Contribution to ISSC 1976 - report cornr'ittee 1-2 FULL-SCALE MEASUP.E!NTS OF WAVE-INDUCED LOADS
V. Ferdinande
Several research centres have been involved in full-scale measurements, with a view to
long--term predictions of wave loads - the results used as a basis for, and finally as a verification of the existing long-term prediction techniques,
(Ref'. 1,2,14,9),
correlation with theoretical calculations and model experi-ments, and verification of some statistical concepts, (ef.5,6,8),
an explanation of observed damage to ship structures, (cracks), (Ref. 6,7).
An attempt to seperate whipping and sringing stresses has been made in Ref. 1, 2, 7.
The results of the measurements still are presented in different units, and in diffrent forms (stress, several R.M.SS values, non-dimensional bending momente, ...). This makes a direct com-parison of various vessels difficult.
The knowledge of measured stresses is significant for the inves-tigated vessel itself, but it is of less value to the assessment of wave loads, if th midship section modulus is not
cornrnuni-cated.
For the reasons given above, the midship stresses, given in various units in the different publications, are converted into a (wave) bending ioment coefficient , where M X x
pLB
M wave bending moment (peak-to--peak), X peak-to-peak wave bending stress (reversall midship section modulus. Where
V
possible, following presentation is given - averag9 RMS
pL.B
bending moment coefficient. (average of the root mean square of
th bending moment coefficint R.M.S. refers here to peak-to-peak reversals on the record, so EM 8 a2 where a2 j the
variance of the bending moment record)., or, whn there is given
1.
N max pgL3B
- avg. max. bending moment coefficient.
The dimensional results ar given here in S.I. units. So, stress 1 kp (or kgf)/mm2 9.807 MN/rn2 (meganewton per sq. m), 1 KPSI
6.895 rmrim2 1 tonhin.sq. 15.145 MN/rn2 section modulus : 1 in2.ft:
0.1966 x
io
rn, etc..This is the same presentation as used in the report of Committee 3 (Wave Loads, Statistical Aporoach) of the 'I.S.SC. Ramburg 1973", so the full-scala data nertaining tO the ships given in the present Tabel I (Ship particulars) can directly be compared with those
given in the latter report. Hence, the here included graphs should be considered as cornolc?menting the correspondinc graphs in the
former r'port. Also for comperison purposes, a few curves from the former graphs are reproduced on the graphs here presented.
Ref. Ship type Tanker (1) "Ship A"
(6)
JADEBulk carrier
(8)
T.V."0ssendr'cht
175.65
20.60
13.25
O.B.O.
(1)
"Ship C"
281.98
44.26
23.39
117.51 Containership(1)
"Sh±p B"
213.36
30.118 16.11619.30 14.90
(1)
"Ship D'
218.00
30.11818.60
i1.78
(3)(11)
S.S. Sea-LandMc Lean
253.11132.16
19.52/20.89
Dry cargo-liner (5) rI!.S. Esquilino137.00
20.00
8.91
5.51
Bulk/Orecarriers
(2)
Yakuinokawa Maru(2)
Shobu Maru
(2)
Kimi Maru
Cable sh4p
(7)
Marcel-Bayard
109.66
15.63
8.50
Ship particulars Table I B D section in in mmodulus
in330.71
51.82
25.60
77.2
329.00
51.80
25.60
220.00
32.20
18.50.
22. 711236.22
31.85
18. 75
23.014238.00
32.20
18.20
23.20
Measuring apparatu
arid measuring technigues
- Ship "A", "B", "C", "D" (Ref.1) - !3.S.R.A. - Data loggers to
measure main hull bending stress, numerical filtering technique
to separate the wave induced stresses from the springing stresses
20 mm. period from every 12 hours. Statistical strain gauges, to
record the number of excedences, at predetermined stress levels
(out-to-out lonc'itudinal stress excursions in deck structure)
- Yakumokawa Maru, Shobu Maru, Kimi Maru (Ref.2) - 118 th.
Research Comm.itte,.Jaoan - straingauges (double bottom - upper
deck), dynamic strain meters - stress cyle counter, filtering
to separate wave induced stresses from whipping stresses. Green
sea max. pressure gauge and counter.
- S.S. Sea-Land Mc. Lean, (ef. 3 &
1i)- S.S.C. - Vertical bending :
longitudinal stress gaues P & S under deck, near midship. Midship
torsional shear
:shear rosettes amidship P & 5, on si.deshel]. at
neutral axis. Hull acc'lerations
:verttcal and transverse
accele-rorrieters located at vessel CG., and forward (rigid body as well
as whipping motions)
indicative of most severe accelerations
on vessel. Horizontal bending
longitudinal stress gauges
,
P & S, near midship at neutral axis. Shear-forward
shear rosettes
near forward quarter point, on sideshell, at neutral axis.
Shear-aft. Deckhouse accelerations (vertical and transverse, possible
springing or higher frequency vibratory effects). Box-girder shear.
Wave-recorder
:Tucker Wave Meter and Ocean Wave Height Radar
System.
"Scratch Gages"
simple extensorneters, measurements of the
length of the scratched line will establish thetotal stress
range experienced due to all causes (will be installed aboard each
vessel in the SL-7 class to gather many ship-years of stress data).
Data interval (or record) length
:30 minutes, 1) for statistical
sampling of data
:automatic data-taking period every four hours,
2) manual
:for higher as B7-8.
M/S Esquilino (Ref.5) - CETENA - Dynamic stresses midship (11
extensometers maindeck. Wave measurements
:pressure pick-ups
- Jade, Emeraude, (Ref.6) - B.V. - Longitudinal deck stresses P & 3, extensometers - torsion - shear stresses on longitudinal bulkheads and transverse members. Spectral analysis (analog
computer), MS.
- Cable-ship Marcel-Bayard, (Ref.7) - B.V. - Quasi-static and dynamic stress measurements by extensometers on castle-deck and in double bottom, measurement of vibrations. Wave measurement by means of accelerometer buoy.
- M/V. Ossendrecht, (Ref.8) -, T.NSO. - Strain recorder able to give punch-tape consecutive recordings of 200 points, at a
maximum speed of 2L points per second. The ordinate-values
were
recorded of the strain-response at each connectedmeasuring
point, equally spaced with a tim-interval of 1.2 seconds.Converted full-scale data
According to the form of prrjs.ntation in the various publications, the corresDonding non-dimensional data are brought into different
graphs, viz, the cumulative probability (of excedence) of the wave bending moment coefficient
-(Fig.2), the average
pgL
B-R.M.S. bending moment coefficient versus Beaufort number pgL B
(Fig.3), and versus significant wave height H113, (Fig.), and Mnax
the average maximum bending moment coefficient
'
versus
pgLB
Beaufort number, (Fig.5). These graphs of
non-dimensional
data are preceded by agraph of
cumulative probability of stresses,(Fig.1), although the latter are less indicative for the purpose. of comparison.
Fig. 3 has been complemented with converted data of the WOLVERINE STATE (North Atlantic) from Ref.9, viz, a correction to the cor-responding curve in Fig.II-1, in the report of Comrnittee3(Wave Loads, Statistical Approach) to I.S.S.C. 73 Hamburg, as already mentioned in a contribution to this
report.
5
Discussion of the results and_conclons.
Pig. 1 contains information on curnative probability (of exce-dence) of stresses on the "B.S.R.A.-ships" (nef.1), the "Kirni Maru" (Ref.2) and the S/L. Mc. Lean (Ref.), converted into ben-ding moment coefficients in Fig. 2. Regarben-ding the ships A, B, C and D (B.S.R.A.), several long-term prediction techniques were applied, but we only retain the results 1) of the method based on the R.M.S. of the wave induced stress reversals in each record
from the data logger, and 2) of the method based on the assumed double exponential distribution of the maximum stress reversals
(wave bending + springing) over a number of watches chosen at random from the statistical strain gaupe. The agreement between the results looks rather insufficient.
Remarkable is the very high maximum stress reversal (after an extreme slam ev3rlt) on the Sea-Land Mc Lean, which (after 1928 hours) was plotted at a probability
io6.
The environmentalconditions : Beaufort.12,.Wi-fld speed 100 knots, avg. wave height..
15 m,hove to condition. Before and after the slam event, in the same weather, the recorded maximum stress is seen to be much lower, the difference being indicative for additional whipping stress. Also according to the stress record of this slamming event in Ref.I, 50 per cent should be added to the low-cycle wave
bending stress. In Fig.5, a clear distinction appears between the curves related to stresses, springing included and excluded res-pectively, (Ref.2).
Fig. shows that the R.M.S. bending moment coefficients for the
cargo-liner 7v.S. Esquilino (Ref.5) agre' well with those of the similar cargo-liners published formerly (I.S.S.C. 73).
Regarding theoretical concepts, the analysis of the full-scale measurements in Ref.1 confirms the assumption (E.V. Lewis). that the R.M.SS of the reversals is normally distributed within each Weather Goup. Also in Ref.1, it is shown that the equation, proposed by Rice, and used by MILES, defines the short-term be-haviour of the combined wave inducd and springing stresses tretty well.
6.
From full-scale measurements on a bulkcarrier, besides the gaussian character of strain-response to wave-loading, therayleighan
character of the short-term distribution of peak-to-peak zero-crossing values could be demonstrated in Ref.8. The relationship
is confirmed by the results. The property that 8 m0-values are equal to the rnan square height values E (valid in case of very narrow spectra) is not valid for broad-bounded spectra. E has to be taken as the parameter in the Rayleigh distribution
formula, arid not 8 m0.
A comparison of measured RMS-values on a tanker (spectral analysis) and theoretically calculated RMS (ef.6), showed a good agreement
for vertical bending stress in bow and quartering seas, but rather poor agreement for shear forces in sideshall and
longi-tudinal bulkheads.
According to Ref.5, theoretical calculations would give higher peak values and lower corresponding peak frequencies in the
spectra than measured full-scale.
The usefulness of full-scale experiments in a general attempt to explain the apocarance of cracks (fatigue) is emphasized in
Ref.7, for special vessels, and in Ref.1 and 6, for container ships. In this respect, the cracking on the fast containership SS. Sea-Land Mc. Lean (Ref.14), noticed after heavy weather en-counters and an extraordinary slam event, is of particular interest.
Recommendations for further research.
The full-scale data, gathered on the SS WOLVERINE STATE and on other similar cargo-liners, have been the basis of the prediction of the long-term distribution of bending moments (vertical, late-ral, combined vertical & latelate-ral, torsion) by the authors of
Re.f.1O. The same may be expected in the near future for the SL-7 class containership, for which up to now only a few data are published. Further systematic communication from all research centres of full-scale data will render a basis for and a means of verification for feasibility of the different techniques for long-term prediction of wave-loads on various ship types, and of
7.
actually adopted theoretical concents as wll. However, the necessary comparison of different ships would be made easier by
adopting a commun system of units (5.1.) and presentation, and each communication should content a more complete information on ship particulars and on. the ship's structure, in particular the section modulus.
REFERENCES
G. WARD and M. KATORY, "Data on Midship Bending Stresses from four Ships", International Symposium on the Dynamics of Marine Vehicles and Structures in Waves, University College London,
May 1971!.
The 118 th Research Committee, "Full Scale Measurements of Hull Stresses of Large Bulk/Ore Carriers", Report of SR-118, no. 72, The
Shipbuilding
Research Association of Japan, Tokyo, May 1972.R.A. FAIN, "Design and Installation of a Ship
Response
Instru-mentation System aboard the SL-7 Class Containership S.S. Sea-Land Mc. Lean", SSC-238 (SL-7-1), Ship Structure Committee 1973. 14 Teledyne Materials Research, "Sample Data recorded duringSecond Season Heavy Weather Periods", Technical Report No.E-1559
(1), March 1971!.
M.G. BRIA et G. SAPTIRANA, 'Comportement la mer du M/S "Esquilino", Bulletinde l'Association Technique Maritime et Aronautique, N°73, Paris 1973.
M.D. HURE, "Mesures dynamiques a bord d'un grand Détrolier, "Bulletin Technique du Bureau Veritas, Paris, avril 1972,
no.14.
J. OSOUF, "Etude exprimentale du comportement dyriamique sur houle du navire cblier "MARCEL-BAYARD", Nouveautés Techniques
Maritimes, Journal de la Marine TMarchande, Paris, 1973.
8 F.X.P. SOEJADI, "Full-Scale Measurements of Stresses in the
Bulkcarrier M.V. "Ossendrecht", 2nd Progress Report, Netherlands Ship Research Center TNO, in preparation.
8.
D. HOFFMAN, R. van HOOFF andE.V. LEWIS, "Evaluation of
Methods for
Extrapolationof Ship
Bending Stress Data??, Techn. Pep. SSC-2314,Ship Structures Coxrniittee,
1972. E.V. LEWIS, D. HOFFMAN, W.M. MACLEAN, P. van HOOFF, i.B.ZUBALY, "Load
Criteria for Ship Structural Designt1,
SSC-21t0,+ Mox.SIL Mc.Lean slam B12 (whipping) 340
-.
U)
LPLP 300 ci'-E1 E 260 g -.2 S/L Mc. Lean 220 E C a' .! D Lii-
-°
jHIP'B
after modif iccit ion
a:: 140 .
!
Li
C)iQq_
S 'kcillQsL_ >%E 100.ON-- II
Uo
LSHIP'A I batlasF MIDSHIP STRESSES KIMI MARU\
Max.Meas.Stress \\ SHIP'D' Max.Meas.Stress\\
\\
SHIP'C' " Max.Meas.Stress 0/
Statistical gauge stress cyclesSHIP 'A'
-
---Trend based on R MS. of stress Max.Meas.Stressreversals logger data.
,20 1twave inauc1ed.). CUMLATIVE POBABILIT'
io io_2 0 1
I
'. SHIP 'A' LOAD CONDITION
I I
/
o .. 'A BALLAST ,, 'B' PORT GAUGE 'B' STBD .. 'C' wave bending + springing (statistical gauge) HIPB' + II D.1010 H pgL3B --4 PU) 15x10 U.. Q)
.. SHIP A LOAD CONDITION
E I SHIPB
U...Y KIMI6efore modification
U) '-'
E MARU
--iep.
U)
0
a KIMI MARU (wave induced)
.2 iOx1(5'j .ü SHIPB D ci fterrnodication KIMI MARU > . I Max. ..' C 0.)< () * ISHIPC attast
-
SHIPA"-o Load I baLLast\
N
. IN..
PP IIN
.
'A'N
N
b'-N
N
S±IIP IbaRast L(wave inducpid8
BENDING MOMENTSHIP 0 Max. Meas. Stress.
\
"N
-9LLgst
SHIPSHIP 'A'
--
Mcix.Meas.St --i.""N
Mcix.Meas. Stress.
4-Statistical gauge stress cycles
Trend based on R M.S of stress reversals, loqqer data.
UMULATIVE PROBABILITY (of excedence)
4
-3
io 10F.2
io2
B PORT GAUGE ,. 'B' STBD o 'C' + 'D' -' wave bending + springing (statistical gauge I 1 IlxlO
KMIb) KIMI MARU. Baltasted
I
KM(fI) ,, , FuU Loaded
/
/N. /
SM SHOBU MARU YM = YAKUMOKAWAMARU i.N
N
EM = ESSO MALAYSIA/
I/
FOTI FONTINI L I/
FOLL = R.G FOLLISi\
/
/
Y7 M.S = MU4ERAL SERAING X - (WoLverine State/
/ /
/
Hoosier State NORTH ATLANTIC/
X
, /
/
/
/
WAVE INDUCED VERTICAL BENDING MOMENT.
Avg. R.M.S Bend Mom. Coëff.
-I
/
./
/
./
/
,
/
/
/
'-.4
/
/
I 1 I I/
I/
I/
I/
pgL3B 3x1O BEAUFORT NUMBER. 1 1 1 6 7 8 9 10 1 2 4 .5 Fig. 33x1O4
2x1O
ixiO4
-
Avg. RJV1.S Bend. Mom. Coéff.pg L3 B
1
WAVE INDUCED VERTICAL BENDING MOMENT.
Head seas
Bow seas
E squitino J Jordciens L Lukuga I I I I I I 2 3 4 5 6 7 F1g4 w/3 (H113)1
pgL3B Avg. Max. Bend. Mom .Coëff. 7x10' 6x1O 5 xl 4 xlO I I 1 2 WS.(w.NA}=Wo[verine State (winter, North Atlantic)
B.O = Boston
WAVE INDUCED VERTICAL BENDING MOMENT
3x104