Progress in tank stabilizers

(1)

PROGRESS IN TANI STABILIZERS A. J. Giddings Naval Architect Bureau of Ships Navy Department Washington, D.C. L INTRODUCTION

Tank stabilizers using the U-tube analogy have been designed for a variety of naval ships. None of these have been combatant ships, but they have been ships which meet one or more of the criteria for passive tank stabilization as expressed in reference (1). Repeating these criteria:

Low speed operations are a major characteristic.

Hull penetrations for retractable fins are not practicable and retraction is required or desired.

Only "some" rather than "a lot of" stabilization Is needed. Cost is more important than the degree of stabilization. Space and weight are available with little penalty to the mission.

Ship types which have met the requirements are icebreakers, missile range ships, oceanographic research and survey ships, and

other special purpose ships. More than 17 designs have been completed. The very virtues of the free-surface type passive stabilizer

tank that make it attractive in application, work against a large investment of time and effort in the scientific analysis of their performance. The tanks are àheap to install, and perform well with-out a great deal of design effort. As little as 3 man hours have been spent (by an experienced designer) on the geometric design of some tanks. More time, of course, is Involved in structural design and

performance prediction. A design data sheet will be Issued in the

JAA

near future (it is awaiting printing) which will simplify the

geometric design process still further.

(2)

In spite of the foregoing de-emphasizing factors, certain results.

are available and a limited amount of analysis of them has been made. References (2) and(5) are some of the published model test results while references. (6) and (7) .report full scale results. Through the cooperation of Military Sea Transportation Service the trials of reference (7) were especially complete in that a number of measured sea states were obtained and variations in tank water level, heading and speed were all obtained. References (2) and (9) are computer predictions of stabilizer performance.

II. FULL SCALE

Analysis of the full scale reports compared with predictions has been limited by a lack of suitable full scale trials.. For example, references (2), (6) and (9) should be directly comparable, but the

trials of reference (6) do not provide sufficient data. The trials. of USNS ELTANIN, reference (7), were conducted in December 1961, and the report issued very recently. . For this reason, little can be said in detail about the quality of the predictions. However, since the ELTANIN trials were so complete, it is expected that cnsiderab1e analysis and comparison will be forthcoming.

The method of analysis of full scale motions which is being used as a design tool is a highly linearized and de-coupled approach to the equations of motion. The full-scale runs selected for analysis of roll performance are low speed beam-sea runs. The measured sea state spectrum Is used as an input in the analysis, so such measurements are necessary.

The unstabilized ship is considered as a single degree of free-dom harmonic oscillator, and the amplitude response curves for

several damping coefficients are plotted. The sea spectrum is con-verted to a wave slope spectrum by:

,2(W)

cn4 H2(cn)

LA1i

(g)2 Lw

The "effective"wave slope on the ship is considered to be a function of the depth of the center of buoyancy and the ratio of

ship beam to wave length.

(3)

The effect of ship beam is obtained by integrating the static moment generated by a sinusoidal wave and relating this to the moment. obtained from a wave of infinite length and slope

_'I'e

or:

3 [2

+.

>4 _)..

where: = effective wave slope

= effective beam (actual beam times the prismatic coefficient) = wave length

This may be sufficiently approximated by

li

r'-(j1

The .-"effective" wave slope spectrum is then:

r'. 8w

L4)

-

/

)

4 - L

1A4J

AW

This spectrum is multiplied by the square of amplitude response to obtain a roll spectrum. The average values of "predicted" roll and measured roll are compared to obtain a value of "full scale" linear ship roll damping.

A similar approach is used in analyzing the stabilized runs, except that the amplitude response curves used are those for the stabilized ship and various values of velocity-squared tank damping coefficient are investigaed.

The same approach has been used to analyze other ship motions. The details of the shape of .the predicted ship motion spectra do not agree with the measured spectra, but the approach is useful for comparison of various designs.

(4)

III. MODEL

The larger number of model test reports has permitted the

corn-p].et ion of some analysis. This analysis has been aimed at verifying

the design process, which is aimed at selecting a tank geometry which will provide the desired frequency. Figure 3 summarizes the approach used in design at the Bureau of Ships. The contribution of the

nozzles to the calculated frequency is usually very small so that the "basic" frequency can often be used alone.

Table 1 shows a comparison of the design frequency versus the measured frequency for 7 designs at different water depths. The measured frequency was determined from phase measurement of tank water transfer or moment as compared to roll angle. The moment and water transfer phase angles are usually close to each other. The tank models were oscillated in each case in 'pure roll," in that the roll axis of the model tank was located to scale, and no sway of that point was permitted.

Table 2 is a similar comparison of the design approach with "exact" theory as taken from water wave theory in a rectangular tank.

Figure 1 is a graph of the values of tables 1 and 2 versus the parameter 1- It can be seen that both the model results and the

exact theory bear the same relation to the design approach using the U-tube analogy. Even those model tanks having narrow cross-over ducts fall in line. (Ships 2, 3, 4, 6 and 7). Figure 2 is a plot of the same data arranged to investigate the dependence on water depth. This Is not a strong dependence except for large values of

Practical considerations of U-tube "moment" terms will lead the design to values of between .4 and .7, so that this effect as shown by "exact" theory is not important.

It is recommended that the curve of figure 1 be used to correct the U-tube frequency in the design process.

E

C

I

a

I

I'

I

(5)

Roll Stabilization by Means of Passive Tanks Vasta, et a).

-SNJ

Annual Meeting, November 196].

Experimental Mode]. Investigation of Heeling Tank Stabilizer for AK253 - Odenbrett and Yamanouchi, ETT Letter Report 725 Experimental Study of a Passive Rolling Tank Stabilizer Installation for a Pacific Missile Range Recovery Ship AG(PMR) - Odenbrett, Davidson Laboratory Report 778

Experimental Investigation of a' Passive Rolling Tank Stabilizer Installation for an 0ceanogaphic Research Ship (AGS)

Odenbrett, Davidson Laboratory Letter Report 799

Model Test of a Passive Anti-Roll Tank for a Class 111 Survey Ship - Russ, DT}IB Report 1618

A Full Scale Evaluation of Passive Anti-Roll Tanks Aboard an AK-Type Ship - Go].ovato, DT Report 1414

Preliminary Evaluation of Passive Roll Stabilization Tanks Installed Aboard the TJSNS ELTANIN (TAX 270) - Foster, DTMB Report 1632

Analog Simulation of a Passive Anti-Rolling Tank System for an Oceanographic Survey Vessel -' Oliver and Church, DT Report 1233 Analog Simulation of a Passive Anti-Rolling Tank System for a Missile Range Ship - Church, DTMB Report 1322

REFERENCES

(6)

TABLE 1, COMPARISON OF MEASCJktED TANK FIIEQUENCY AND U-TUBE ANALOGY FREQUENCY

C')

SHIP DES MEAS

4

A. 1-1

.800

.695

..0

1.00 .053

.633

.869

.902

.779

.606

1.00 .068

j

.633

.864

2

.511

.460

.51

.57

.068

.74

.900

3

.666.

.675

.389

.50

.111

.849

1.016

4 .462

.470

.462

.50

.082

.838

1.017 .504

.501

.462

.50

.098

.838

.994

.542

.533

.462

.50

.114

.838

.983

5

.550

.452

.709

'1,00

.042

.497

.822

.632

.535

.709'

1.00 .056

.497

.846

.701

.603

.709

1.00 .070

.497

.860

6

.661

.671

.486

.396

.108

.764

1.015 .775

.743

.486

.396

.149

.764

.959

.867

.912

.486

.396

.189

.764

1.051

7 .570

.504

.51

.57

.085

.74

.884

TABLE 2. COMPARISON OF ACT WAVE THEORY IN A

FtECTANGTJLAR TANK WITH U-TUBE ANALOGY

WJ_tu

-

I V .

Ai-f

,4.-t

44

I

C

eA33

_t

Ak

0

.4

.5

.6

8

1.0 .04

1.112 1.Ô20

.964

.892

.671

.086

.06

1.112 1.b21

.966

.894

.676

..132

.08

1.114

1.023 .968

.897

.682

.176

.10

1.116

1.026 .972

.902

.692

.

.219

.15

1.121

1.035 .983

.917

.720

.322

[

L

(7)

WOLS. OR U LO S LUu-yueg. I I I I I I I -.1 .2.

.S

.4

.5

.0

1

.8

.9

1.0

I_f

A COMPARISON

OF DESIGN VALUES WrrH MEASURED OR

)' .4

UEXACT7S VALUES

_{AS A rUNCTIOPI OF t-Y}

_F

RANGE OF MEASURED VALUES.

RANGE OF EXACT VALUES FOR

.04 TO

- .10.

FIGI

(8)

OR

W

!XACT DES. (1)

j.

WOES.

.6

-Y=I.O I I I I 1 1 I I I

.02

.04

.06

.08

.10 .12 .14 .16 .18

20

THE DEPENDENCE OF AND WIXfr$CT ON .

AND Y.

QJDES,

NUMBERS IN PARENTHESES ARE Y.VALLJES FOR THE

a

I

(9)

TANK NATURAL FREQUENCY- U- TUBE

ANALOGY

NOZZLE PLAN

j

₀

/28

43. kt tt= -:+-c

7

"Mapu(

"F'1

ARE GRAPHED

FI&3

A. J. Giddings

g

-.

RAD/5rO.

L.t

-=

Ci

{-F

_-

_i]

A

SURFACE AREA OF ONE WING" TANK

= CROSS SECTION AREA AT '&!'

4. DISTANCE ALONG. "i" OF EQI)IVALENT U-TUBE.

_28

("BAsic"

PLUS "NOZZiJ")

M

₄

. 'C N