Rules For Classification and Construction V Analysis Techniques
Rules For Classification and Construction V Analysis Techniques
Rules For Classification and Construction V Analysis Techniques
V Analysis Techniques
1 Hull Structural Design Analyses
Edition 2004
The following Guidelines come into force on August 1st, 2004
Alterations to the preceding Edition are marked by beams at the text margin.
Head Office
Vorsetzen 35, 20459 Hamburg, Germany
Phone: +49 40 36149-0
Fax: +49 40 36149-200
[email protected]
www.gl-group.com
"General Terms and Conditions" of the respective latest edition will be applicable
(see Rules for Classification and Construction, I - Ship Technology, Part 0 - Classification and Surveys).
Table of Contents
Section 1 General
A. Objective of the Guidelines ........................................................................................................ 1- 1
B. Objective of a Fatigue Strength Analysis ................................................................................... 1- 1
C. Analysis Methods ....................................................................................................................... 1- 1
D. Stress Types ............................................................................................................................... 1- 2
Section 1
General
The proof of sufficient fatigue strength, i. e. the 1. The objective of a fatigue strength analysis is
strength against crack initiation under dynamic loads normally:
during operation is useful for judging and reducing
– the identification of structural areas with high
the probability of crack imitiation of structural mem-
cyclic stresses
bers during the design stage.
– the analysis and assessment of critical structural
Due to the randomness of the load process, the details
spreading of material properties and fabrication fac- – the determination of structural alternatives
tors and to effects of ageing, crack initiation cannot be
completely excluded during later operation. Therefore
among other things perodical surveys are necessary. 2. The identification of structural areas with
high cyclic stresses requires the calculation of stress
ranges and mean stresses in the areas under considera-
tion for one of the stress types mentioned in D. to-
gether with a suitable representation of the results that
A. Objective of the Guidelines allows the identification of critical details.
spectrum for the whole service life is determined with from relatively coarse finite element models. The
statistical methods for the subsequent fatigue strength effective breadth has to be considered. Further guid-
assessment. ance for the computation of nominal stresses by finite
element analysis is contained in Chapter 1 – Guide-
4. Contrary to the spectral method, the simula- lines for Strength Analyses of Ship Structures with the
tion of the stress history is performed in the time do- Finite Element Method.
main, making it possible to consider the direct calcula-
tion of nonlinear effects during the load and damage 3. Welded structures which cannot be related to
process. a classified detail or where additional stresses occur
which are not or not fully considered by the cata-
5. The selection of the method depends primar- logued detail can be assessed with respect to fatigue at
ily on the question as to whether acceptable results can weld toes on the basis of the structural or hot-spot
be expected in view of the respective simplifications stress σs, which contains the stress increase due to the
and assumptions. In the case of the simplified deter- structural geometry, but not that due to the weld toe.
ministic method, which is at present mostly applied In some cases, structural stress concentration factors
for practical reasons, the main question is whether the Ks can be used for the determination of the structural
selected load situations consider sufficiently the rele- or hot-spot stress σs,
vant stresses and whether the shape and number of
cycles of the stress spectrum can be estimated with
satisfactory accuracy on the basis of experience and/or σs
similar computations. When applying the spectral Ks =
σn
method, it is necessary to check how far simplifica-
tions such as the neglect or approximate consideration
of nonlinear effects can be justified. In the case of a which result from parametric investigations. Here, the
simulation of the stress history, the length of simula- definition of the nominal stress and the validity ranges
tion time which allows extrapolation to the whole of the parametric formulae or diagrams have to be
service life with sufficient accuracy is of primary considered. The fatigue strength assessment is per-
interest. formed using specific S-N curves or detail categories
for structural stresses, as given in the Construction
Rules for Seagoing Ships.
Section 2
1. The modelling of the structure has to be per- 1. The calculation of nominal stresses is nor-
formed in such a way that, at the detail considered, the mally performed by using beam models of the ship’s
type of stress on which the fatigue assessment is to be hull girder, of the primary structural members as well
based can be calculated with sufficient accuracy. as secondary structural members, or by using rela-
tively coarse finite element models to represent these
structures. The latter method is recommended espe-
cially for more complex structures and load cases.
2. Two methods are normally applied in the
calculation: 2. It is common to distinguish between global
stresses in primary structural members and additional
– analytical methods, e.g. based on beam theory, local stresses in secondary structural members
which usually result in nominal stresses (frames, girders, stiffeners), see Fig. 2.1. The latter
include bending stresses in stiffeners and plates. In
– numerical methods, such as the finite element fatigue analyses, the combined effect of both compo-
method, which are particularly well suited for nents is to be taken into account.
the stress analysis of geometrically complex
structures as well as local stress increases
3. The extent of the model is primarily deter-
mined by the question of whether and how reasonable
boundary conditions can be formulated for partial
3. In the following special information is given models of the structure. Fig. 2.2 shows a typical over-
about the modelling of a structure in view of calculat- all model of the ship structure with a relatively coarse
ing the desired type of stress, see Section 1, D. Further mesh. As a result, global nominal stresses are obtained
information is contained in Chapter 1 – Guidelines for in the primary structural members; these have to be
Strength Analyses of Ship Structures with the Finite superimposed with local stresses as well as increased
Element Method. stresses due to reduced effective breadths.
Neutral
Axis
A B C
a)
b)
Type a)
Fig. 2.2 Overall model of a ship's structure a)
ally, plate or shell elements having improved in-plane 5. For both types of modelling, the dimensions
deformation behaviour should be chosen. of the first two elements in front of the weld toe
should be chosen as follows:
If solid elements are used, the weld can be modelled
simply with prismatic elements, see Fig. 2.5. With – Weld toes of type a):
isoparametric 20-node elements, one element in plate The element length should correspond to the
thickness direction is sufficient due to the quadratic plate thickness t. In the transverse direction, the
shape function and the linear stress distribution. In plate thickness t may be chosen as well for the
connection with reduced integration order, the stresses breath of the plate elements. However, the total
can be extrapolated directly to the element surface and breath of the two shell elements or the solid
then to the weld toe. element in front of the weld toe should not ex-
ceed the attachment width w, where w is the
shell elements thickness of the attached plate above or below
(without welds) plus 2 times the weld leg length, see 2.5.
– Weld toes of type b):
At plate edges, the element length and depth
should be 10 mm. Due to the steep stress in-
crease in these areas, generally higher-order
elements with mid-side nodes or shell elements
with improved in-plane deformation behaviour
have to be used.
Section 3
– assumption of the loading condition that is most 3. Generally, load combinations (cf. Fig. 3.1)
unfavourable for the fatigue strength of the de- with
tail considered, i.e. with respect to the mean
stress and the stress ranges – waves from ahead/astern (max. vertical bending
moments in the hull girder, correspondingly in-
– assumption of several representative loading creased or reduced side and bottom pressure in
conditions which on the one hand are typical for the forward, midship and aft part of the hull
the ship being considered and on the other in- girder), and
clude different cargo and ballast distributions – oblique head/stern waves (reduced vertical
and also different draughts. Frequently, 4 to 12 bending moments in the hull girder, max. hori-
loading conditions are sufficient, if they zontal bending and torsional moments, max.
transverse acceleration or tank pressures, with
– cause high hogging and sagging bending corresponding opposing pressure from outside)
moments in the hull girder,
should be considered. In the latter case, the considera-
tion of the upright and heeled ship position might be
– include the ballast and fully-laden condi-
necessary, depending on the structural detail under
tions, as well as an intermediate condition,
consideration. Unfavourable load combination may
if appropriate,
occur at port or starboard side.
– include cargo distributions which cause
high stresses in the bottom and in the bulk- 4. After calculating the highest stress range
heads. Δσmax and the associated mean stress σm for each
loading condition, further characteristics of the stress
Regarding consumables, consideration of the condi- range spectrum are estimated on the basis of the Con-
tion "50 % consumables" is normally sufficient. struction Rules for Seagoing Ships. For wave-induced
stresses, usually a straight-line spectrum in semi-
In extensive fatigue strength analyses for different logarithmic representation and a total number of load
structural areas, various loading conditions normally cycles nmax = 5 ⋅ 107 during the service life of about
have to be analysed, because unfavourable conditions 20 years are assumed, see also I – Ship Technology,
can result from different loading conditions for differ- Part 1 – Seagoing Ships, Chapter 1 – Hull Structures,
ent details. Section 20, A.2.4.
Chapter 2 Section 3 B Definition of Load Cases and Calculation of the Stress Spectrum V - Part 1
Page 3–2 GL 2004
1. Head Seas
pD
(+) (-)
az = 0 az = 0
T
z z
(-) (+)
max ps min ps
(+) (-)
z z
(-) (+)
(+) (-)
az = 0 az = 0
(+) ay > 0 (-) (-) ay < 0 (+)
j > 0 (for tanks) j < 0 (for tanks)
z z
(-) j y (+) j y
Remarks:
- In cases 1. and 2., az ¹ 0 may be considered in connection with mean external pressure.
- The signs (+) and (-) apply to the longitudinal stresses due to vertical and horizontal wave bending of the hull
girder acc. to Construction Rules for Seagoing Ships.
- Torsion related stresses are to be included for ships with large deck openings.
Fig. 3.1 Typical Load Situations for the Calculation of the Highest Stress Range
V - Part 1 Section 3 D Definition of Load Cases and Calculation of the Stress Spectrum Chapter 2
GL 2004 Page 3–3
2. The wave load cases in the first step (see 1.1) 5. It is recommended that the long-term distri-
should be calculated with a wave amplitude which lies bution of stress ranges at the structural detail consid-
within the range of the highest damage due to the ered be calculated separately for all different loading
number of load cycles (approx. 1 m). The ship’s speed conditions, because they are assessed with different
is usually assumed to be 2/3 of the service speed if no mean stresses. For each loading condition, the entire
service profile is specified. The wave lengths should service life of 20 years should be assumed in a first
be chosen such that the pressure fluctuations due to step. In this way, the results can be directly compared
shorter waves as well as the roll, heave and pitch reso- and it is possible to assess them for various combina-
nances of the ship are covered, and such that enough tions of loading conditions as well as for the most
values are available to allow for a sufficiently accurate critical loading condition.
representation of the transfer function between wave
amplitude and structural response (normally 15 – 25
wavelengths). The angle of encounter should be varied
in steps of 30 degrees, whereby waves running trans- D. Simulation of the Stress History
versely to the ship’s course usually require a special
consideration. If only angles of encounter from one 1. Different possibilities exist for the simulation
side of the ship are considered, the results for the of the stress history, inter alia:
corresponding detail on the other ship’s side have to
be included in the evaluation. For each wave situation, a) Simulation of the stress history on the basis of
two cases with a phase difference are considered in computed stress range spectra for the individual
order to obtain real and imaginary part of the har- sea states
monic load process.
b) Simulation of the complete motion and load
process for the whole ship structure.
3. In the second step (see 1.2), the stress spectra
are calculated for all sea states considered by the long- In any case, different sea states and angles of encoun-
term statistics, with all loading conditions and angles ter are considered for all loading conditions investi-
of encounter, by multiplying the quadratic transfer gated, from which the long-term distribution of stress
function with the individual wave spectra. For the ranges is evaluated in a similar way as that described
wave spectra, a two-parameter standard distribution in C.4.
(e.g. according to Pierson-Moskovitz) can be assumed.
The frequency distribution of the stress ranges results 2. Method a) is based on calculations of the
approximately from the spectral moments and the structural response, as described in C.2. – C.3. in con-
assumption of a narrow-banded Gaussian process nection with the spectral method, so that the prerequi-
(Rayleigh distribution of the maxima). A correction sites are the same. The advantage of simulation lies in
for the wide bandedness of the loading process can be the computation of the course of the stress history,
performed on the basis of published methods, whilst which allows a refined assessment of the damage
its neglect generally leads to slightly conservative process, e.g. by considering sequential effects or mate-
results. rial-related nonlinearities in corresponding concepts.
Chapter 2 Section 3 D Definition of Load Cases and Calculation of the Stress Spectrum V - Part 1
Page 3–4 GL 2004
3. In addition to that, method b) permits realistic such that significant statistical evaluations can be
consideration of nonlinearities in the load process, made with respect to the entire service life.
which may be significant for certain ship types.
5. From the stress history evaluated, the relevant
parameters for the assessment (e.g. largest stress range
Δσmax, shape of stress spectrum and associated mean
4. The loading conditions and sea states consid- stress σm) are to be determined using the rainflow
ered as well as the simulation time are to be chosen counting method.
GL 2004
V - Part 1
hs \ tz 1.5 2.5 3.5 4.5 5.5 6.5 7.5 8.5 9.5 10.5 11.5 12.5 13.5 14.5 15.5 16.5 17.5 18.5 SUM
Section 3
0.5 0.0 0.0 1.3 133.7 865.6 1186.0 634.2 186.3 36.9 5.6 0.7 0.1 0.0 0.0 0.0 0.0 0.0 0.0 3050
1.5 0.0 0.0 0.0 29.3 986.0 4976.0 7738.0 5569.7 2375.7 703.5 160.7 30.5 5.1 0.8 0.1 0.0 0.0 0.0 22575
D
2.5 0.0 0.0 0.0 2.2 197.5 2158.8 6230.0 7449.5 4860.4 2066.0 644.5 160.2 33.7 6.3 1.1 0.2 0.0 0.0 23810
Table 3.1
3.5 0.0 0.0 0.0 0.2 34.9 695.5 3226.5 5675.0 5099.1 2838.0 1114.1 337.7 84.3 18.2 3.5 0.6 0.1 0.0 19128
4.5 0.0 0.0 0.0 0.0 6.0 196.1 1354.3 3288.5 3857.5 2685.5 1275.2 455.1 130.9 31.9 6.9 1.3 0.2 0.0 13289
5.5 0.0 0.0 0.0 0.0 1.0 51.0 498.4 1602.9 2372.7 2008.3 1126.0 463.6 150.9 41.0 9.7 2.1 0.4 0.1 8328
6.5 0.0 0.0 0.0 0.0 0.0 12.6 167.0 690.3 1257.9 1268.6 825.9 386.8 140.8 42.2 10.9 2.5 0.5 0.1 4806
7.5 0.0 0.0 0.0 0.0 0.0 3.0 52.1 270.1 594.4 703.2 524.9 276.7 111.7 36.7 10.2 2.5 0.6 0.1 2586
8.5 0.0 0.0 0.0 0.0 0.0 0.7 15.4 97.9 255.9 350.6 296.9 174.6 77.6 27.7 8.4 2.2 0.5 0.1 1309
9.5 0.0 0.0 0.0 0.0 0.0 0.2 4.3 33.2 101.9 159.9 152.2 99.2 48.3 17.7 6.1 1.7 0.4 0.1 626
10.5 0.0 0.0 0.0 0.0 0.0 0.0 1.2 10.7 37.9 67.5 71.7 51.5 27.3 11.4 4.0 1.2 0.3 0.1 285
11.5 0.0 0.0 0.0 0.0 0.0 0.0 0.1 3.3 13.3 26.6 31.4 24.7 14.2 6.4 2.4 0.7 0.2 0.1 124
12.5 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 4.4 9.9 12.8 11.0 6.8 3.3 1.3 0.4 0.1 0.0 51
13.5 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.3 1.4 3.5 5.0 4.6 3.1 1.6 0.7 0.2 0.1 0.0 21
14.5 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.1 0.4 1.2 1.8 1.8 1.3 0.7 0.3 0.1 0.0 0.0 8
15.5 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.1 0.4 0.6 0.7 0.5 0.3 0.1 0.1 0.0 0.0 3
Wave Scatter Diagram for the North Atlantic
16.5 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.1 0.2 0.2 0.2 0.1 0.1 0.0 0.0 0.0 1
SUM: 0 0 1 165 2091 9280 19922 24879 20870 12898 6245 2479 827 247 66 16 3 1 100000
Section 4
Alternatively, the additional load cycles due to chang- 5. It is recommended that the results be assessed
ing loading condition may be included in the damage on the basis of the usage factor U with respect to stress
calculation, by taking the rainflow counting method ranges:
into account and making conservative assumptions for
the probability of occurrence of the individual loading
conditions. Δσmax
U =
Δσp
4. For standardized stress spectra, the criterion
for the assessment is When applying Palmgren-Miner's Rule, Δσp results
Δσ max ≤ Δσp from the condition that with an equivalent spectrum of
stress ranges, i.e. having the same shape and number
of load cycles, the limit damage ratio D* according to
Δσmax = highest stress range within the spectrum B.2. is reached. The application of the usage factor has
the advantage that - due to the relationship with
Δσp = permissible stress range
stresses - direct conclusions can be drawn with respect
The permissible stress range results directly from the to the necessary reduction of stresses or a possible
Construction Rules for Seagoing Ships, considering increase of the detail category. For example, in case of
the appropriate detail category or design S-N curve, as U = 1,2, the stress has to be reduced by 20 % or the
well as various other factors, such as the mean stress. detail category increased by 20 %.