Sunteți pe pagina 1din 3

Ñer E c.

ci+ aÇ rccTh cAr4


rarsA#crr'..S
The Radii of Gyration of Merchant Ships
TECHNISCHE UNIVERSITEIT R W PEACH, MSE, PE, CEng. Member
Laboratorium vooc Na va/Architect and Marine Engineer, USA
A K BROOK, BSc, MSc, CEng, British Maritime Technology Ltd
Scheepshydromechanlcs
Archief Report on a presentation and discussion
Mekelweg 2,2628 CD D&ft
TeL: 015- 786873 - Fax: 015- 781836 1 6th February 1987

Predicting the motion response of a ship at sea requires, So far all the equations refer to radius of gyration for
among other things, values for the respective radii of roll. The computer program was also used to obtain values
gyration with reference to the three principal co-ordinate of radii of gyration associated with pitch and yaw, and Mr
axes through the centre of gravity of the ship. Mr Peach Peach went on to describe the computations in some detail.
began his presentation by pointing out that, until the Three ships, each of a characteristically different type,
development of modern computers, direct calculations for
radii of gyration were considered to be too time consuming were used in the investigation: a Roll On/Roll Off (Ro/Ro),
in relation to the value of their application. Instead, various a Lighter Aboard Ship (LASH), and an LNG carrier.
empirical formulae were developed and have been used for Particulars of the ships are given in Table i and the
many years. In fact, as late as 197.7 Mr Peach discovered, calculated radii of gyration are given in Table 2. Mr Peach
from an approach to the US Maritime Administration Table I-Ship characteristics
(MarAd), that no directly calculated values were available.
He suggested that a computer program should be developed SHIP TYPE Ro/Ro LASH LNG
for this purpose and MarAd agreed, adding two require-
ments. The program was to be produced as an entity i.e. not Builder Avondale Bath Avondale
trying to add radii of gyration calculations to any existing Design C7-5-95A C8-5-8lb LG9-5-107a
program for weights and centre of gravity. They also La, Feet 684.00 772.00 931.50
wanted the program to produce values for weight per foot of Metres 208.48 235.31 283.92
ship length as a longitudinal distribution. L, Feet 640.00 724.00 887.00
Metres 195.07 220.68 270.36
Apart from the use of the program to perform direct B, Feet 102.00 100.00 140.50
calculations, Mr Peach thought that it was useful to Metres 31.09 30.48 42.82
consider the results in comparison with some existing D. Feet 69.50 60.00 94.00
empirical formulae. First he presented the simplest type Metres 21.18 18.29 28.65
Displ. Tons 15805 14874 30298
K =C1 B Tonnes 16059 15113 30784
where K is radius of gyration for roll LCG Feet 382.20 400.94 477.05
B is breadth of ship Metres 116.49 122.21 145.40
C1 is a coefficient dependent upon ship type. VCG Feet 38.57 36.59 48.68
Metres 11.76 11.15 14.84
Traditionally, values have been quoted for C between 0.40 TCG Feet -0.16 0.14 0.18
and 0.44. References i and 2 give corresponding values in Metres -0.049 0.043 0.055
the form of C = l.108C1. He then considered the radii of KM Feet 52.20 58.30 118.24
gyration for a thin-walled rectangular tube, which has a Metres 15.91 17.77 36.04
GM Feet 13.63 21.71 69.56
radius of gyration Metres 4.154 6.617 21.20
K = (B+D)/jT Roll Period, Seconds
or C1 = (1 + B/D)/AJ12 T=2 7TKr

This form led him to suggest that perhaps (B + D) may be a /g GM 10.89 8.81 6.70
better parameter than the frequently used (B2 + D2)+
* Not including cantilevers at stern
Another equation given by K ato3 is At side
La = Length overall
(K/B)2=Cl2=F[CbCU+ l.IOCU(l-Cb) = Length between perpendiculars
(HIT -2.20) + (H/B)2 J
where
Cb = Block coefficient Table 2-Radii of gyration
C = Deck area coefficient = (Deck Area)/LB
H = Equivalent hull depth = D + A/Lp SHIP Kr (Roll) K (Pitch) K (Yaw)
A = Profile projected areas of erections and deck- Feet Metres Feet Metres Feet Metres
houses RO/RO 36.2732 11.06 187.8276 57.25 187.3549 57.11
T = Draft LASH 37.0488 11.29 205.0905 62.5 I 204.9517 62.47
B = Beam LNG 50.4408 15.37 239.2502 72.92 239.9572 73.14
F = Constant for the ship type
= 0.125 for passenger, passenger and cargo, and
cargo ships explained that the calculation used relevant information for
= 0.1.33 for tankers every individual piece of structure, outfit and equipment,
= 0.177 for whalers involving between 20000 and 40000 items. The results are
Finally, Mr Peach quoted from his own report4 to MarAd: converted into coefficient form associated with various
parametric terms, and presented in Table 3. For compari-
K = 0.30 (B2 + D2)+ son, values of Kr/B estimated from Kato's equation are
based on the results of the full-length calculations by included in Table 3. These are considerably higher than the
computer. computed values and can only be correlated by using much
115
Table 3-Radii of gyration coefficients showed that, although a 10% variation in radius of gyration
caused a significant change in the natural period of roll, the
SHIP TYPE Ro/Ro LASH LNG roll amplitude did not have a direct correlation. This was
Kr/B 0.3556 0.3705 0.3590 based upon results from calculations giving RMS roll
K/(B2 + 0.2939 0.3177 0.2979 amplitude in various sea states as shown in Table 4.
F
Following this, he suggested that highly accurate and
0.125 0.4278 0.3400 0.5831 detailed calculations were not essential, although logically
Kato's 0.133 0.4413 0.3508 0.6014 the best possible estimate should be used. He was more
K /B 0.17 7 0.5090 0.4046 0.6938 concerned about roll damping coefficients and considered
Required F 0.0483 0.0707 0.0336 these to be a more serious problem than radii of gyration
because damping coefficients are difficult to estimate from
K p/La 0.2746 0.2657 0.2568 available data. In his opinion, more experimental work was
KI(La2 + D2)+ 0.2732 0.2649 0.2555 necessary in this respect. Accurate radii of gyration can be
0.2935 0.2833 0.2697 calculated if required, but the same is not true for damping
K /(L2 + D2)+ 0.29 18 0.2823 0.2682 coefficients at the present time.
Ky/La 0.2739 0.2655 0.2576
K/(L + B2)+ 0.2709 0.2633 0.2547
0.2827 0.2831 0.2705
THE DISCUSSION
Ky/(Lp2 + B2)+ 0.2891 0.2804 0.2672 Dr I L Buxton, opening the discussion, referred to the
long history of estimation of radii of gyration and reiterated
L = Length between perpendiculars
La = Length overall Mr Brook's warning about the misuse of empirical data
especially, for example, Kato's equation based upon
smaller values of the coefficient F, compared with those information from the 1950s. He considered it timely to
recommended by Kato. update such information because of the increasing aware-
ness of the need to predict ship motions. This is very
From the same table it can also be seen that the important, for example, in the way that pitching affects ship
calculated values of radii of gyration Kp and Ky, for pitch performance. He referred to a paper by Swaan and Rijker5
and yaw, are significantly higher than the frequently used K which demonstrated that speed loss at sea could be a
=0.25L. consequence of particular longitudinal weight distributions.
Mr Peach concluded by emphasising that the computed Dr Buxton asked Mr Peach whether loaded conditions had
values did not take into account the effect of entrained been examined and was informed that they had not been
water in the context of added virtual mass. included in the study. The calculations were for 'empty'
ships. This aspect was referred to by other contributors
A complementary presentation was then given by Mr during the discussion and Mr O M Clemmetsen pointed out
Keith Brook of British Maritime Technology. He started that typically the cargo weight was about two thirds of the
by identifying himself as a user of this type of data as total loaded ship weight.
distinct from an analyst such as Mr Peach. His experience
has been particularly in the context of using estimated Mr J Whatmore, doing post-graduate work on midship
values of radii of gyration in computer calculations predic- section scantlings by computer at Newcastle University,
ting ship motion response and his comments were largely in indicated that he expected the work to progress towards the
relation to rolling. Mr Brook had tried Kato's equation but inclusion of radius of gyration. He asked Mr Peach if his
was more confident using that proposed by Bureau Ventas computation had indicated whether the results were sensi-
for radius of gyration in roll tive to particular items, i.e. could some items be treated in
less detail than others? Mr Peach indicated that all items had
K/B = 0.289-/ i + 4 (KG/B)2 been treated in full detail. However, Mr W Hills reworded
He expressed the view that it should be remembered that Mr Whatmore's query and asked if some degree of
Kato's equation was derived some years ago and it was 'lumping' was not possible, i.e. using blocks or groups of
probably unfair to expect it to relate to modern vessels of items, without significant loss of accuracy while achieving a
special types; the limitations of empirical formulae are not worthwhile reduction in data preparation. In reply, Mr
always appreciated. In comparison with the range of K/B Peach simply reiterated that no lumping of data had been
from 0.40 to 0.44 given in References I and 2, Mr Brook considered.
mentioned that values used in the UK were usually in the Mr D Brown referred to the importance of good values of
range 0.35 to 0.40 for a loaded cargo ship. The significance roll radius of gyration for warships and fleet auxiliaries
of this was raised in the subsequent discussion. because MOD (Navy) ask for design estimates of the
Describing a recent sensitivity study with which he had motions of specific locations such as helicopter decks. Dr
been involved, Mr. Brook presented information which Buxton added a comment in a similar context, referring to
the importance of all three radii of gyration in the current
Table 4-Effect of error in roll radius of gyration on roll debate on the 'short fat' versus 'long thin' warship hull.
response It was then pointed out by Professor J B Caidwell that
RMS ROLL (DEGS) though roll amplitude may not be very sensitive to accuracy
VESSEL K/B T SIG.WAVEHEIGHT(m) of radius of gyration, rolling accelerations were likely to be
(SECS) more relevant because they are dependent upon radius of
3 5 8 gyration squared. This could be important in relation to
acceleration forces on such items as containers stowed on
OFFSHORE 0.36 6.8 5.5 8.3 10.7 deck.
SUPPLY 0.40 7.5 5.1 7.9 10.6
0.44 8.2 4.6 7.4 10.3 Mr J Davison suggested to Mr Peach that while (B + D)
FISHERY 0.36 9.4 4.8 8.6 13.0 may be a relevant parameter for the radius of gyration of a
PROTECTION 0.40 10.5 3.7 7.3 12.0 hollow rectangular shell, it might be more appropriate to use
0.44 11.5 2.7 5.8 10.6 (B2 + D2) for loaded ships by analogy with the radius of
0.36 12.7 1.1 2.2 4.5
CONTAINER 0.40 14.1 0.8
gyration of a solid rectangle. Referring to added virtual
1.7 3.5
0.44 15.5 0.6 1.3 2.7
mass, he pointed out that it was not just radius of gyration
which was affected, but also the total moving mass. it
116
would seem preferable, he suggested, to treat added virtual Dr T Svensen, in proposing a vote of thanks to Mr Peach
mass separately as in vibration calculations. Both Mr Peach and Mr Brook, mentioned that he would treat radii of
and Mr Brook agreed that there was scope for more gyration with much greater respect in future, and summed
experimental work in this respect. up the discussion by referring to the general consensus that
The President, Dr Mime, asked if any data had been the presentations had re-emphasised the need for further
collected from the actual behaviour of ships at sea. Mr research on this important topic, especially in relation to
Brook replied that consideration had been given by the increasing interest in the prediction of slamming, deck
offshore industry to continuous monitoring of the metacen- wetness and similar seakeeping problems.
tric height of a vessel by measuring a vessel's natural roll
period, i.e. using an estimated radius of gyration in the REFERENCES
equation T = 2 r KA/g GM. However, the difficulties of Principles of Naval Architecture, SNAME, 1941
determining the vessel's natural period accurately mean this Principles of Naval Architecture, SNAME., 1967
method is problematic. Mr Whatmore asked if Mr Peach KATO, H. On the Approximate Calculation of Ship's Rolling Period,
was able to provide, for publication, information on the JSNAJ, Vol 89, 1956
weight per foot of length distribution which had been PEACH. R. W. ENGINEERING ASSOCIATES, FinaiReport on Study of
Ship Radii of Gyration for US Department of Commerce, Maritime
produced as a supplement to the calculations for radii of Administration, April 1979
gyration. Mr Peach readily agreed and Fig. 1 is included for SWAAN, W. A. and RIJKER. H. Speed Loss at Sea as a Function of
this purpose. Longitudinal Weight Distribution, Trans. NECIES, 25 Jan 1963

TONS PER FXT


- 220

- 200

- 180

- 160

- 140

-120 CONVENTIONAL TRAPEZOIDAL


WEIGHT DISTRIBUTION CURVE

- 100

-80

-60

i 20

0
h50 0O 350 300
DISTANCE FROM FORE PERPENDICULAR IN FEET

Fig. 1Distribution of weight per foot of length for LNG (LG9S-107a)

117

S-ar putea să vă placă și