Alexander von Graefe
ISMT,
University Duisburg-Essen,
Bismarckstrasse 69,
Duisburg 47057, Germany
e-mail: alexander.von-graefe@dnvgl.com
Ould el Moctar
ISMT,
University Duisburg-Essen,
Bismarckstrasse 69,
Duisburg 47057, Germany
e-mail: ould.ei-moctar@uni-due.de
Jan Oberhagemann
DNV GL,
Brooktorkai 18,
Hamburg 20457, Germany
e-mail: jan.oberhagemann@dnvgl.com
Vladimir Shigunov1
DNV GL,
Brooktorkai 18,
Hamburg 20457, Germany
e-mail: vladimir.shigunov@dnvgl.com
Linear and Nonlinear Sectional
Loads With Potential and Field
Methods
A Rankine source method is applied to predict linear and weakly nonlinear sectional
loads o f a modern container ship. The method uses solution in the frequency domain, lin
earized with respect to wave amplitude about the nonlinear steady flow due to forward
speed, which accounts for the nonlinear free-surface conditions, ship wave, and dynamic
trim and sinkage. Weak nonlinearity of the sectional loads in waves (e.g., hogging
sagging asymmetry) is taken into account by pressure extrapolation and integration up to
the estimated actual water line. The sectional forces obtained with this method are com
pared with the results of other methods, including (1) linear Rankine panel method,
where flow due to waves is linearized about the double-body flow, (2) linear zero-speed
Green function method with correction for forward speed, (3) fully nonlinear simulations
based on field-based solution of Reynolds-averaged Navier-Stokes (RANS) equations,
and (4) model tests. Comparison with RANS solution and model tests shows, that the pro
posed method can accurately predict sectional loads for small to moderate wave heights.
[DOI: 10.1115/1.4026885] "
Introduction
A ship hull girder is dimensioned with respect to extreme loads
and required fatigue life. Extreme loads strongly depend on non
linear effects; sometimes, nonlinearities are classified as weak
nonlinearity (taking into account real wet surface) and strong
nonlinearity (impact loads and whipping). Accurate computation
of strong nonlinear effects is only possible with field-based meth
ods, such as finite-volume methods for Reynolds-averaged
Navier-Stokes equations, and therefore requires much computing
time. Besides, there are no straightforward statistical techniques
for the definition of long-term extreme loads using such methods.
Therefore, simplified approaches, based on a combination of lin
ear hydrodynamic analysis (and linear statistics) with nonlinear
analysis in selected regular design waves, are currently used in
design and approval. Current developments of this technique
include nonlinear simulations in irregular design wave trains [1]
and nonlinear Monte Carlo simulations in selected irregular
“design seaways” [2]. All these methods are based on the assump
tion that linear computations are an accurate identifier of the com
binations of seaway conditions and operational parameters that
lead to extreme nonlinear responses. Therefore, accurate calcula
tion of linear responses in the frequency domain is an important
prerequisite for the assessment of nonlinear maximum global
loads when such methods are used.
For the estimation of fatigue life, the loads are frequently sepa
rated into two components: so called wave loads (i.e., rigid-body
responses) and dynamic loads (elastic responses, such as springing
and whipping). Wave loads produce the main contribution to fa
tigue damage; besides, it is important for the designer and ap
proval body to know what contribution to fatigue damage comes
from rigid-body responses and what from elastic reactions.
Calculation of elastic responses will be addressed elsewhere; for
fatigue-relevant rigid-body loads, linear loads are again important.
’Corresponding author.
Contributed by the Ocean, Offshore, and Arctic Engineering Division of ASME
for publication in the Journal of Offshore M echanics and A r c t ic E ngineering .
Manuscript received June 18, 2013; final manuscript received February 11, 2014;
published online April 1, 2014. Assoc. Editor: Dominique Roddier.
Finally, an efficient way to take into account weak nonlinearity
in potential methods is to use linear hydrodynamic analysis in the
frequency domain to compute added masses, damping, and wave
excitation forces (and use impulse-response theory or state-space
model to compute linear hydrodynamic forces in the time do
main), combined with the calculation of nonlinear forces and
moments due to the incident wave and nonlinear hydrostatic
restoring forces and moments by pressure integration over the in
stantaneous wetted surface in the time domain; when using this
approach, accurate linear analysis in the frequency domain is
required.
Therefore, the aim of this paper is to address the prediction of
linear sectional loads. Nonlinear loads and their accurate predic
tion are not considered here; however, a simple nonlinear correc
tion in the frequency domain is tested to check its applicability.
Two linear methods are used: (1) the zero forward speed free-
surface Green function with encounter-frequency correction to
consider forward speed effect and (2) the new Rankine source
method [3]. The Rankine source method directly takes into
account the effect of forward speed, whereas the zero-speed Green
function approach uses a simple approximate correction to take
this effect into account; comparison of these two methods is one
of the aims of this paper. Both methods are implemented in the
code GL Rankine (GL stands for Germanischer Lloyd).
Few open benchmarking data for linear methods in the fre
quency domain are available. Reference [4] compares strip-theory
computations from 24 organizations of midship vertical shear
force and bending moment in head and following waves, and mid
ship lateral force, bending moment, and torsional moment in quar
tering (60 deg off stem) and bow (30 deg off bow) waves with
model tests for an S-175 container ship at Fr = 0.275. Vertical
shear force and bending moment in head waves show moderate
scatter between different calculations and agree with the measure
ments. In following waves, different computational results show
significant scatter in the entire frequency range; the tendency of
numerical results with increasing frequency differs from the meas
urements. Midship lateral shear force and bending moment, as
well as torsional moment, show moderate scatter between differ
ent computations in bow waves but significant scatter and
Journal of Offshore Mechanics and Arctic Engineering AUGUST 2014, Vol. 136 / 031602-1
Copyright © 2014 by ASME