ANNEX M SPECIAL DESIGN PROVISIONS FOR COLUMN STABILIZED UNITS
M.5 FATIGUE LIMIT STATES (FLS)
M.5.1 General
General requirements and guidance concerning fatigue criteria are given in Chapter 8, and Annex C.
Evaluation of the fatigue limit states shall include consideration of all significant actions contributing to fatigue damage both in non-operational and operational design conditions.
The minimum fatigue life of the unit (before the design fatigue factor is considered) should be based upon a period of time not being less than the planned life of the structure.
Assumptions related to the resistance parameters adopted in the fatigue design, e.g. with respect to corrosion protection, (see Annex C) shall be consistent with the in-service structure.
Local effects, for example due to :
• slamming,
• sloshing,
• vortex shedding,
• dynamic pressures, and,
• mooring and riser systems,
shall be included in the fatigue damage assessment when relevant.
In the assessment of fatigue resistance, relevant consideration shall be given to the effects of stress raisers (concentrations) including those occurring as a result of :
• fabrication tolerances. (Including due regard to tolerances in way of connections of large structural sections, e.g. as involved in mating sequences or section joints),
• cut-outs,
• details at connections of structural sections (e.g. cut-outs to facilitate construction welding).
M.5.2 Design fatigue factors
Criteria related to Design Fatigue Factors, are given in the Chapter 8.
Structural components and connections designed in accordance with this Annex should normally satisfy the requirement to damage condition according to the Accidental Limit State with failure in the actual joint defined as the damage. As such Design Fatigue Factors selected in accordance with Chapter 8, Table 8-1 should normally fall within the classification: "Without substantial
consequences".
When determining the appropriate Design Fatigue Factor for a specific fatigue sensitive location, consideration shall be given to the following :
• Consideration that economic consequence of failure may indicate the use of larger design factors than those provided for as minimum factors.
• The categorisation : ‘Accessible / Above splash zone’ is, intended to refer to fatigue sensitive locations where the possibility for close-up, detailed inspection in a dry and clean condition exists. If any of these requirements are not fulfilled, the relevant design fatigue factor should be considered as being that appropriate for ‘Accessible / Below splash zone’, or, ‘No access or in the splash zone’ as relevant to the location being considered.
• Evaluation of likely crack propagation paths (including direction and growth rate related to the inspection interval), may indicate the use of a different Design Fatigue Factor than that which would be selected when the detail is considered in isolation, such that ;
Annex M Rev. 1, December 1998
• Where the likely crack propagation indicates that a fatigue failure, from a location satisfying the requirements for a 'Non-substantial' consequence of failure, may result in a 'Substantial' consequence of failure, such fatigue sensitive location is itself to be deemed to have a 'Substantial' consequence of failure.
• Where the likely crack propagation is from a location satisfying the requirement for a given
‘Access for inspection and repair’ category to a structural element having another access categorisation, such location is itself to be deemed to have the same categorisation as the most demanding category when considering the most likely crack path. For example, a weld detail on the inside (dry space) of a submerged shell plate should be allocated the same Design Fatigue Factor as that relevant for a similar weld located externally on the plate –see Figure M.5-1.
Implications of the above are that for internal structures, below or above splash zone, the Design Fatigue Factor (DFF) shall normally be 1 where access for inspection and repair is possible. Non- accessible areas, like welds covered with fire proofing and other areas inaccessible for inspection, shall be designed with DFF=3.
The hull shell below splash zone shall normally be assigned a DFF=2. This applies also to any attachment, internal or external, from where a fatigue crack can grow into the shell.
Figure M.5-1 Example illustrating considerations relevant for selection of design fatigue factors (DFF’s) in locations considered to have ‘non-substantial’ consequence of failure.
M.5.3 Splash zone
The definition of ‘splash zone’ as given in NORSOK N-003, Section 6.6.3, relates to a highest and lowest tidal reference. For column stabilised units, for the evaluation of the fatigue limit state, reference to the tidal datum should be substituted by reference to the draught that is intended to be utilised when condition monitoring is to be undertaken. The requirement that the extent of the splash zone is to extend 5m above and 4m below this draught may then be applied. (For application of requirements to corrosion addition, however, the normal operating draught should generally be considered as the reference datum).
If significant adjustment in draught of the unit is possible in order to provide for satisfactory accessibility in respect to inspection, maintenance and repair, account may be taken of this possibility in the determination of the Design Fatigue Factors. In such cases however, a sufficient margin in respect to the minimum inspection draught should be considered when deciding upon the appropriate Design Fatigue Factor in relation to the criteria for ‘Below splash zone’ as opposed to
‘Above splash zone’. Where draft adjustment possibilities exist, a reduced extent of splash zone may be applicable. (See Commentary for further details).
The entire unit may be regarded as being above the splash zone if the unit is to be regularly dry- docked every 4-5 years.
M.5.4 Fatigue analysis M.5.4.1 General
The basis for determining the acceptability of fatigue resistance, with respect to wave actions, shall be appropriate stochastic fatigue analyses. The analyses shall be undertaken utilising relevant site specific environmental data and take appropriate consideration of both global and local (e.g.
pressure fluctuation) dynamic responses. (These responses do not necessarily have to be evaluated in the same model but the cumulative damage from all relevant effects should be considered when evaluating the total fatigue damage.)
Simplified fatigue analyses may form the basis of a ‘screening’ process to identify locations for which a detailed, stochastic fatigue analysis should be undertaken (e.g. at critical intersections).
Such simplified fatigue analysis shall be calibrated –see M.5.4.2.
Local, detailed FE-analysis of critical connections (e.g. pontoon/pontoon, pontoon/column, column/deck and brace connections) should be undertaken in order to identify local stress
distributions, appropriate SCF’s, and/or extrapolated stresses to be utilised in the fatigue evaluation –see Annex C2.8 for further details. Dynamic stress variations through the plate thickness shall be documented and considered in such evaluations.
Explicit account shall be taken of any local structural details that invalidate the general criteria utilised in the assessment of the fatigue strength. Such local details may, for example be access openings, cut-outs, penetrations etc. in structural elements.
Principal stresses (see C2.2) should be utilised in the evaluation of fatigue responses.
M.5.4.2 Simplified fatigue analysis
Provided that the provisions stated in M.5.4.1, are satisfied, simplified fatigue analysis may be undertaken in order to establish the general acceptability of fatigue resistance. In all cases when a simplified fatigue analysis is utilised a control of the results of the simplified fatigue evaluation, compared to the stochastic results, shall be documented to ensure that the simplified analysis provides for a conservative assessment for all parts of the structure being considered.
Annex M Rev. 1, December 1998 Simplified fatigue analysis is particularly well suited for preliminary design evaluation studies. In such cases, a space frame model of the structure may be considered as being adequate. For final design however the basis model for the ‘screening’ evaluation should be a model of similar (or better) detail to that described in M.4.2.2.
Simplified fatigue analyses should be undertaken utilising appropriate design parameters that provide for conservative results. Normally a two-parameter, Weibull distribution (see C2.9) may be utilised to describe the long-term stress range distribution. In such cases the Weibull shape
parameter (‘h’ –see Eqn.(M.5.1) below) should normally have a value between 1.0 - 1.1.
( )
( )
m 1
m 0 1 h 1 0 n
mh 1 Γ n
a (DFF)
Δσ lnn
0
= +
(M.5.1)
where :
n0 is the total number of stress variations during the lifetime of the structure
∆σn0 is the extreme stress range that is exceeded once out of n0stress variations. (The extreme stress amplitude Δσampl_n0 is thus given by 2
Δσn0
).
h is the shape parameter of the Weibull stress range distribution
a is the intercept of the design S-N curve with the log N axis (see, for example, C2.3.3)
(1 mh)
Γ + is the complete gamma function (see C2.9) m is the inverse slope of the S-N curve (see C2.9) DFF is the Design Fatigue Factor.
Generally, the simplified global fatigue analysis should consider the ‘F3’, SN class curve (see C2.3), adjusted to include any thickness effect, as the minimum basis requirement. Areas not satisfying this requirement are normally to be excluded from the simplified fatigue evaluation
‘screening’ procedure. (This implies that connections with a more demanding SN fatigue class than F3, are not to be applied in the structure, e.g. if overlap connections are applied then the fatigue SN class is to be suitably adjusted.) The cumulative fatigue damage may then be obtained by
considering the dynamic stress variation, ∆σActual(n0), which is exceeded once out of ‘n0’ cycles (see eqn. (M.5.2)) with the allowable equivalent stress variation, ∆σn0, calculated from equation (M.5.1). The fatigue life thus obtained is found from eqn. (M.5.1)).
m
=
) Actual(n
n Design
Actual
0 0
Δσ N Δσ
N
(M.5.2)
where :
Actual
N is the actual (calculated) fatigue life
Design
N is the target fatigue life
n0
Δσ is described in Eqn.(M.5.1)).
( )n0
Actual
Δσ is the actual design, dynamic stress variation which is exceeded once out of ‘n0’ cycles.
When the simplified fatigue evaluation involves utilisation of the dynamic stress responses resulting from the global analysis (as described in M.4.2), the response should be suitably scaled to the return period of the basis, minimum fatigue life of the unit. In such cases, scaling may normally be
undertaken utilising the appropriate factor found from eqn (M.5.3).
h n
n n
n
1
100 0
log log
100
0
∆
=
∆σ σ
(M.5.3)
where :
n100 is the number of stress variations (e.g. 100 years) appropriate to the global analysis
n100
σ
∆ is the extreme stress range that is exceeded once out of n100stress variations.
(Other parameters are as for eqns. (M.5.1) and (M.5.2)).
M.5.4.3 Stochastic fatigue analysis
Stochastic fatigue analyses shall be based upon recognised procedures and principles utilising relevant site specific data.
Analyses shall include consideration of the directional probability of the environmental data.
Providing that it can be satisfactorily documented, scatter diagram data may be considered as being directionally specific. Relevant wave spectra shall be utilised. Wave energy spreading may be taken into account if relevant.
Structural response shall be determined based upon analyses of an adequate number of wave directions. Generally a maximum radial spacing of 15 degrees should be considered. Transfer functions should be established based upon consideration of a sufficient number of periods, such that the number, and values of the periods analysed :
• adequately cover the site specific wave data,
• satisfactorily describe transfer functions at, and around, the wave ‘cancellation’ and
‘amplifying’ periods. (Consideration should be given to take account that such ‘cancellation’
and ‘amplifying’ periods may be different for different elements within the structure).
• satisfactorily describe transfer functions at, and around, the relevant excitation periods of the structure.
Stochastic fatigue analyses utilising simplified structural model representations of the unit (e.g. a space frame model) may form the basis of a ‘screening’ process to identify locations for which a, stochastic fatigue analysis, utilising a detailed model of the structure, should be undertaken (e.g. at critical intersections). When space frame, beam models are utilised in the assessment of the fatigue strength of large volume structures, (other than for preliminary design studies) the responses resulting from the beam models should be calibrated against a more detailed model to ensure that global stress concentrations are included in the evaluation process.
Annex M Rev. 1, December 1998