in the oblique flow from the characteristics in axial flow condition, if we know the distri-bution functions 0(x), and V(x).For simplicity, however, if we take the 0.7R section as
repre-sentative, the thrust and torque of one blade are given by
the following equations.
T'/z= (pfl2Dl/Z).(ze/uo)2.Krn(J o) (10)
Q/z=(pn2D°/z).(u/u0)2.K0o(J, a0) (11)
K0 and K00 corresponding to [J10, ço). a0(O, so)] can be read for each angular position of the blade from the charac-teristics in the axial flow condition. The circumferential
nO 2 uO ) 2
mean values of KTOk ' and Keo1 ' will be
u0 u0
taken as K' in the direction of propeller shaft) and K0 in
oblique flow condition respectively, viz.
KT' f2KTOJ. a5).(u/u0)2.dçt (12)
(13)
Thrust T in the direction of general flow is obtained by assuming the side force is equal to sin p,
K = K' cos O - 2Lf K00(J, a) sin O. SIflÇD(u/u)'dç
(14)
2
Fig. 2 Propeller P. 1329
Table i Propeller particulars
3. Test Scheine
3. 1 Propeller modelsFive propeller models with Tulin's supercavitating blade section were tested in the present study. As described schematically in the diagram below, the particulars of the
propellers were varied systematically on the basis of P. 1329, which is a scale-model of the propellers of a high-speed boat.
p 1.286 Ae/Ad- 0.514 P. 1372 D=230 mm p= 1.286 Ae/Ad= 0.411 E.A.R. Series D=23O.Orn P = 2 05 .7 P=1 .286
P. 1369, P. 1329 and P. 1370 constitute a series of pitch
ratio, while P. 1329, P. 1371 and P. 1372 constitute a series of
expanded area ratio.
The particulars of the propeller models are presented in Table 1, and the drawing of P. 1329 is shown in Fig. 2.
3.2 Kinds of test (a) Cavitation tests
Cavitation tests in oblique flow condition were carried out with a special attachment as shown in Fig. 3. A propeller 1329 1369 1370 1371 1372 Diameter (mm) 230.00 230.00 230.90 230.00 230.00 Pitch (0.7 R) (mm) 295.71 230.00 368.00 295.71 295.71 Pitch ratio (0.7 R) 1.2857 1.0000 i 1.6000 1.2857 1.2857 Disc area (m2) 0.04155 0.04155 0.04155 0.04155 0.04155 Expanded area (m2) 0.02572 0.02572 0.02572 0.02136 0.01708
Expanded area / disc area 0.6190 0.6190 0.6190 0.5141 0.4110
Bossratio 0.1819 0.1819 0.1819 0.1819 0.1819
Thick-chord ratio at
0.7 R (%) 6.118 6.118 6.118 6.516 7.088
a e sec i°fl
BI d t S.C.typeS.C.typeS.C.typeS.C.typeS.C.type
Number of blades 3 3 3 3 3 P. 1369 P. 1329 P. 1370 D=230 mm ,=1.0Oo Ae/.4d= 0.619 D=-230mm p.= 1.286 Ae/Ad= 0.619 D=230 mm p=1.600 Ae/Ad==0.619 PitchSeries P. 1371 D=230 mm General flow Direction of rotation u
'I
IHandle for changing shaft inclination
Universal joints
Observation window
shaft ç900mm long and 38mm in diameter) was connected to the ordinary propeller shaft by means of a couple of universal joints, and the downstream end of the shaft was supported by a vertical strut, which enabled the variation of
the inclination of the shaft.
The test conditions were as follows; Inclination of the shaft: O
O=0, 4° and 8 Advance ratio: J=v/nD
J=0.7l.l
for P. 1329, P. 1371 and P. 13721=0.5-0.9 for P. 1369 J=0.9-1.5 for P. 1370
each covering the range 10-45% in slip ratio. Cavitation number: r
a,,=0.3-1.0 and atmospheric condition
for P. 1329, P. 1369, P. 1371 and P. 1372
011=0.5-1.5 and atmospheric condition
for P. 1370
Thrust and torque were measured by the
propellerdynamometer for ordinary Cavitation tests so that the
measured thrust was in the direction of general flow.
(b) Open.water tests
In order to obtain a reasonable method of Correction for the tunnel wall effect and for the rotational wake of the rotating propeller shaft upstream of propeller, open.water tests in oblique flow condition were carried out on all the models. A special propeller dynamometer of a strain gauge type was developed and it was so arranged that the propeller shaft could be inclined in the vertical plane. It should be borne in mind that the thrust measured by this dynamometer is in the direction of the propeller shaft, while in the cavita tion tunnel it is in the direction of general flow.
4. Test Results
4. 1 Correction to the measured results in tunnel
The thrust T and torque Q measured in tunnel were
Comparison of the test results in tunnel in non.cavitating conditions with those in open-water showed that it was not
necessary to correct the results for tunnel wall effect in terms
of the equivalent inflow velocitythe water speed was
measured by a pitot tube as shown in Fig. 4, but it was
Attachment shaft
Fig. 3 Sketch of special attachment shaft
Static pressure hole Direcflonof flow
Tunnel wall
Observation window
Total pressure tube
Static pressure hole
Total pressure tube
Fig. 4 Pitot tube
observed that the rotational wake of the rotating propeller shaft upstream of propeller had a significant effect on the propeller characteristics5. The correction for the rotational wake was applied on the basis of the assumption that KQ J curves in tunnel under atmospheric condition and in open. water should agree with each other at the same Reynolds
number. In other words the correction factor in/n for the
rotational speed was obtained in such a way that KQJ
curve in tunnel should coincide with KQ*_J* curve (after the correction for Reynolds number) satisfying the relations
.
(18)
1* =iv(i+)
(19)Using this correction factor, we obtain the corrected propeller
characteiistics as (20) Miß 010045 MARCH 1967 o o Observation window 3
reduced to the non-dimensional coefficients K1° and K0°,
KrT/pn2Dl
(15)KQ0=QJpnODS (16)
J0v/nD
(17)4 0.20 Q 0.15 0.10 0.05 0 0.15 U1 Q o u-b 0.10 0.05 O 0.25 e 0.20 0.25 0.15 0.20 0.10 0.15 0.05 0.10 0.15 o 0.10 0.05 . O 0 0.05 0.15 Q u-b 0.10 0.05 07 8 4 8 Marks 08 P. 1329
Cavitation test results ir, oblique flow
09
J-1 'ID
Fig. 5 P. 1329 Kr-I curves 10
Kr J
11 12
by the variation of local cavitation number and they are nearly linear with respect to advance coefficient J. In such
a case, the quasi-steady calculation can be simplified and we
obtain, as shown in Appendix,
KT' = KTO+ [--- 70J+ - a,0 (J/O.Th)2O2
KQ = K55+ [--- JP-J+--ano (J/O.71v)2]02 (26)
or referring to (23) and (24)
1 dK0
1Aeai=
--B0i --
SKQ0J1U/07)2
(28)where a,0 and a5, show KTO and K50, respectively, as extra-polated linearly to J=O.
Comparison of the coefficients A and B as obtained by experiments and theory will be a good measure for checking the reliability of the quasi-steady calculation. In the table below, the comparison is given in terms of ratio a and b
(25)
(27)
KQ=KQ0(1+ 21)
=10(1+ An) (22)
The correction factor An/n obtained in non-cavitating con-dition was applied also for cavitating concon-ditions, because significant variation of the rotational wake is not expected
by the change of the pressure in tunnel.
Thus corrected K and K5 were plotted to the base of the corrected J. An example of the corrected results is given in Figs. 5 & 6.
4.2 Comparison with a quasi-steady calculation The variation of KT' and K5 due to the effect of oblique flow can be expressed by an even function of the shaft
inclination O. For a small angle of shaft inclination, Ky.' and K5 are approximated by
KT'KTo+AO2 (23)
KQ*KQO+BO2 (24)
0.04 N. o
Ub
0.03 0.02 0.01 O 0.04 o o u. b 0.03 0.02 0.01 O 0.05 e o Ub 0.04 0.03 0.02 0.01 0.03 0.02 0.01 0.03 o o 0.02 0.01 07 08 8 a=Aexp/Aci (29)bBexp/Bi
(30) where Aexp=(KT'K70)/02 Bexj, (KQ_KQO)/02A5 and
were obtained from open-water test results.It is to be noted that a and b are generally larger than unity and b increases with pitch ratio. If we compare KQ itself, however, instead of its increment due to the inclination of
OQ
O.4
P. 1329
Cavitation test results in oblique flow
K0 -J L..
OQ3
j = V/nD Fig. 6 P. 1329 KQ-J curves 0 9 1.0 1 1 1 2 MIB 010045 MARCH 1967the shaft, the differences between the measured values and those calculated by the quasi-steady method are about 2 except for P. 1370 with the highest pitch ratio (p= 1.6), for which the difference amounts to 3-5%. Further study both on theory and on experiments will be necessary to make clear this trend.
In cavitating condition, the estimation of the propeller characteristics in oblique-flow from those in axial flow con-dition is much more difficult and complicated than in non-cavitating condition. The instantaneous K- and K53 in eqs.
ib) and (111 respectively, at each angular position of a blade are functions of local cavitation number as well as local
advance coefficient. Besides, K0 and K50 versus J curves can no longer be approximated by a linear relation due to thrust and torque breakdown caused by cavitation on the 5
P. No. p AelAd a(6=8°) 6(8=8°)
P. 1329 1.286 0.619 3.74 1.92 P. 1369 1.000 0.619 2.07 1.55 P. 1370 1.600 0.619 1.29 2.34 P. 1371 1.286 0.514 3.15 1.97 P.1372 1.286 0.411 2.28 1.92 0.03 o o
ub
0.02 0.01 o Marks o oo Measured Calculated o=0.S P 1329 p 1.2857) or. = 04 P 1369(p=l.O
\
=0.5 K0 K0 P 1370 (p=I.6) 0.5 0.6 0.7 0.8 0 9 1,0 1.1 1.2 J V ltDFig. 7 Comparison between the measured values and those calculated by quasi-steady method O = 8°
blades. At present, therefore, the quasi-steady calculation as stated before in 2, will be the most useful, and may be the only available method to estimate the effect of oblique flow, although the method yields slight discrepancy from the experimental results even in the simplest condition viz, in
non-cavitating condition. In Figs. 7 & 8, the calculated KT corrected to the direction of general flow) and K6 are coni-pared with those measured at O=81. The measured K9's agree quite well with those calculated by the quasi-steady method except for P. 1370, for which the measured torque is slightly larger than that calculated. As for KT, the measured values are slightly smaller than those calculated. In general,
however, it may he said that the agreement is pretty good and is within the accuracy of measurement. The remaining discrepancies, though considerably small, may be due to the approximate analysis by means of the quasi-steady method. Better agreement will be achieved by the refinement of the measuring technique as well as the development of a more
rigorous theory.
For a propeller blade in oblique flow condition, local advance
ratio and cavitation number change with its angular position. It is interesting to compare the cavitation pattern on the blade in oblique flow condition with that in axial flow at the corresponding J and o. Fig. 9 shows a comparison of the cavitation patterns between oblique flow and axial flow
con-ditions. The sketch in the middle of this figure represents the cavitation pattern in oblique flow conditions (O=8.
çD=900). The local advance ratio J(O. çs) and the local
cavi-tation number o(O, ço) calculated by (2) and (9) are, J=0.93 and a,=O.44
Comparing this cavitation pattern with the sketches in axial flow )at the corners of this flgure, vie may say that the propeller blade operates in nearly the same condition as is
predicted by the quasi-steady calculation.
0.8 0.9 1.0 0.8 0.9 1.0 0.8 0.9 1.0
jV eD
Fig. 8 Comparison between the measured values and those calculated by quasi-steady method O = 8°
Fig. 9 Cavitation pattern in oblique flow
4. 3 Effect of pitch ratio and expanded area ratio In non-cavitating condition the effect of oblique flow can be expressed, as mentioned above, by the increase of thrust and torque which is in proportion to the square of the angle
P 1371 :L' li 0= Q, 0.514 K0 Measured Calculated A P 1372 ,4a-0 411 0.5 0.04 0.03 0.02 0.2 0j P 1329 A/Aa=O,616 S' c-. 0.5 0.04 0.03 0.02 0.2 0.1 o =0.6 K1 P 1370
of shaft inclination. From (25), (27) and (29), we obtain,
Kr(J,0)Kro_--02[aIj0J ato(J/0.7r)21
..
31)and similarly from (26), (28) and (30)
KQ(J, 0)=Ko_02[
aJJ
aqo(J/0.72r)2] (32)Since a propeller with large pitch ratio operates in general at large advance ratio and the factor b increases with pitch ratio as mentioned before, the increments of K7 and KQ
increase with pitch ratio.
As for expanded area ratio, there is no significant variation
3K70
of the factors a and b. and -a-,- increase slightly with expanded area ratio. In non-cavitation, therefore, the effect of oblique flow varies little with expanded area ratio.
In cavitating condition, such a simplified analysis is not suitable for the discussion on the effect of pitch ratio and expanded area ratio.
In general the increments of KQ due to shaft inclination decreases with the decrease of cavitation number, and for the range of e<O.5 and s>0.3 the design point of super-cavitating propellers usually lies in this range no appreciable variation in K is found with the angle of shaft inclination
except for P. 1370. For P. 1370, which has the highest pitch
ratio p= 1.6) among the propellers tested, K increases still at aj=O.5 with the angle of shaft inclination.
On the other hand, K)J. (measured in the direction of general
flow) decreases in oblique flow cavitating condition in con-trast with non-cavitating condition. The decrement of K7 increases slightly with pitch ratio, while expanded area ratio does not have a definite influence on the decrease of K7 due to shaft inclination.
5. Conclusions
The results of the above-mentioned theoretical and ex-perimental studies may be summarized as follows:
(a) Quasi-steady calculation can be used to predict the
propeller characteristics in oblique flow in cavitating
con-dition as well as in non-cavitating concon-dition.
)b) Observation of the photographs of the cavitation
Simplification of the quasi-steady calculation in non-cavitatifig condition
In non-cavitating condition K7 and K are not affected by cavitation number and are expressed approximately by a linear function of advance ratio, viz.
K7=a,+h7J (A.1)
KQ=aQ±b4J (A.2)
As mentioned in section 2, the quasi-steady K is obtained
by u(f1,ç) 2dço (13) no J aq0+b50J(0,ça) u(0, ç,)
Substituting (2) and (8) for J(0, ç) and
respective-I u I
ly, we obtain
KQ(JQ, O)=a00+b00.J0.cos
a0
(J0/irx)' sifl2 Q --. (A. 3)where x=0.7.
Expanding (A. 3) into the series of 0, we obtain KQ(JO, 0) =a00+ hqoJo+ [_ hq0J0+ J0/0.7 7)2]02
patterns on the blades suggests that the propeller blades in oblique flow condition operate in nearly the same manner as is estimated by means of a quasi-steady calculation from the behaviour in axial flow condition.
(c) In non-cavitating condition, the increments of K7 and KQ due to shaft inclination increase with pitch ratio, but little with expanded area ratio. In cavitating condition,
the effect of shaft inclination is not in such a simple relation with the geometry of a propeller. In general it was observed that the increment of KQ decreases with the decrease of cavitation number and that K7 decreases
in oblique flow in contrast with non-cavitating condition.
As an application of these investigations, the trial data of some high-speed crafts were analyzed based on the propeller characteristics which were estimated for the oblique flow conditions using the present method. The model-ship cor-relations seem to be improved to some extent by such
treat-ment.
6. Acknowledgements
This investigation has been carried out under the sponsor-ship of the Office of Naval Research, Department of the
Navy, Contract No. Nonr 5002)00).
The authors wish to express their gratitude to all the members of Mitsubishi Experimental Tank (Nagasaki) who cooperated in carrying out this investigation.
References
R. Yamazaki: On the Theory of Screw Propellers in Non-Uniform
Flows, Memoirs of the Faculty of Engineering, Kyushu University, XXV, 2, (1966)
K. Taniguchi and K. Watanabe: An Experimental Study on Propeller Characteristics in Oblique Flow, Journal of Seibu Zosen
Kyokai, 8, (Aug. 1951)
F. Gutsche: Untersuchung von Schiffschrauben in schräger
Anströmung, Schiffbauforschung, 3, (3/4/1961
K. Taniguchi and N. Chiba: Investigation into the Propeller Cavitation in Oblique Flow (ist report), Mitsubishi Experimental
Tank, Report 1800, May 1964)
K. Taniguchi, H. Tanibayashi and N. Chiba: Investigation into
the Propeller Cavitation in Oblique Flow 2nd report Mitsu-bishi Experimental Tank, Rep3rt 2221, (May 1966)
Appendix
MIB 010045 MARCH 1967
Referring to (A. 2), we can write
KQ(J0, 0)=KQO+[--- 1°j0+ OO(J/O7 r)2]02 -.. (A.4)
Similar discussion holds also for K7', and
K7'(J0, 0) = K70+ [- - aK.0+ ---U0/0.7 7)2102 -- - (A. 5)
List of Symbols
Ag/Ad Expanded area ratio
A, B Coefficient of increment of thrust coefficient and
torque coefficient
b= B051,/Brai
C07 Chord length at 0.7 R
D Diameter of propeller dT, dQ Thrust and torque element
J
Advance coefficientk07 Reduced frequency at 0.7 R
K7 Thrust coefficient in the direction of general flow K7' Thrust coefficient in the direction of propeller axis
KQ Torque coefficient
n Number of revolution of propeller
8
p Density of fluid
pse
Cavitation number based on the rotationalp(nD)2
speed of propeller
Angular position of blade (QO at the top)
Thrust distribution function Torque distribution function
w Frequency of oscillation of the velocity relative to
propeller
u (Subscript refers to the quantities in axial flow condition. o (Superscript refers
to the quantities as measured in
cavitation tunnel in oblique flow condition.
* refers to the open-water propeller characteristics.
Reprntin or reproduction without written permission prohibited.
We would appreciate receiving technicol literature published by you.
An/n Correction factor for the rotational wake of the
upstream shaft
p Pitch ratio of propeller
Q Torque
R Radius of propeller
T Thrust in the direction of general flow T, Thrust in the direction of propeller axis
u Circumferential velocity relative to propeller blade
without induced velocity
V Resultant inflow velocity to propeller V Advance speed of propeller
X Non-dimensional radius of propeller z Number of blades
Hydrodynamic pitch angle without the correction
for the induced velocity