Some remarks on vibration problems occuring in the design of propellers for sea-going single screw ships

SOME REMARKS ON VIBRATION PROBLEMS

OCCURRING IN THE DESIGN OF

PROPELLERS FOR SEA-GOING SINGLE-SCREW SHIPS

by

J. A. van Aken

Lips Propeller Works, Drunen, Holland

Introduction.

The following article does not claim to be of a strictly scientific nature, but only touches on a few data which have to be thoroughly

consi-dered in designing marine propellers.

The path trodden by a designer of ship

pro-pellers is strewn with so many pitfalls that the problems which are being broached in this ar

tide only constitute a minor part.

The first section of the article deals with the

number of propeller blades chosen in connection with vibrations set up by the impulses of the revolving blades.

The second part enables the reader to gain an insight into the choice of the minimum clearan-ces which the propeller may have in the screw aperture, which clearances have been recorded in a table.

In the third section the causes and the results

of the singing of propellers are discussed, to-gether with an investigation quite recently made into the manner in which cracks due to this singing were detected in the blades.

These three sections cover the discussion of

all vibration problems that the author has

ex-perienced as occurring with the propeller at the ship's stern.

The problems discussed in the first and second section can only be solved by making a new screw design, whereas, fortunately, the third

problem can be solved in a less expensive way, namely by providing anti-singing edges.

I. Numberof blades chosen in connection with

vibrations set up by the impulses

of the

rotating propeller.

For some years past more attention has been

paid by classification societies, shipowners,

shipbuilding yards and screw manufacturers to

the choice of the number of blades and to the

correct position of the propeller in its aperture than was done before.

By raising the power and the number of re-volutions of the propelling machinery and, con-sequently, the ship's speed, inconvenient

vibra-tions were caused in the after bodies of various

ship types, for instance tankers, which

vibra-tions could not be the direct result of any

syn-chronism of an interferring frequency of the

propelling machinery with the natural frequen-cy of the ship's hull.

It appears, in fact, that in general the

vibra-tions set up by the propelling machinery (in this case an internal-combustion engine is specially

thought of) - either one or two node

vibra-tions - then synchronise with one of the types

of vibrations of the ship's hull (either vertical

or horizontal flectional vibrations with two or more nodal points, or torsional vibrations).

The present tendency of the development of technical science is, however, of such a nature

that these resonances can be determined with

sufficient accuracy beforehand.

The vibrations, however, which may be set up by a rotating propeller are due to an entirely different cause. During each revolution of a

four-bladed propeller of a single-screw ship, a

blade completes a cycle which may be described

as follows. If we start from an upright propeller

blade, that is a blade in top position (blade no. 1

in Fig. 1) the tip of this blade when moving in

right-handed direction, first passes the top of the screw aperture and after having revolved

180 degrees, the bottom of the aperture, while the leading edge of the blade section twice

pas-ses the stern post, and the trailing edge of the

blade section twice passes the leading edge of

the rudder or the back of the stern frame. Since in the case of a four-bladed propeller

the tip of blade No. 3 passes along the bottom

of the screw aperture at the moment when

blade No. 1 passes the top of the aperture, four impulses of the four propeller blades are to be reckoned with at each revolution of the

propel-1er, so that the base frequency is equal to the

product of the number of blades and the num-ber of revolutions of the propeller per minute. Multiples of this base frequency are

general-ly so high that they are of little importance in

Table 1. Frequency Diagram for the Full-load Condition

Table 2. Frequency Diagram for the Ballast Condition

4

I I Amplitudes stated with a three-bladed propeller.

/

LRevolutions

SHIP's HULL

of the

Propeller per Minute

/ PROPELLING

Propeller SYSTEM Revolutions of the per Minute

/

3- 4- 5-

/

3- 4-

5-Bladed Bladed Bladed Bladed Bladed Bladed

Number of impulses

per revolution

--

3 2 4 5 _...-

-Order 3rd 2nd 4th 5th

Node -/min Node

i

-

_i

407-5 135-8 203-7 1019 8F5 2 96.8 96-8 2 7305 2435 3653 1826 1461 3 206 68-7 1030 515 41-2 4 293 97-6 146-5 733 29-3

/

\ /

_{/' \}

/

SHIP's HULL Revolutions of the Propeller per minute

-/

,,/'N..\

/

PROPELLING Propeller SYSTEM Revolutions of the per Minute

-Bladed Bladed Bladed Bladed Bladed Bladed

Number of impulses

3 2 4 5 Order - _{3rd 2nd 4th 5th}

per revolution

-Node -/min Node

-/min

1

i

i

407-5 1358 203.7 10F9 815

2 108-5 1085 2 730-5 243-5 3653 182-6 146-1

3 224-5 74-s 1122 561 449

In the case of a four-bladed propeller,

how-ever, care should be taken with regard to the

semi-base frequency still being capable of call-ing forth transverse vibrations in the ship's hull.

In the case of three and five bladed propel-1ers, the base frequencies are three and five

times the number of revolutions of the propel-ler per minute respectively.

The leaving free vortex of the screw tip on the one hand causes the impulses referred to

above, while on the other hand the passage of the leading and trailing edges of the propeller

blade along the front and back of the stern

frame respectively - the inequality of the

ve-locity field existing there being of considerable

importance - sets up the impulses.

Besides a satisfactory clearance of the pro-peller in its aperture, the avoidance of

exces-sive peaks, both radial and circumferential, in this velocity field is, therefore, of great

impor-tance.

Tables Nos. 1 and 2 show examples of the

interrelation between the natural frequencies of

the torsional vibrations of a propelling instal-lation, the transverse vibrations in the ship's

Fig. 1.

Situation of the propeller in the aperture.

Top of Aperture Front of stern frame 07 R Centre line of shaft Bottomof Aperture

gard to a single-screw tanker in full-load and

ballast condition respectively.

In the full-load condition the maximum trans-verse amplitudes which were measured in the wheelhouse, were 0.71, 1.15 and 1.36 mm, and in the ballast condition 0.43, 1.21 and 0.9 mm for the 2nd, the 3rd and the 4th node vibrations

respectively.

Though these amplitudes, which, therefore,

were merely due to the impulses of the

three-bladed propeller, are of little importance in

respect of the forces to which the ship's hull is subjected, they should be avoided in the

work-ing condition, varywork-ing herefrom about 100 tc

110 revolutions of the propeller per minute.

According to the tables Nos.

i and 2 the

employment of a five-bladed propeller would certainly be recommended here.

II. Choice of the Minimum Clearances of the Propeller in its Aperture.

a) As far back as 1942 Prof. van Lammeren, in his book, «Weerstand en Voortstuwing van

Schepen», pp. 112 and 255, gave a guide for

diagrammatically (see also Fig. 1). The base of this diagram indicates the ship's length in feet. Van Lammeren gives the clearances as a

func-tion of the ship's length, assuming that the

boundary-layer thickness increases with this length.

This choice is only correct if it is presumed

that with an increase of the ship's length the engine power, hence also the screw diameter,

increases proportionally.

Besides, no allowance is made for the fineness of form of the after part of the ship, as, with

a fine after body, the boundary layer will leave the hull later than with a full stern, to the effect that, the clearances being the same, vibrations will be more readily set up in the case of a full

stern than with a fine after end.

b) In a circular letter dated May 1953, Lloyd's

Register calls attention to clearances of

propel-1ers in their apertures in the case of single-screw

i C o C 'u 6 4 o 5 01 o u, Ii -c o C C o o e u Fig. 2.

Minimum clearances of the propeller in its aperture according to van Lammeren.

vessels in connection with ship vibrations, and vibrations caused by the propeller in the case of geared turbine installations.

Lloyd's Register, too, gives the provisional minimum clearances for the propeller in its

aperture.

c) In a detailed report No. 6742 on

«Disposi-tion des Hélices dans les Cages d'étambot»,

issued by the «Institut de Recherches de la

Con-struction Navale», this matter is further dis-cussed, in 22 cases the results have been ana-lysed, from which conclusions have been drawn.

while a proposal for minimum clearances is put forward. The report further discusses an article

written by Allan, who also gives values for

minimum clearances.

Table 3 embodies four proposals put forward,

while a fifth proposal reflects the views held by the present author on the question.

The data were derived from a 400-foot-long

ship, the screw diameter being 4,800 mm (15.75 feet).

Of the balanced rudder it was assumed that

l/t = 7, or t/l = 0.143.

The nomenclature, taken from the diagram in Fig. 2 (see also Fig. 1), is as follows:

screw diameter . . . D

screw radius R

distance of trailing edge of blade from

leading edge of rudder or leading face of

rudder post . a

distance of leading edge of blade at 0.7 R

from propeller post b

distance of propeller-blade tip from top of

stern frame . . . c

distance of propeller-blade tip from bottom

of stern frame d

distance of the back of the stern-post boss

from the front of the propeller boss e

maximum thickness of rudder t

breadth of rudder (inclusive of rudder post) I

From Table No. 3 we see that only van Lam-meren gives fairly small values for the dimen-sion «a», because this clearance is based on the

influence of the rudder, which results in an

increased screw efficiency. When this clearance

becomes excessive, the screw efficiency will de-crease considerably.

On the other hand, there is a possibility of

vibrations being caused by a clearance which is

too small, so that a compromise will always

have to be looked for.

for nome nclatu e

;Z

p

--/

//

-

--/

,

/

/

/

/

/

--

/

-

--4 200

feet-Table 3. Minimum Clearances of the Propeller in its Aperture in the Case of Single-Screw Ships

*) _{In using a balanced rudder, if} t/1 _{> 0.11, «a'> 0.08 D, its maximum being 0.15 D.}

**) For any rudder, if t/1 > 0.11, «a» =

From the above remarks the conclusion may be drawn that if in the case of an existing ship, consequently with a screw aperture that is to be

kept unchanged, vibrations are set up in the

ship's hull which are due to insufficiently large clearances of the existing screw in its aperture,

these vibrations may be considerably reduced or even eliminated altogether by increasing the

number of blades and decreasing the screw dia-meter. If, for instance, a five-bladed propeller

is substituted for a four-bladed one, and the

power of the propelling machinery remains the

same, the thrust per blade of the five-bladed propeller will be 20 per cent less than that of

the four-bladed one. The leaving free vortex of the propeller-blade tip will . be considerably

smaller, through which, naturally, the impulses will be reduced as well.

Since the resonance phenomenon will occur with a lower number of revolutions of the pro-peller per minute, this thrust per blade will be

reduced once more, as with the lower number

Df revolutions the propeller will absorb less

power and, consequently, yield less thrust.

0.06 D, its maximum being 0.10 D.

from a smaller optimum screw diameter of a

five-bladed propeller as compared with a

four-bladed one, will have a favourable influence on the clearances at the top and at the bottom

of the screw aperture. With an absolute blade area remaining practically unchanged the

clear-ances in respect of the propeller post and the leading edge of the rudder will be larger for

the five-bladed propeller than for the

four-bladed one. Owing to this, there will, in

conse-quence of the inequality of the velocity field

and, consequently, of the resulting impulses, be

less variation in torque and thrust when the

leading and trailing edges of the blades pass the propeller post and the leading edge of the

rud-der respectively.

In the last few years the author has success-fully dealt with several similar vibration pro-blems with regard to both vessels of the inland

shipping trade and sea-going ships. He may add that in a few special cases concerning ships

en-gaged in the inland trade not only the number of propeller blades was increased, but the type

of blade section was also changed. In these cases Data Van Lammeren for 400-ft. ship Lloyd's Institut Register's de

Circular Letter Recherches

Van Aken's Allan Proposal 0.08D..0.15D*) 0.06 D -0.10 D ) "a" 0.056 D _{here 0.104 D} 0.06 D - 0.10 D 0.08 D - 0.15 D here 0.078 D "b" 0.134 D 0.15 D 0.15 D - 0.17 D 0.20 D 0.15 D "C" 0.082 D 0.08 D 0.07 D 0.08 D - 0.10 D 0.08 D "d" 0.025 D 0.04 D 0.02 D - 0.03 D 0.03 D - 0.04 D

"e" 8"-lO" 8"-lO"

tu

series were replaced by the aerofoil sections

having shock-free entrance calculated with the

aid of the vortex theory.

From experiments recently carried out in the Trondheim model basin it appeared that in the

case of a p:ro.peller having aerofoil sections with shock-free entrance, the percentage of

vari-ations between maximum positive and negative

allowances on the mean torque dropped from 22 % to 13 % as compared with a propeller

fitted with ordinary blade sections.

These results fully agree with those achieved in practice by the author.

III a. Causes of the singing of marine propellers. The inconvenient noise frequently noticed

with revolving marine propellers and generally termed «singing», is essentially a critical

con-dition of vibration, also, in the case of the su-bject on hand, with propeller blades.

If we want to explain the phenomenon of

singing propellers we have to consider the

trail-ing edge of a blade profile. Here the flow leaves

the blade profile and forms the leaving vortex.

On any propeller blade that is not provided with an anti-singing edge, the releasing point of the leaving vortex shifts within certain limits. This releasing point may be on the pressure side

as well as on the suction side of the blade

ele-ment, while the said limits are, in any case,

situated at a very short distance from the trail-ing edge. If the frequency of the displacement

of the releasing point moving up and down

agrees with the natural frequency of vibration of the propeller blade, singing will occur.

Since the natural frequency of vibration of a propeller blade depends on the shape of the blade, the form of the blade profiles, and the curve of the maximum blade thickness of the profiles, while the frequency of the

displace-ment of the releasing point of the vortex is

un-known, it is very difficult to establish before-hand that the phenomenon of singing will occur.

It sometimes happens that a damaged

propel-ler which was not singing before being damaged,

will sing after having been repaired, which, in the authors' opinion, is due to the accurate re-conditioning and smooth finishing of the blade profiles at the trailing edges, which enables the

releasing point of the vortex to shift.

On the other hand, it is not necessary for

every propeller to sing when revolving behind the ship. The best method of preventing singing of the propeller blades is to grind anti-singing edges on the trailing edges of the blades in such

Direction of Turning

D Diameter of the Propeller

m Sucton S,de Omm hmmßmmdmm 7000 95 30 300 6.000 90 29 275 5,000 65 28 275 4.000 75 25 250_ 250 3000 65 22 0 mm h mm B mmd mm 3.000 650 22 250 2.500 575 20 225 2,000 500 lB 200 f500 400 16 175 1000 300 14 l'SO D 3.000 rom D.i.000-3.000rnrn 0<1,000mm

NOTE Th Valul of h e oand for 079-0 BR end O 99

The Negative Tolerance of "h" may Not Eoceed /z mro

Fig. 3.

Different types of anti-singing edges.

a way that the releasing point of the leaving

vortex is fixed.

Fig. 3 shows the different types of anti-sing-ing edges, dependanti-sing-ing on the screw diameter, for

bronze, cast-steel and nodular-iron propellers

manufactured according to the Lips' design.

From this figure it is seen that for propeller

diameters of upwards of 3,000 nun, the trailing edge between 0.5 R and the tip of the blade has

to be cut chisel-shaped on either side over a

breadth of a few centimetres (I), while the

breadth of the anti-singing edge has to be such

that the leaving vortex will be forced to leave

the blade profile somewhat earlier than would be the case if there was no anti-singing edge.

For propeller diameters of upwards of 1,000 and less than 3,000 mm, when the angle a

be-comes too small, it is sufficient to grind the

anti-singing edge only on the pressure side (II). For smaller propellers, of less than 1,000 mm diameter, it is recommended to grind the anti-singing edge as given under (III).

III b. Harmful consequences that may arise

from singing propellers.

Until recently only one case had been estab-lished by the author in which a singing propel-ler showed small cracks along the trailing edges

of the blades. As this propeller had not been

designed by Messrs. Lips, so that no

anti-sing-n I Suction Side

Photo No. i

Crack in propeller-blade 3.

ing edges had been provided, and there was not

sufficient opportunity to investigate the pro-blem more closely, the blades were repaired and provided with anti-singing edges. In this way

the fault was remedied.

Some months afterwards, however, a unique opportunity presented itself for facing the pro-blem and holding a thorough investigation into the cause of the cracks in the blades of a

four-bladed propeller of about 5 metres diameter, which cracks were probably due to singing.

This bronze propeller, which had been cast by

Messrs. Lips from drawings furnished by the

client, had, at the latter's

requ'est, not been provided with anti-singing edges. When the

ship, which was a single-screw vessel, made her maiden voyage, the propeller appeared to be singing.

In the course of time, however, the singing

became less apparent, till at length it stopped. Shortly thereafter, the ship was put in drydock,

where it appeared that all four blades were

showing cracks between 0.95 and 0.85 R,

start-ing from the trailstart-ing edges.

Photo No. 2

Propeller with cracked blades due to singing.

Photo No. 3

Assemblage of the vibration apparatus.

The ship was drydocked about five months

after the propeller had been put into use. The Photos Nos. i and 2 give a clear idea of the development of the cracks.

From blade No. 2 a portion even proved to

have been struck out (see photo No. 2).

According to the strength calculation the

maximum compression stresses in the blade sections at 0.85 and 0.90 R amounted to 490 and 310 kg/sq.cm. respectively.

At the point in the blade section where the crack started - at 0.95 R in the trailing edge

- the maximum compression stress was 115 kg/

sq. cm..

Considering the low stresses at the point

where the crack had originated, M'essrs. Lips conceived the idea that this crack might be due to the original singing of the propeller.

ill c. investigation into the nature of vibration.

It was decided to carry out experiments on

the vibration phenomena in collaboration with the Netherlands Ship Model Basin at

Wagenin-gen.

First one of the blades (blade No. 3) had to

be carefully welded in order to imitate its initial

stage as accurately as possible.

Photo No. 3 shows the assemblage of the vi-bration apparatus, and Fig. 4 a diagram of the switch gear used for this apparatus.

A metal terminal (see Fig. 4) was fitted to the propeller blade (blade No. 3) in way of D (see

Fig. 5).

By means of a pin lying against the bottom

of the terminal (see Fig. 4) the propeller blade

was caused to vibrate with a particular

lo Exciter--. u-

-i.

i

Tuned Amplifier

Pin of the Exçjjr

000000

₀₀₀₀₀

Generator

R t-tiv VI. - u.n P

Propeller blade

o

o 0 o

0000

000

L-F Electronic Oscil ograph

Fig. 4.

Diagram of the switch gear and the L-F electronic

oscillograph.

This was done by making use of a tuned ge-nerator transmitting a definite alternating

cur-rent to an amplifier.

The power of this amplifier was introduced

into the exciter, in consequence of which the

above-mentioned pin was m.ade to vibrate. The vibrations were again absorbed by a re-lative vibration pick-up, which can be moved

along the surface of the propeller blade by

hand.

Via the amplitude-measuring apparatus the

vertical amplitude of these vibrations were

re-produced on the screen of a low-frequency

electronic oscillograph.

By coupling this vertical amplitude to its own time base - by which is meant the horizontal movement on the screen of the oscillograph, representing the period of the vibration - the

well-known sinusoidal curves were produced. When the relative vibration pick-up is moved about the surface of the propeller blade the ver-tical amplitude of the vibration, when intersec-ting a nodal line, must practically be zero- on

the screen of the oscillograph, so that this screen

only shows a horizontal line.

If, on the other hand, a so-called «maximum deflection» is met with, a maximum amplitude is obtained, which does assume the shape of a

sine curve.

The above method of experimenting on

vibra-tion was followed in the case under

consider-ation.

After a prolonged and scrupulously careful

course of experimentation, during which

vari-ous frequencies were used - the propeller

tested was very sensitive to singing in every

key - the natural frequency of the propeller

blade proved to be 235 Herz. At this frequency the pitch was, therefore, followed by the longest resonance when the apparatus was switched off.

Fig. 5 gives a true picture of the nodal lines

and the position of the «maximum deflection»

at a frequency of 235 Herz (see also photo No. 4).

The nodal line No. 1 indicates torsional

vibra-tion, the nodal lines Nos. 2 and 3 flectural vi-brations.

At the point where the nodal lines Nos. i and 2 intersected an entire field was even observed in which no amplitude was visible on the screen

of the oscillograph.

It then appeared that at C (trailing edge) the

amplitudes were maximal, so that there was a

«maximum deflection>' in the vibration area

passing from A via C to B.

This point C was the same point where the

crack had originated.

Fig. 5.

Photo No. 4

Propeller-blade 3 with nodal lines (Knoop) and

welded crack (Scheur).

All this proved that the cracks had been

caused by the singing of the propeller blades.

Later on, the three other blades were auto-genously welded and the trailing edges provided

with anti-singing edges.

Note. The results obtained during the

in-vestigation described here do not prove that every singing propeller must, sooner or later, show cracks. The author knows instances in

which singing propellers have been in use for

a period of 4 or 5 years without their blades

having shown any traces of cracks.

The conclusion may, however, be drawn from

the investigations made, that, in the course of time, cracks may occur in a singing propeller. Shipowners are.

therefore, advised not to

wait too long before having propellers, if they should be singing, provided with anti-singing