Jack-up

Guidelines for Offshore Structural Reliability -DNV Application to Jackup Structures

Page No. 1

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

LIST OF CONTENTS Section

Title

Page

1.0 1.1 1.2 1.3 1.4 1.5

INTRODUCTION Objective Jack-ups in General Modes of Operation Important Structural Design Parameters Arrangement of Report

3 3 3 3 4 6

2.0 2.1 2.2 2.3 2.3.1 2.3.2 2.3.3 2.3.4 2.3.5 2.3.6

RESPONSE General Jack-up Response in the Floating Mode Jack-up Response in the Elevated Mode of Operation Time Domain Analysis Methods of Evaluating Response Static Load Components Sea Loadings 14 Wind Loadings Foundations

7 7 7 10 11 12 14

3.0 3.1 3.2 3.2.1 3.2.2 3.3 3.3.1 3.3.2 3.3.3 3.4

UNCERTAINTY MODELLING General Loading Uncertainty Modelling Aleatory Uncertainty Epistemic Uncertainty Response Uncertainty Modelling Analysis Uncertainty Damping Foundation Resistance Uncertainty Modelling

19 19 19 19 20 21

4.0 4.1 4.1.1 4.2 4.2.1 4.2.2 4.2.3 4.2.4 4.2.5 4.2.6 4.3

LIMIT STATES General Limit States Appropriate to Jack-up Structures The Ultimate Limit State Leg Strength Foundation Bearing Failure Holding System Global Deflections Global Leg Buckling Overturning Stability Literature Study

25 25 25 27 27 30 30 32 32 32 33

5.0 5.1 5.2 5.3 5.4

SUMMARY OF APPLICATION EXAMPLES General Overview of Analytical Procedure Structural Reliability Example Foundation Reliability Example

34 34 34 36 38

15 16

21 21 22 24


Page No. 2

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

Section

Title

Page

6.0 6.1 6.2 6.3

RECOMMENDATIONS FOR FURTHER WORK General Elevated Condition Floating / Installation Phase Conditions

41 41 41 42

7.0

REFERENCES

44


Page No. 2

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

Section

Title

Page

6.0 6.1 6.2 6.3

RECOMMENDATIONS FOR FURTHER WORK General Elevated Condition Floating / Installation Phase Conditions

41 41 41 42

7.0

REFERENCES

44


Page No. 3

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

1.0

INTRODUCTION

1.1

Objective

The objective of this report is to document offshore structural reliability guidelines appropriate to self-elevating unit structures (hereafter referred to as ‘jack-ups’). With this intention the following items are addressed ; - characteristic responses - modes of failure and related reliability analysis characteristics and parameters - typical examples of reliability analysis. The guidelines are intended for application of Level III structural reliability where the joint probability distribution of uncertain parameters is used to compute a probability of failure.

1.2

Jack-ups in General

The term ‘Jack-up’ covers a large variety of offshore structures from small liftboat structures, Stewart (1991), to large deepwater designs, e.g. Bærheim (1993). The purpose of the jack-up design is to provide a mobile, self-installing, stable working platform at an offshore (or offland) location. The jack-up platform itself may be designed to serve any function such as, for example ; tender assist, accommodation, drilling or production. Thus, the term jack-up may represent a structure that has a mass of a few hundred tonnes and is capable of elevating not more than a few metres above the still water surface, to a structure that has a mass of over 20,000 tonnes and is capable of operating in water depths in excess of 100 metres. ·

It is evident, for the above stated reasons, that statistics representing jack-up structures should be treated with a good deal of suspicion as they may not be representative for the type of structure required to be considered.

·

These guidelines are intended to deal primarily with conventional design, larger size jack-ups, namely those intended to operate in waterdepths in excess of, say, 50 metres. A typical arrangement of such a unit is shown in Figure 1.1 below, Bærheim (1993).

1.3

Modes of of Op Operation

A jack-up generally arrives on location in the self-floating mode. The transportation of the jack-up to the site may, however, have been undertaken as a wet, or dry (piggy-back) tow, or, may have been undertaken by the use of self-propulsion. Once on location installation will take place, which will typically involve elevating the hull structure to a predetermined height above the water surface, preloading, and then elevating to an operational height. Characteristically the jack-up will then remain on location for a period of 2-4 months, before jacking down, raising the legs to the transit mode condition, and transferring to the next location. ·

This short-term contracting of jack-up units has historically resulted in that, within its life cycle, the jack-up rarely operates to its maximum design environmental criteria.

·

There is a current tendency to design jack-up units for extended period operation at specific sites, Bærheim (1993), Scot Kobus (1989), e.g. as work-over or production units.


Page No. 4

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

Such units may been designed to operate in extreme environmental conditions, at relatively large waterdepths for a period in excess of 20 years.

Figure Figure 1.1

1.4 1.4

: Arrang Arrangeme ement nt of a Typical Typical Harsh Harsh Enviro Environme nment nt Jack-u Jack-up p

Impo Import rtan antt Str Struc uctu tura rall Des Desig ign n Par Param amet eter ers s

Jack-up designs varying from being monotower structures (single leg designs) to multiple leg designs, e.g. up to six legs, although units with sixteen legs are not unknown, Boswell (1986). The supporting leg structures may be a framework design, or, may be plate profile design. ·

The conventional jack-up design has three vertical legs, each leg normally being constructed of a triangular or square framework.

Jack-up basic design involves numerous choices and variables. Typically the most important variables may be listed as stated below. Support Footing The legs of a jack-up are connected to structure necessary to transfer the loadings from the leg to the seafloor. This structure normally has the intended purpose to provide vertical


Page No. 5

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

support and moment restraint at the base of the legs. The structural arrangement of such footing may take the following listed forms; -gravity based (steel or concrete), -piled -continuous foundation support, e.g. mat foundations -individual leg footings, e.g. spudcans (with or without skirts). Legs The legs of a jack-up unit are normally vertical, however, slant leg designs also exist. Design variables for jack-up legs may involve the following listed considerations ; -number of legs -global orientation and positioning of the legs -frame structure or plate structure -cross section shape and properties -number of chords per leg -configuration of bracings -cross-sectional shape of chords -unopposed, or opposed pinion racks -type of nodes (e.g. welded or non-welded (e.g. forged) nodes) -choice of grade of material, i.e. utilisation of extra high strength steel Method of transferring loading from (and to) the deckbox to the legs The method of transferring the loadings from (and to) the deckbox to the legs is critical to design of the jack-up. Typical design are ; -utilisation and design of guides (e.g. with respect to ; number, positioning, flexibility, supporting length and plane(s), gaps, etc.) -utilisation of braking system in gearing units -support of braking units (e.g. fixed or floating systems) -utilisation of chocking systems -utilisation of holding and jacking pins and the support afforded by such. Deckbox The deckbox is normally designed from stiffened panel elements. The shape of the deck structure may vary considerably from being triangular in basic format to rectangular and even octagonal. The corners of the deckbox may be square or they may be rounded. Units intended for drilling are normally provided with a cantilever at the aft end of the deckbox, however, even this solution is not without exception and units with drilling derricks positioned in the middle of the deckbox structure are not unknown. There are a large number of solutions available to the designer of a jack-up unit and, although series units have been built, there exist today an extremely large number of unique jack-up designs.


Page No. 6

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

1.5

Arrangement of Report

Response of jack-up structures is described in Section 2, together with relevant methods for computation of the resulting load effects. Model uncertainties associated with the computation of these load effects are discussed in Section 3. Important limit states together with stochastic modelling of failure modes are described in Section 4. Section 5 provides a summary of two example reliability analyses undertaken for the ultimate limit state, DNV (1996b). Recommendations for further work are given in Section 6. Note : This report should be read in conjunction with the following listed documentation ; - “Guideline for Offshore Structural Reliability Analysis -General”, DNV Technical Report no.95-2018, DNV (1996a) - “Guideline for Offshore Structural Reliability Analysis- Examples for Jack-ups”, DNV Technical Report no.95-0072, DNV (1996b) Companion application guidelines are also documented covering for jacket and TLP structures.


Page No. 7

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

2.0

RESPONSE

2.1

General

Jack-up units are normally designed to function in several different operational modes. These modes may be characterised as follows ; -transit -installation -retrieval -operational (including survival) condition. Response of a jack-up in the floating mode of operation is, obviously, far different from that of the jack-up in the as-installed, elevated condition. Both of these modes are critical to the safe operation of a jack-up unit as each mode of operation may impose its own limiting design criteria on certain parts of the structure. To provide relevant guidance with respect to the stochastic properties and probabilistic analytical procedures for both of these modes of operation, is considered to be too large an undertaking to be handled by this example guidance note. ·

This section is therefore mainly concerned with jack-ups in the elevated mode of operation whilst it deals only in general terms with jack-ups in the floating mode.

2.2

Jack-up Response in the Floating Mode

A jack-up unit may transfer from one location to another by a number of methods. For ‘field’ moves a jack-up would, normally, transfer in the self-floating mode utilising either its own propulsion system, or, be ‘wet’ towed to the new location. For ‘ocean’ tows, on the other hand, it is common practice to transfer by means of a dry-tow. Three major sources of accident have been identified in respect to a jack-up in the transit condition, Standing and Rowe (1993), namely those due to; -1Wave damage to the unit structure leading to penetration of watertight boundaries. -2Damage to the structure as a result of shifting cargo (usually caused by direct wave impact, excessive motions and/or inadequate seafastenings). -3Structural damage in the vicinity of the leg support structures. In the jack-up installation phase there are normally two main areas of concern, these being ; -1Impact loadings upon contact with the seabed. -2Foundation failure (i.e. punch-through) during preloading. Impact loadings occur when the jack-up unit is operating in the floating mode, whilst foundation failure is a condition occurring when the jack-up is normally elevated above the still water surface. The retrieval phase of a jack-up has not traditionally been considered as providing dimensioning load conditions. However, when a leg is held fast at the seabed, e.g. due to large penetrations, there may be large loadings imposed upon the jack-up structure. Such loadings may result from the action of waves, current, wind, deballasting and jacking up loadings. Few model tests, or full-scale measurements, have been undertaken for jack-ups in the floating mode. Indeed, recent record searches and enquiries with model basins to establish


Page No. 8

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

relevant model test data, Standing and Rowe (1993), have only been able to identify six relevant model tests in total, with published papers on only two of these cases, Fernandes (1985, 1986). These experiments include free decay tests to provide estimates of damping and natural periods, measurements in heave, roll and pitch motions in regular and irregular waves at zero speed, and measurements of resistance, heave, roll and pitch in regular and irregular waves at 6 knots tow speed. A number of the tests were repeated with the legs raised or lowered various distances. Some full scale results were also published. Comparisons with linear wave theory, based upon potential flow assumptions, predict roll and pitch responses in regular wave sea states very well at frequencies away from resonance, but may tend to overpredict the responses at the natural period (dependent upon damping assumptions). The results from the published jack-up model test data seem to be consistent with findings from ships and barges, i.e. that roll response at resonance is overestimated unless due account is taken of the increased damping resulting from viscous effects. Generally, levels of measured and predicted heave motions in regular waves agreed reasonably well although there may be marked differences in the shapes of the curves. Measurements in regular waves at 6 knots showed a considerable increase in the pitch damping, compared with similar results at zero speed, with reduced response at the natural period. Heave response was similar to that at zero speed. ·

Conventional wave diffraction theory will, in general, predict motion responses of a jackup unit with a reasonable degree of accuracy. If non-linear loading effects e.g. water on deck (‘green seas’), slamming, damping (especially at and around resonance periods), non-zero transit speed etc. are significant, then it is necessary to utilise time-domain simulation and/or model test data.

·

The use of strip theory or Morison formulation to compute the total sea loadings on a jack-up in transit will normally be inappropriate.

·

In connection with the prediction of motion responses, notwithstanding account taken of relevant non-linear loading effects, it seems reasonable to refer to ship or barge related reliability data (e.g. Frieze (1991), Lotsberg (1991), Wang and Moan (1993)).

·

When evaluating leg strength at critical connections, transfer functions for element forces and moments (or stresses) may be calculated directly from the rig’s motions analysis. A model similar to that shown in Figure 2.1 may, typically, be utilised for such purpose.

·

Generally, the following loads will be necessary to consider in respect to any ultimate strength analysis of a jack-up in the transit condition ; -static load components -inertia load components (as a result of motion) -wind load components.

·

If any significant structural non-linearities are present in the system then such nonlinearities should be accounted for in the model. One such non-linearity that may be significant is the modelling of any gaps between jackhouse guides and chords.

·

Reliability analysis of seafastening arrangements is documented, DNV (1992). The generalities of this documented example and the procedure utilised may also be applied to seafastenings for a jack-up unit under transit. If direct wave impact on the item held by the seafastening is a possible designing load, then such loading and associated load uncertainty should additionally be included within the analysis.


Page No. 9

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

Figure 2.1

: Typical Hydrodynamic/Structural Model of a Jack-up in the Transit Condition.


Page No. 10

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

2.3

Jack-up Response in the Elevated Mode of Operation

Response of jack-up structures in the elevated condition has previously been extensively studied, Ahilan (1993), with relevant analytical methodology being described in detail in the Jack-up Recommended Practice, SNAME (1993). The response of jack-up structures, when subjected to random sea excitation, is found to be non-Guassian in nature. Due to the non-linearities in the structural system the extreme responses are generally found to be larger than the extremes of a corresponding Gaussian process, Karunakaran (1993). Relevant, non-linear effects that may be significant in respect to response of jack-up structures are given as ; non-linear loading components (e.g. drag force loadings) bottom restraint (non-linear foundation characteristics) damping (e.g. due to the motions of the jack-up structure, there may be significant hydrodynamic damping as a result of the relative velocity of the water particles and the leg member) dynamics of the structure (as the natural period of the structure is typically relatively high, e.g. 5-8 seconds, there may be significant wave energy available to excite the structural system and hence relatively large inertial forces may result) second order effects (such effects may significantly influence the response in the considered structure) non-linearites of structural interfaces (e.g. gaps between the leg structure and guides) ·

For reliability analysis, in order to account for the non-linearities in jack-up loading and response, it is considered necessary that explicit time domain analysis, utilising stochastic sea simulation, is undertaken.

·

Foundation modelling assumptions have been shown to be an important aspect in respect to the resulting response from analytical models of jack-up units, Manuel et al. (1993). Hence, unless it can be demonstrated that the effects are not significant, non-linear characteristics in the foundation system should be explicitly modelled when undertaking analyses in connection with reliability studies.

·

Guidance provided in the guideline example for jacket structures, DNV (1996c), in respect to the fatigue limit state covers the state-of-the-art knowledge with respect to fatigue reliability analysis. Response in respect to the fatigue limit state is therefore not explicitly covered in this section. Due to the non-linear characteristics of jack-up loading and response, frequency domain solution techniques are however not recommended unless, either it can be demonstrated that such effects are insignificant, or, due account has been taken of such effects.


Page No. 11

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

2.3.1

Time Domain Analysis

Two general methods may be utilised in time domain analysis. These two methods being ; -use of simple, single degree of freedom (SDOF) models, and, -use of multi-degree of freedom models. In both cases however the following general guidance may be given for the analysis, SNAME (1993) ; 1.

The generated random sea should consist of superposition of, at least, 200 regular wave components utilising divisions of equal energy of the wave spectrum.

2.

In order to obtain sufficiently stable response statistics, simulation time for a single simulation should generally not be less than 60 minutes.

3.

The integration time step should not normally be taken greater than the smaller of the following ; - one twentieth of the zero up-crossing period of the wave spectrum - one twentieth of the jack-up natural period.

4.

When evaluating the response of the jack-up, the transient effects at the start of the analysis should be removed. At least the smallest of 100 seconds, or 200 time steps should be removed in this connection.

5.

The method of evaluating the response (e.g. the Most Probable Maximum (MPM) response) should be compatible with the simulation time and sea qualification procedure adopted for the analysis. -Further guidance in connection with this item is provided in the Commentaries to the Jack-up Recommended Practice, SNAME (1993).

The asymmetry of crest heights and troughs, accounted for by higher order wave theories, is not reproduced in methods based upon random wave simulation techniques. Linear wave theory, Sarpkaya (1981), utilised in random wave simulation, accounts for particle kinematics upto the still water surface and ‘kinematic stretching’ is undertaken to compute the kinematics to the instantaneous free surface. It is recommended, Gudmestad and Karunakaran (1994), that Wheeler stretching, Wheeler (1969), is utilised in this connection. The extent of wave asymmetry is a function of waterdepth. For waterdepths less than 25 metres, in extreme environmental conditions, irregular wave simulation is normally considered to be inappropriate and regular wave analysis should be considered. For waterdepths greater than 25 meters wave asymmetry may be accounted for by the formulation given in equation 2.1 below, SNAME (1993). Hs = ( 1 + 0.5 e

(-d/25)

) Hsrp

Where : Hs : adjusted significant wave height to account for wave kinematics (metres) Hsrp : significant wave height (metres) d : waterdepth (metres)

(2.1)


Page No. 12

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

As time domain analyses are usually fairly resource demanding procedures, it is normal practice to utilise simplified structural modelling techniques (see Figure 2.2) ·

A full description of the methodology and procedure utilised in creating both a simplified hydrodynamic and simplified structural model for a jack-up is included in DNV( Feb 1992) and SNAME (1993).

Figure 2.2 2.3.2 ·

: Typical Simplified Model of a Jack-up Structure.

Methods of Evaluating Response

Reliability analysis of jack-up structures will generally be undertaken based upon the following considerations ; -1- Site specific environmental and foundational data should be utilised. -2- Directional and seasonal data may be utilised. In order to reduce the amount of analytical work involved, wind, wave and current load components may however normally be assumed to be coincident. -3- The selected (governing) environmental load direction may be initially identified by evaluation of relevant deterministic, ‘quasi-static’ response analyses of the jackup structure under consideration. The standard procedure of treating wind, waves, currents and seawater level separately and combining the independent extremes as if these extremes occur simultaneously, is conservative. In most cases however, jackup environmental loading is wave dominated and the assumption of simultaneity of the extremes of the environmental parameters is found to be satisfactory.


Page No. 13

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

The probability of failure is estimated during a reference period significantly longer than the analysed, simulated time period. An extrapolation procedure for determining the extreme values for the reliability analysis is therefore required when several environmental variables are to be combined. ·

The reference period for extreme environmental data is normally selected as being equal to the one year return period such that the results may be directly compared with annual target reliabilities.

·

For jack-ups, the two most appropriate procedures for estimation of extreme load events would seem to be ; -1By use of long term statistics of independent sea states -2By use of conditional extreme event analysis.

These procedures are described in detail in Chapter 6 to the guidelines, DNV (1996a). For conventional jack-up structures, in general, the long term response is controlled by the extreme sea states and, as such, both of these procedures are normally acceptable. An example of the estimation of extreme load events by use of long term statistics of independent sea states is provided in the jack-up examples guidelines DNV (1996b). Karunakaran (1993) documents that the short term extreme storm response is marginally higher than the long term response if the long term response is controlled by extreme sea states. If however the long term response is controlled by resonance sea states, the short term extreme storm response is about 10% lower than the long term response for those case studies considered. Response from time history simulations may be characterised by the normalised statistical moments ; mx, sx, sx’, g3, g4, which are the mean, standard deviation, standard deviation of the time derivative, skewness and kurtosis of the response respectively. A limit state may then be defined from the statistical moments of the response and the estimated reliability thus obtained by the resulting response surface, DNV (1996b). ·

Response surface techniques are considered to provide the most appropriate methodology in the estimation of the reliability of jack-up structures for extreme load events.

In order to model how the statistical moments change with realisations of the basic variables, the derivatives of these moments may be estimated by finite differences of the variables at one estimation point. As the limit state functions are highly non-linear this technique will only give satisfactory results if a good fit is obtained around the design point. Generally, reliability analyses of jack-up structures may be undertaken by use of first and second order solution methods (FORM/SORM), Madsen (1986). -See also DNV (1996a), Chapters 2 and 3, for further guidance concerning utilisation of reliability methods.


Page No. 14

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

2.3.3

Static Loading Components

Previous jack-up reliability analyses, Karunakaran (1993), Løseth et al. (1990), have identified that response uncertainty is not significantly affected by the choice of the static mass model. This is further demonstrated in the example documented in DNV (1996b). ·

Permanent loads and variable loads are generally lumped together. For structural assessment the upper bound of this sum is normally conservatively modelled. For overturning assessment the mean variable load is combined with the permanent load.

2.3.4

Sea Loadings

Sea loadings on conventional jack-up structures are calculated utilising Morison’s equation, Sarpkaya (1981) ;

Fn ( r, t ) = r

p D

4

2

Cma n ( r, t ) +

1 2

r DCd v n ( r , t ) v n ( r , t )

(2.2)

Wave and current velocity components in the Morison equation are obtained by combining the vectorial sum of the wave particle velocity and the current velocity normal to the member axis. (When relative motions are involved, eqn 2.2 may be modified to reflect such motions in the terms an(r,t) and vn(r,t)). Epistemic uncertainties related to Morison’s equation are documented in Section 3. Wave Loadings The basic stochastic sea description is defined by use of a wave energy spectrum. The choice of the analytical wave spectrum and associated spectral parameters should reflect the width and shape of the spectra and significant wave height for the site being considered. Generally, either the Pierson-Moskowitz or the Jonswap spectra will be appropriate. See DNV (1996a), Section 5. ·

Due to the possibility of inducing greater dynamic response at lower wave periods than that necessarily associated with storm maximum significant wave height, a range of periods and associated significant wave heights should normally be investigated.

·

The simulated storm length is normally to be taken as 3 hours, SNAME (1993) or 6 hours, NPD (1992).


Page No. 15

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

For the extreme load event it is normally, conservatively assumed that a long crested sea simulation is undertaken, NPD (1992), however, in accordance with SNAME (1993) the following directionality function F(a) may be utilised ; F(a) = C. cos2na

for -p/2 £ a £ p/2

(2.3)

where ; n : 2.0 for fatigue analysis 4.0 for extreme analysis C

: constant chosen such that :

å-

/2

p

p

F (a )d a = 10 .

/2

Current Loadings ·

Current velocity should include all relevant components, DNV (1996). Normally, however, it is acceptable to divide the total current into two components, namely, that of wind and wave generated current, V(w,w) and that of residual (e.g. tidal) current, Vr . The first of these two current components may be assumed to be fully correlated with the significant wave height, whilst the latter current component, Vr , is assumed to be completely independent of the other environmental characteristics. See DNV (1996a), Section 5.1.3.2, for a full description of this procedure.

Unless site specific data indicate otherwise the current profile should be described according to the procedure documented in SNAME (1993). 2.3.5

Wind Loadings

Singh (1989) has found a number of inconsistencies in existing wind loading calculation procedures. Based upon this finding it has been concluded that wind tunnel measurements appear to provide the only viable method for accurately estimating loads on complex offshore structures. ·

For jack-up structures, if it is not possible to utilise model test data, either by direct testing, or from scaling of geosim models, then, assuming that wave loading is the dominating load effect, it is normally acceptable to base such loading on simplified, direct calculation methods.

SNAME (1993) documents an acceptable procedure for the calculation of wind loadings, where the wind loading, Fwi , is calculated as a static load contribution by use of the equation ; Fwi = ½ r Vref ² Ch Cs Aw

(2.4)

where r

Vref Ch Cs Aw

: density of air : the 1 minute sustained wind velocity at 10 meters above sea level : height coefficient : shape coefficient : projected area of the block considered

In locations where wind loading may be the dominating load effect (e.g. due to cyclones etc.) this load effect should be specially considered.


Page No. 16

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

2.3.6

Foundations

The uncertainty in jack-up response is greatly influenced by the uncertainties in the soil characteristics that determine the resistance of the foundation to the forces imposed by the jack-up structure. Ronold (1990) showed that, for a jack-up, the total uncertainty governing the safety against foundation failure is dominated by the uncertainty in the loading. Nadim et al. (1994), on the other hand, showed that the response of a jack-up structure subjected to a combination of static and cyclic loads is just as much influenced by the uncertainties in the loads as by the uncertainties in the soil resistance. The significant discrepancy between these results is due to the different assumptions made with respect to the uncertainties in the variables. One should therefore be careful in generalising the results obtained for a specific site to other environmental and soil conditions. For traditional jack-up foundation solutions, the stability and performance of a jack-up foundation is primarily determined by the installation procedure for the unit. This operation involves elevating the hull and pumping water ballast into the preload tanks, causing the spudcans to penetrate into soil and thereby increasing their bearing capacity. ·

The geotechnical areas of concern for jack-up foundations are: -Prediction of footing penetration during preloading. -Jack-up foundation capacity under various load combinations after preloading. -Foundation stiffness characteristics under the design storm.

The recent trend in using jack-up structures in deeper waters and on a more permanent basis has resulted in another type of foundation solution, namely spud-cans equipped with skirts. The installation of skirted footings is normally achieved by suction, not preloading. The skirted footings not only provide more predictable capacity, they also increase the footing fixity significantly. The procedure for estimating the capacity of the individual footings is based upon analytical procedures similar to that undertaken for foundation of gravity based structures. For jack-up foundation systems, however, it is important to look at the complete foundation ‘system’ because at loads close to failure, significant re-distribution of reactions among the footings may take place. (Refer to the foundation example in DNV (1996c) for more information in respect to this item.) It is evident from statistics, Sharples et al. (1989), Arnesen et al. (1988), that punch-through during preloading is the most frequently encountered foundation problem for jack-ups. Punch-through occurs when a weak soil layer is encountered beneath a strong surficial soil layer. ·

The only way to avoid punch-through is to undertake a thorough site investigation at the jack-up location prior to installation in order to identify the potentially problematic weak soil layers.

The total amount of preload used in the installation is often used as a checking parameter for the spudcan capacity to withstand extreme loads. The so-called “100% preload check” requires that the foundation reaction during preloading on any leg should be equal to, or greater than, the maximum vertical reaction arising from gravity loads and 100% of environmental loads. The preload defines the static foundation capacity under pure vertical loading immediately after installation. Under the design storm the footing is subjected to simultaneous action of vertical and horizontal loads, and overturning moment. The storm induced loads are cyclic with a short duration and the supporting soil may have a higher reference static shear strength than right after installation due to consolidation under the jack-


Page No. 17

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

up weight. On the other hand, for equal degrees of consolidation, the vertical capacity of a footing will be greater during pure vertical loading than during a combination of vertical, horizontal and moment loadings. Having regard to the oversimplification of the l00% preload check, SNAME (1993) suggests a phased method with three steps, increasing in the order of complexity, for the evaluation of foundation capacity, as follows : Step 1. Preload Check The foundation capacity check is based on the preloading capability - assuming pinned footings. Step 2. Bearing Capacity Check Bearing capacity check based on resultant loading on the footing under the design storm. Step 3. Displacement Check The displacement check requires the calculation of displacements associated with an overload situation arising from Step 2. Any higher level check need only be performed if the lower level checks fail to meet the foundation acceptance criteria. It is difficult to quantify the uncertainties associated with the “preload check” approach. Nadim and Lacasse (1992) developed a procedure for reliability analysis of the foundation bearing capacity of jack-ups. The procedure, which may be categorised as a Step 2 approach, is based on a prior calculation of the bearing capacity under different load combinations (interaction diagram) and updating the interaction diagram from the measured vertical preload. The bearing capacity calculations are performed probabilistically using the FORM approximation. The procedure developed by Nadim and Lacasse (1992) was used by Nadim et al. (1994) to study the reliability of a jack-up at a dense sand site in the North Sea. An important result of the FORM analyses is the correlation between the foundation capacity under a given combination of horizontal and vertical loads (and overturning moment if spudcan fixity is significant) and the foundation capacity under pure vertical loading. The degree of correlation determines the significance of the measured preload on reducing the uncertainty associated with foundation capacity for a given load combination. ·

For a given loading combination (vertical, horizontal and moment), the lognormal distribution function appears to provide a good fit to the foundation capacity, Nadim and Lacasse (1992).

·

The properties of the volume of soil under the footing fluctuate spatially and can be represented by a random field. The effects of this are accounted for by spatial averaging, Vanmarcke (1977, 1984), and by using stochastic interpolation techniques, Matheron (1963), if enough data exist.

·

Otherwise, the uncertainties in the soil parameters are based on the statistics of the available data. Mean and standard deviation are calculated by ordinary statistical methods, e.g. Ang and Tang (1975). Usually the probability distribution function used to represent geological processes follows a normal or lognormal law. More often than not however, and especially in the case of jack-up structures, there are not enough data


Page No. 18

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

available, and the designer needs to use correlations or normalised properties as a function of the type of soil to establish consistent soil profiles. See also DNV (1996a), Section 7.3. As an example the undrained shear strength of soft sedimentary clay normalised to the in-situ overburden stress is about 0.23 ± 0.03 for a horizontal failure mode; the friction angle of sand can be selected on the basis of its relative density and an in-situ penetration test.


Page No. 19

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

3.0

UNCERTAINTY MODELLING

3.1

General

This section provides general guidance in respect to uncertainty modelling as appropriate to the extreme load event for a jack-up structure.

3.2

Loading UncertaintyModelling

Uncertainty in the load process may be attributed to either aleatory uncertainty (inherent variability and natural randomness of a quantity) or epistemic uncertainty (uncertainty owing to limited knowledge). In respect to jack-up reliability analysis, guidance appropriate to the most significant of the uncertain variables associated with the load process is given below. 3.2.1

Aleatory Uncertainty

Tables 3.1 to 3.3 below document a summary of recommended distributions for selected stochastic variables. It should be noted however that site specific evaluation of environmental variables may dictate use of variable distributions other than those recommended in the tables below. For further guidance see also DNV (1996a), Chapter 5.

Description Randomness of storm extremes Waterdepth (D)

Marine Growth Table 3.1

: General Environmental Variable Distributions

Description Significant wave height (Hs) Zero up-crossing period (Tz) Spectral peak period (T p) Joint distribution (Hs,Tz) or (Hs,T p)

Tidal current speed (Vt) Wind generated current speed (Vw) Average wind speed (U10m) Table 3.2

Distribution Poisson Uniform (tidal effects), or, Normal (storm surge effects - conditional on Hs) Lognormal

Distribution 3-parameter Weibull/Lognormal Lognormal (conditional on Hs) Lognormal (conditional on Hs) 3-parameter Weibull for Hs and Lognormal for Tz or T p (conditional on Hs) Uniform Normal (conditional on U10m) Weibull (conditional on Hs)

: Long Term Analysis Variable Distributions


Page No. 20

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

Description Significant wave height (Hs) Total current speed (Vc) Average wind speed (U10m) Table 3.3

Distribution Gumbel *1, 2 Gumbel *1, 2 Gumbel *1, 2

: Extreme Analysis Variable Distributions

KEY :

*1 :

Normally it is sufficient to consider the extreme dominating variable being either ; -the significant wave height, -the current, or, -the wind speed, in combination with this extreme distribution the remaining two variables are assigned the distribution according to Table 3.2.

*2 :

Instead of a Gumbel distribution, a Weibull distribution (see the long term analysis variables in table 3.2), raised to the power of the number of considered seastates in one year, N Sea, may be utilised in practice. (See DNV (1996a), Section 6.7.)

3.2.2

Epistemic Uncertainty

·

The following listed time independent, basic load variables have been identified as being possible significant contributors to the overall reliability of a jack-up structures, Løseth (1990), Karunakaran (1993), Dalane (1993) ; -Drag coefficient -Inertia coefficient -Marine growth -Mass of structure.

Guidance to selection of distribution type and distribution parameters for random model uncertainty factors associated with these basic load variables is given in Table 3.4 below. Basic Variable Name Drag coefficient 2 (CD) 3 Inertia coefficient (CI) 4 Marine growth Mass of structure 5 Table 3.4

Distribution Lognormal Lognormal Lognormal Lognormal

1

1.0 1.0 1.0 1.0

C.o.V. 0.2 0.1 0.2 0.14

: Load Model Uncertainty Variables

KEY :

1:

The absolute value of the distribution variables are given relative to the value applied in the structural analysis.

2:

The selection of appropriate drag coefficients for the structural analysis are stated in SNAME (1993).

3:

For extreme value jack-up analysis, without loss of any generality, it is normally considered acceptable to select the inertia coefficient as a fixed quantitiy. An inertia coefficient of 1.8 may be uti lised.

4:

The selection of the appropriate value for the marine growth should be evaluated based upon a site specific evaluation, e.g. NPD (1992).

5:

See also section 2.3.3


Page No. 21

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

3.3 ·

Response Uncertainty Modelling Significant contributions to response model uncertainty may be attributed to the following causes, Nadim (1994), Løseth (1990), Karunakaran (1993); -Analytical uncertainty -Damping ratio -Foundation stiffness

3.3.1

Analysis Uncertainty

Analytical uncertainty accounts for the model uncertainty resulting from the statistical accuracy of a single analytical simulation (i.e. the variability resulting from different engineers, utilising different software, undertaking exactly the same analysis). With respect to jack-up response analysis this uncertainty is documented in DNV (1996a), Chapter 6. Guidance to selection of distribution type and distribution parameters for random analytical uncertainty factors is given in Table 3.5 below. Basic Variable Name Analytical uncertainty Table 3.5

3.3.2

Distribution Lognormal

1.0

C.o.V. 0.18

: Analytical Model Uncertainty Variables

Damping

Damping model uncertainty may vary depending upon the procedure adopted for including damping within the response analysis, Langen (1979). Relative velocity, hydrodynamic damping should generally not be used if Eqn. 3.1 below is not satisfied, SNAME (1993). uTn/Di ³ 20

(3.1)

where u : water particle velocity Tn : first natural period in surge/sway Di : diameter of leg chord ·

For extreme response analysis, in general, hydrodynamic damping may normally be explicitly accounted for by use of the relative velocity formulation in Morison’s equation.

·

A value for total global damping may be obtained by summation of those appropriate damping component percentages stated in Table 3.6, SNAME (1993).


Page No. 22

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

Damping Source

Structure, holding system etc. Foundation Hydrodynamic Table 3.6

Global Damping (% of critical damping) 2% 2% or 0% 1 3% or 0% 2

: Table of Recommend Critical Damping

KEY :

1:

Where a non-linear foundation model is adopted the hysteresis foundation damping will be accounted for directly and should not be included in the global damping.

2:

In cases where the Morison, relative velocity formulation is utilised the hydrodynamic damping will be accounted for directly and should not be included in the global damping.

Guidance to selection of distribution type and distribution parameters for random damping uncertainty factor associated with the response basic variables is given in Table 3.7 below. Basic Variable Name Damping ratio Table 3.7

Distribution Lognormal

1

1.0

C.o.V. 0.25

: Damping Model Uncertainty Variables

KEY :

1:

The absolute value of the distribution variables are given relative to the value applied in the structural analysis.

3.3.3

Foundation

For geotechnical analysis, model uncertainty is difficult to assess as there are few comparable full scale prototypes that have actually gone to failure and where there was enough knowledge about the site conditions and the load characteristics to enable calculation of the uncertainty. ·

Therefore to evaluate model uncertainty, comparisons of relevant scaled model tests with deterministic calculations, expert opinions and information from literature, in addition to any field observations that are available for similar structures on comparable soil conditions, are normally utilised.

Using "traditional" analysis methods to undertake the bearing capacity analysis of the spudcan of a jack-up foundation results in large model uncertainties, as was documented by Endley et al. (1981). They compared, for 70 case studies on soft clays and 15 case studies on layered profiles consisting of soft clay over stiff clay, predicted rig footing penetration with observed penetrations. The comparisons suggest a model uncertainty with mean value 1.0 and standard deviation 0.33, as based on the 70 cases studied. The observed data ranged between 0.4 and 1.55 times the predicted values. McClelland et al. (1982) undertook similar comparisons for jack-ups on uniform clay profiles and for jack-ups on layered profiles. In this study the standard deviation was about 0.20 to 0.25 about a mean of 1.0.


Page No. 23

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

The “traditional" methods of analysis are the so-called "bearing capacity formulas” which do not account for strength anisotropy, cyclic loading, soil layering, nor variation of soil properties with depth or laterally. The model uncertainty values quoted above are valid for a failure mode under vertical loading only. In the method proposed by Nadim and Lacasse (1992), a more rigorous bearing capacity approach than the "traditional" approach is used. The analysis uses a limiting equilibrium method of slices. Effects of anisotropy and cycling loading, the uncertainty in the calculation model for both vertical and horizontal (moment) loading and combined static and cyclic loading are included. The uncertainty in this calculation model was studied in detail with series of model tests at different scales. On the basis of the work carried-out by Andersen and his co-workers, Andersen et al. (1988), (l989), (1992), (1993), Dyvik et al. (1989), (1993), model uncertainty for bearing capacity of a footing in clay may be mean 1.00, standard deviation 0.05 for failure under static loading only, and mean 1.05, standard deviation 0.15 for failure under combined static and cyclic loading. For footings installed in sand, much less information exists, and tentative values may be mean 1.00, standard deviation 0.20 to 0.25, based on engineering judgement and the results of recent centrifuge model tests, Andersen et al. (1994). The model uncertainty may vary according to the failure surface. It should be noted that the mean of model uncertainty factor for most offshore foundations (e.g. piles in sand and clay, shallow foundations on sand) is greater than 1.0, i.e. the analytical models tend to be conservative. The methods developed for shallow foundations on clay, however, have been fine-tuned and calibrated against large-scale tests in the past 20 years, and much of the inherent conservatism in the methods has been removed. Little information exists on the model uncertainty associated with the foundation displacement of a jack-up structure (see step 3 in section 2.3.6) and the model uncertainty can only be guessed for those cases. A model uncertainty with a coefficient of variation of at least 50 % is expected. Guidance to selection of distributions associated with the foundation parameters is given in Table 3.8 below. Reference should also be made to DNV (1996a), Section 7.3. Description Rotational stiffness Horizontal stiffness Vertical stiffness Table 3.8

: Foundation Parameter Distributions

KEY :

*1 :

See also section 2.3.6

Distribution*1 Lognormal Lognormal Lognormal


Page No. 24

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

3.4

Resistance Uncertainty Modelling

The level of reliability of jack-up structures is “load driven”, Ronold (1990), Dalane (1993), that is to say that the importance of the uncertainties in the loading is much greater than the importance of the uncertainties in the capacities. As a consequence of this it is most likely that a structural failure event will result from the load being high, rather than the strength capacity being low. ·

Uncertainties associated with resistance are dependent upon the resistance model included in the limit state under consideration. Modelling of the uncertainly parameters associated with the resistance model should be relevant to the formulation of the resistance model utilised in the limit state. See section 4.0 for further guidance.

·

General resistance uncertainty information is given in DNV (1996a), Chapter 7.

·

A realistic analysis of the ultimate (‘push-over’) capacity of a jack-up structure can in many cases only be performed by using advanced non-linear finite element software, e.g. USFOS (1996).


Page No. 25

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

4.0

LIMIT STATES

4.1

General

Limit states are formulations of physical criteria beyond which the structure no longer satisfies the design performance requirements. Limit state categorisation is generally defined as follows, ISO 13819, Part 1, ISO (1995) ; a).

The ultimate limit states that generally correspond to the maximum resistance to applied actions.

b.)

The serviceability limit states that correspond to the criteria governing normal functional use.

c.)

The fatigue limit states that correspond to the accumulated effect of repeated actions.

d.)

The accidental damage limit states that correspond to the situation where damage to components has occurred due to an accidental event.

Some code of practices, e.g. Eurocode 3 (1992), however, defines only two limit states, these being ; the Ultimate Limit State, and the Serviceability Limit State. In such cases the states prior to structural collapse which, for simplicity are considered in place of the collapse itself, are also classified and treated as the ultimate limit state. 4.1.1

Limit States Appropriate to Jack-up Structures

Serviceability Limit State (SLS) · For steel structures, the serviceability limit state is not normally a designing criterion and is therefore not further discussed within this section. Fatigue Limit State (FLS) · The fatigue limit state is a relevant limit state to consider for jack-up structures. Both for long term site engagements and for the transit condition, the fatigue limit state may be designing. ·

The guidance provided in the guideline example for jacket structures, DNV (1996c), in respect to the fatigue limit state, although utilising frequency domain solution techniques, covers the state-of-the-art knowledge with respect to fatigue reliability analysis of jackup structures. The fatigue limit state is therefore not explicitly covered in this section and reference should be made to DNV (1996c) for appropriate guidance concerning the fatigue limit state.

Ultimate Limit State (ULS) ISO 13819, Part 1, ISO (1995), lists the following examples of ultimate limit states ; a.)

loss of static equilibrium of the structure, or of a part of the structure, considered as a rigid body (e.g. overturning or capsizing),

b.)

failure of critical components of the structure caused by exceeding the ultimate strength ( in some cases reduced by repeated actions) or the ultimate deformation of the components,


Page No. 26

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

c.)

transformation of the structure into a mechanism (collapse or excessive deformation),

d.)

loss of structural stability (buckling etc.),

e.)

loss of station keeping (free drifting), and

f.)

sinking.

·

The ultimate limit state for jack-up structures is difficult to describe through simple design equations. Additionally, general guidelines on how to perform structural system collapse analyses are lacking, hence limit state functions for reliability analysis of jackup structures are general based on design equations for single components.

For a jack-up in the elevated mode of operation the following listed ultimate limit states may be considered as designing ; Component Level -leg local structural strength -hull local structural strength -foundation capacity (local) -holding system loadings ·

The following listed limit states may therefore be considered as being relevant component limits states for reliability analyses ; -1Leg element yield -2Leg element buckling -3Leg joint capacity -4Foundation bearing failure -5Holding system capacity

System (Global) Level -leg global structural strength -hull global structural strength -overturning stability -horizontal deflections -foundation capacity. Accidental Damage Limit State (ALS) The accidental damage limit state check ensures that local damage or flooding does not lead to complete loss of integrity or performance of the structure. ·

4.2

The intention of this limit state is to ensure that the structure can tolerate the damage due to specified accidental events and subsequently maintain integrity for a sufficient period under specified environmental conditions to enable evacuations to take place. The accidental events and the consequences of such events are normally based upon Quantitative Risk Analyses (QRA). For further details on QRA reference should be made to DNV (1996a), Chapter 2.

The Ultimate Limit State


Page No. 27

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

This subsection describes in more detail Ultimate Limit State criteria documented in subsection 4.1.1. 4.2.1

Leg Strength

General As previously mentioned, (see Section 3.4), reliability of a jack-up structure in the ultimate limit state condition is found to be ‘load driven’, i.e. the importance of the uncertainties associated with the loading dominates. When describing the uncertain quantities associated with the limit state it is generally therefore not necessary to breakdown the individual uncertainties associated with, for example, a buckling resistance code formulation, and code criteria may be utilised with generalised randomisation parameters. ·

Suitable strength resistance criteria, may be found in a wide variety of structural codes and standards. The following references may be recommended ; -AISC (1984) -API (1993) -DNV (1995) -Eurocode 3 (1992) -NPD (1990) -SNAME (1993)

When utilising standard codes and Practices the following issues should be considered ; (i)

The formulations contained in these codes may only be applicable within certain limits (e.g. R/t ratio between given limits). It should therefore be ensured that the resistance formulation utilised in the limit state is satisfactory for the structure under consideration.

(ii)

The resistance formulations contained within these codes are based upon analytical approximations to the physical behaviour where characteristic values are defined at some fractile value or lower bound value. For reliability analysis the capacity formulation in the limit states should be based on the 50 percent fractile (median) values. The basis for buckling curves in different codes and standards are different. The API buckling curve, API (1993) is derived as a lower bound value for low slenderness while it is equal to the Euler stress for high slenderness values, which may be considered as an upper bound value in that region. Another definition of a buckling curve is used in AISC (1984). The background for the buckling curves used in design of steel structures in European design standards is based on work carried out within the European Convention for Constructional Steelwork which is presented in The Manual on Stability of Steel Structures, ECCS (1976). The design curves are presented by their characteristic values which are defined as mean values minus two standard deviations along the slenderness axis. The test results are assumed normal distributed.

(iii)

Effective buckling lengths are dependent upon joint flexibilities. Buckling lengths may normally be measured in relation to centreline to centreline for chords, whilst, face to face lengths are normally acceptable for the braces. X-brace buckling lengths depend upon the amount of tension loading in the crossing member. The effective lengths may be derived from analytical considerations. The effective buckling


Page No. 28

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

lengths derived from tests of frame structures until collapse are generally shorter than those derived from theoretical calculations. (iv)

Different allowable requirements to fabrication tolerances (eccentricity) are associated with the various buckling curves. For European buckling curves a straightness deviation at the middle of the column equal to 0.0015 times the column length is allowed, while for API (1993) and AISC (1984) the corresponding numbers are 0.0010 and 0.00067 respectively.

For conventional design jack-up structural elements the effect of external pressure may, normally be disregarded. The susceptibility of local buckling of tubular members is a function of the member geometry and yield strength. For jack-up structures, it may normally be assumed that leg elements are stocky, beam elements. Yield strength control is implicitly covered by the buckling limit state for members in compression, whilst, for tension members, the limit state is given by, for example, eqn. 5.1, NPD (1990), NS3472 (1984).

G = f

y-

[

s

a+ s

by+ s

2

bz

]

+

3[

t

xy+ t

xz+ t

2 t

]

(5.1)

where

f y t t

= material yield strength = axial stress component = torsional shear stress component

, s bz t xy , t xz

= bending stress components = plain shear stress components

sa

s by

The capacity criterion stated in SNAME (1993) is an example of an expression applicable to describe resistance of jack-up elements subjected to compressive loadings. Such formulation may be described in the limit state format as ;

é

G = 1 - X bias [

ù

ìï M uey üï P u 8 ì M uex ü + êí ý +í ý ú P n 9 êî M nx þ ïî M ny ïþ ú h

ë

h

1 h

]

û

Where ;

P u P n M uex M uey

is the chord axial load is the chord nominal axial strength in compression is the chord local effective applied bending moment about the local x-axis is the chord local effective applied bending moments about the local y-axis

M nx M ny

is the chord local nominal bending strength about the local x-axis

h

is the exponent for biaxial bending.

is the chord local nominal bending strength about the local y-axis

(5.2)


Page No. 29

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

A full description of limiting criteria, the parameters utilised in Equation 5.2, and the methodology utilised in calculating the specific values of these terms are documented in SNAME (1993), Section 8.1.4. The SNAME (1993) formulation for buckling resistance is based upon AISC (1978). The uncertainty parameters stated in Galambos (1988) may therefore be utilised in describing the uncertainty parameters including X bias. Joint Capacity Joint capacity design equations have been established for the static strength of tubular joints. The equations in API (1993) and NPD (1990) show a similar shape although the coefficients are different as also might be expected as the API (1993) are based on allowable stresses and NPD (1990) has based the design on the partial coefficient method. Jack-up brace/chord connections are, however, normally non-standard, due to the rack structure inclusion in the chord section. Static strength capacity formulation for standard tubular/tubular connections may give erroneous results for brace/chord connections. Work on joint capacities is currently being performed in development of a new ISO standard on design of steel offshore structures. This work should be considered as basis for limit state functions when it is available. As an example limit state Eqn 5.3 documents the static strength of tubular joints formulation based on the NPD guidelines, NPD (1990) and the limit state function for the static capacity of tubular joints can then be formulated, NPD(1990) as ; 2

G = 1 - X bias [

N æ M IP ö M +ç ÷ + OP N k è M IPk ø M OPk

]

where X bias

= bias (See DNV (1996a), Chapter 7.2) = brace axial force

N k M IP M IPk M OP M OPk

= characteristic capacity of the brace subjected to axial force = brace in-plane moment = characteristic capacity of the brace subjected to in-plane moments = brace out-of-plane moment = characteristic capacity of the brace subjected to out-of-plane moments

A detailed description of this limit state is given in DNV (1996c).

(5.3)


Page No. 30

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

4.2.2

Foundation Bearing Failure

The limit state function for the ultimate limit state of foundation bearing capacity is defined as : G = R - L, where R and L are respectively the lengths of resistance and load vectors as shown in Fig 4.1. The origin of the vectors on the vertical axis, P w, is the static load on the footing due to submerged weight of the jack-up. The end point of vector L, point A, is the coordinate in the load space under the design storm. The end point of vector R, point B, is the foundation bearing capacity along the load path Pw®A. For the limit state function, G, the lengths of resistance, R, and load vectors, L, are defined as follows ; L= R= Vex Hex Mex Vcy,f

= = = =

Hcy,f

=

Mcy,f

=

Pw

=

r

=

(Vex

-

Pw ) 2

(Vcy, f - Pw ) 2

+

+

( Hex ) 2

(H cy , f ) 2

+

( Mex / r ) 2

(5.4)

( Mcy , f / r) 2

(5.5)

+

Vertical load on footing under the extreme load combination Lateral load on footing under the extreme load combination Moment load on footing under the extreme load combination Vertical capacity of footing along the path defined by load vector starting at (Pw,0,0) in direction of (Vex, Hex, Mex) Lateral capacity of footing along the path defined by load vector starting at (Pw,0,0) in direction of (Vex, Hmax, Mex) Moment capacity of footing along the path defined by load vector starting at (Pw,0,0) in direction of (Vex, Hmax, Mex) Mean vertical load on footing during the storm (mainly due to submerged weight of jack-up) Radius of footing (reference length used for normalising the moment)

The values of Vcy,f , Hcy,f , and Mcy,f are obtained by extending the load vector starting at (Pw,0,0) in the direction of (Vex, Hex, Mex) until it intersects the bearing capacity interaction diagram as shown on Fig. 4.1a. L and R are the lengths of the extreme load and resistance vectors shown on Fig. 4.1b. 4.2.3

Holding System

The limit state function for the ultimate limit state of holding system capacity is defined as : G = R - S, where R is the ultimate holding capacity of the jacking system and S is the response loading. The ultimate capacity of the holding system is usually obtained by detailed finite element analysis (F.E.M. analysis) in combination with relevant prototype testing.


Page No. 31

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

Figure 4.1 :

Definition of Limit State Function for a Footing on Clay with Moment Fixity.


Page No. 32

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

4.2.4

Global Deflections

The limit state function for the ultimate limit state of global deflections is defined as : G = R - S, where R is a stated value (some prescribed threshold), e.g. chosen from considerations in respect to proximity to another offshore installation, and S is the response displacement. 4.2.5

Global Leg Strength

The structural behaviour beyond first member failure depends not only on the ability of the structure to redistribute the load, but also on the post-failure behaviour of the system, e.g. the ductility of the individual members and joints. For a balanced structure, i.e. where all members, in a linear analysis, have the same utilisation at the time of first member failure, the first member to fail and the system effects for overload capacity beyond the first member failure are determined by randomness in member capacity. As the uncertainty in the structural capacity is much less than that in the loading, Dalane (1993), and the structure is not balanced, there will normally be only a few failure modes that will dominate. The identification of such members is however, complicated by simplicities made in the analysis e.g. at the interfaces between the hull and the leg structures, and at the foundation interfaces. There has been little previous workings undertaken concerning jack-up collapse analysis related to reliability analysis, however, by referring to jacket experience, it is considered that the collapse capacity may be directly related to the global overturning moment. This implies that the collapse capacity can be represented by a single random variable. The loading may also be represented by a single random variable, and, as such, the limit state function for the ultimate limit state of global leg strength capacity may be defined as : G = R - S, where R is the strength capacity of the leg (i.e. the overturning moment) and S is the loading. Guidelines related to the total collapse of jacket structures are given in (1995c). Such guidelines may form the basis for considerations relevant for the collapse (‘push-over’) analysis of a jack-up structure. 4.2.6

Overturning Stability

Jack-up overturning stability criteria are documented in various publications, e.g. SNAME (1993), DNV (Feb 1992). An example of this limit state is given by SNAME (1993) as ; G = ( MD + ML + MS ) - ( ME + MDN ) MD ML MS ME MDN

(5.6)

= the stabilising moment due to weight of structure and non-varying loads (at the displaced position) = the stabilising moment due to the variable loads(at the displaced position) = the stabilising moment due to the seabed foundation fixity = the overturning moment due to the extreme environmental load condition = the dynamic overturning moment

When considering the moments in connection with this limit state it is important to ensure that the axis of rotation of the system is fully considered.


Page No. 33

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

4.3

Literature Study

From a literature review it may be concluded that there have, in the past, been few public papers issued concerning structural reliability of jack-up units. From an extensive documentation review the following listed reliability studies have been identified in respect to jack-up structures ; General Structural Reliability Papers ; 1.) 2.) 3.) 4.) 5.) 6.) 7.) 8.)

Løseth, R., Mo, O., and Lotsberg, I, (1990) Leira, B.J., and Karunakaran, D. (1991) Mo.O., et.al. (1991) Ahilan, R.V. et.al. (1992) Gudmestad, O.T., et.al. (1992) Karunakaran, D., et.al. (1993) Ahilan, R.V., Baker, M.J., and Snell, R.O., (1993) Dalane J.I.(1993)

The majority of the papers referred to above may be considered as providing information concerning general reliability. Løseth et.al. (1990) and Karunakaran et al.(1993) document the global limit state criteria of maximum axial force and base shear in one leg. Karunakaran et al.(1993) also documents considerations with respect to deck displacement and foundation limit states. Ahilan et al.(1992), (1993) covers reliability code calibration studies undertaken in connection with SNAME (1993). Mo et al. (1991) and Dalane (1993) document structural leg strength capacity considerations. Foundation Reliability Papers ; 1.) 2.) 3.)

Ronold, K.O., (1990) Nadim, F., Lacasse, S., (1992) Nadim, F., Haver, S., and Mo, O. (1994)


Page No. 34

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

5.0

SUMMARY OF APPLICATION EXAMPLES

5.1

General

This section documents a summary of the reliability analyses undertaken to analyse the response of a jack-up structure in a typical North Sea environment at a waterdepth of 81 metres as documented in DNV (1996b). In order to assess change in reliability as a function of time, the reliability examples are undertaken for a jack-up exposed to multi-year operation at the same location. The following listed time dependent effects have been considered in the analyses ; - Soil Consolidation The foundation rotational stiffness was increased by a factor of 2.5 to account for soil consolidation. - Drag Coefficient Drag coefficients were increased by a factor of 15% to account for the change in drag due to increased roughness. - Marine Growth Marine growth diameter thickness’ according to the values recommended by the NPD (1992) were applied. - Deckbox Mass The total mass of the rig was assumed to have increased by a factor of 10% to account for weight growth in the deckbox. Two limit states have been considered covering the structural strength of the jack-up leg and the foundation capacity. In both of these cases the effects on reliability of long term operation at the specific site have been evaluated. The reliability analyses documented in DNV (1996b) have been undertaken by the methodology generally known as ‘Long Term Statistics by Independent Seastates’, Bjerager et al. (1988), and were based upon response resulting from time domain simulations in irregular seastates. Report DNV (1996b) fully documents the following items ; introduction to the problem stating assumptions and provisions theory of the models for representation of the problem a description of the limit state formulation and the formulation itself probabilistic and deterministic modelling descriptions the reliability analysis procedures results of the analysis, including reliability indices, failure probabilities, uncertainty importance factors, and parametric sensitivity factors discussion and conclusions.

5.2

Overview of Analytical Procedure

Utilising site specific criteria, detailed deterministic and simplified dynamic, non-linear analyses were undertaken in order to determine appropriate jack-up response statistics.


Page No. 35

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

Long term statistics were established by use of PROTIM (1989). PROBAN (1989) was utilised to solve the probabilities of failure of the limit state functions. For the foundation example a probabilistic bearing capacity model was established in order to account for the different combinations of force and moment at the foundation footing. An overview of this procedure is shown schematically in figure 5.1.

DESIGN CRITERIA

DETAILED MODEL ANALYSES (Deterministic Sea)

SIMPLIFIED MODEL ANALYSES (Stochastic Sea)

WAJAC

FENRIS FENSEA

PROBAN

STRUCTURAL RELIABILITY OUTPUT : The annual probability of failure for the most critically loaded structural element. (Determined by establishing the long term statistics considering independent seastates)

FOUNDATION RELIABILITY OUTPUT : The annual probability of failure for the most utilized footing.

Figure 5.1

: Overview of Analytical Procedure

ESTABLISH CRITICAL PARAMETERS ; -Load Direction -Design Criteria -Element -Foundation soil springs

ESTABLISH THE RESPONSE STATISTICS ; -Force and moment for the most critically loaded structural element -Force and moment for the most utilized footing

PROTIM

ESTABLISH PROBABILISTIC BEARING CAPACITY MODEL for the different combinations of force and moment on most utilized footing

ESTABLISH DISTRIBUTIONS OF ANNUAL EXTREMES for force and moment on most utilized footing


Page No. 36

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

5.3

Structural Reliability Example

This example documents an ultimate limit state reliability analysis undertaken for a jack-up structure exposed to multi-year operation. The foundation description ‘unconsolidated soil’ is intended to reflect a cohesive soil condition (e.g. clay) at the time of the initial placement of the jack-up unit. The ‘consolidated soil’ condition is a condition where, at the same location, after a given period of operation, say 10 years, the foundation is considered to have settled and consolidated. Failure probability of leg, chord buckling provided the measure of the change in reliability with time. An overview of the analytical methodology adopted in the reliability analysis is shown in figure 5.1. The main results from the undertaken reliability analysis are presented in table 5.1. Table 5.2 presents results from the sensitivity evaluation, where the mean and standard deviation have been increased by 10% over those values utilised in the undertaken reliability analysis.

SORM Reliability index - Unconsolidated Soil : = 4.35 - Consolidated Soil : = 4.41 Variable

Significant Wave Height, Hs Randomness of Storm Extreme, Uaux Drag Coefficient, CD Critical Stress, Fcr Heading, q Wave Spreading, n Foundation Rotational Stiffness, K r Tidal current, VT Damping Deckbox Mass Table 5.1

Unconsolidated Soil Importance Factor 56% 16% 11% 9% 3% 2% 2% <1% <1% <1%

Consolidated Soil Importance Factor 44% 15% 15% 10% 1% 4% 9% 1% 1% <1%

: Structural Reliability Importance Factors

Unconsolidated

Soil

Consolidated Soil Condition


Page No. 37

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

Variable

Environment Rotational Stiffness, K r Vertical Stiffness, K v Lateral Stiffness, K h Drag Coefficient, CD Tidal current, VT Marine Growth Damping Deckbox Mass Wave Spreading, n Waterdepth, D Spectral Peak Parameter, g Heading, q Yield Strength, f y Critical Stress, Fcr Duration, D No. of Seastates, Nsea Table 5.2 value

Condition Mean / Lower Bound N/A 0.0202 -0.0063 -0.0008 -0.1686 -0.0340 -0.0025 -0.0090 0.0193 -0.0037 0.0074 -0.0008 -0.0424 0.0054 0.2525 -0.0195 -0.0214

CoV / Upper Bound -0.3138 -0.0038 0.0002 0 -0.0379 0.0007 0.0001 0.0003 -0.0007 -0.0377 0.0075 -0.0040 -0.1728 0 -0.0421 N/A N/A

: Sensitivity Analysis of Results (

Mean / Lower Bound N/A 0.0286 0 0 -0.1847 -0.0316 -0.0186 -0.0369 0.0600 -0.0048 0.0322 -0.0019 -0.0260 0 0.2643 -0.0197 -0.0214

CoV / Upper Bound -0.2611 -0.0227 0 0 0.0497 0.0005 0.0001 -0.0013 -0.0043 -0.0764 0.0315 -0.0134 -0.0637 0 -0.0467 N/A N/A

for a 10% increase in the mean

and CoV for selected variables) Key : N/A

: Not applicable

The reliability levels resulting from the example seem to be relatively high for a jack-up unit when compared to other relevant studies for jack-up units, e.g. SNAME (1993). The main reason for this is that the jack-up chord element under investigation in the example, although being the most heavily loaded structural element, is not loaded up to the allowable deterministic capacity of the element in the designing storm condition. The condition analysed was however based upon an actual loading situation for the jack-up unit. This example would therefore tend to confirm the in-service experience that jack-up units generally operate at reasonably high levels of reliability in respect to structural strength due to the fact that, in the normal mode of operation, the jack-up is not utilised to the maximum capability of the jack-up unit in respect to the leg strength ultimate limit state condition. For jack-up units designed to operate as production units over a longer period of time at a single location, where the jack-up is designed and optimised for site specific criteria, such a conclusion can not however be made from the investigation performed in the example. Over the period of time considered, the reliability of the jack-up is found to remain fairly constant in the example presented. It would appear that the time varying negative effects of increased static and environmentally induced loadings are offset by the effects of soil consolidation. In the case represented in the example study, consolidation of the foundation has lead to an increased bottom restraining condition. Other soil conditions may however lead to degradation of the foundation restraint. In all cases site specific data should be utilised as the basis for evaluating the long term effects of the foundation.


Page No. 38

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

5.4

Foundation Reliability Example

The foundation reliability example documented in DNV (1996b) demonstrates ultimate limit state analyses undertaken for the stability of the most utilised footing for 'unconsolidated' and 'consolidated' soil conditions. Each leg of the jack-up considered in the presented case study was supported by a 20 m diameter footing with 6 m skirts. The site consisted of 2 clay layers: a soft clay layer down to 5 m depth and a stiff, overconsolidated clay layer underneath. The mechanical model for evaluating the capacity of skirted footings in clay was assumed well developed and the modelling uncertainty relatively small. The limit state function for the ultimate limit state of bearing capacity for the most utilised footing was defined as G = R - L, where R and L were respectively the lengths of the resistance and load vectors as shown on Fig. 4.1. The distribution of the resistances was estimated by specifying a deterministic load on the foundation and evaluating the probability of failure using FORM. By varying the load, the probability of failure at different load levels was computed. The results showed that a lognormal distribution provides an excellent fit for the static foundation capacity. The CoV's and distributions of the foundation resistance parameters used in the analyses are given in Table 5.3 (see Section 4.2.2 and Fig. 4.1 for definitions).

VARIABLE

Distribution Mean CoV Unconsolidated clay (all layers) V pre Lognormal 212 MN 12% Hs,max Lognormal 40 MN 13% Ms,max Lognormal 640 MNm 14% Consolidated clay (all layers) V pre Lognormal 253 MN 12% Hs,max Lognormal 51 MN 13% Ms,max Lognormal 777 MNm 14% Other variables (same for consolidated and unconsolidated conditions) F1 Normal 1.06 3% F2 Normal 0.72 3% F3 Normal 0.78 3% Table 5.3

: Foundation Resistance Parameters

The extreme loads on the most utilised footing were computed by PROBAN (1989). Table 5.4 shows the load parameters used in the foundation reliability calculations. The CoV of Pw was assumed to be identical to the CoV of the deck mass.


Page No. 39

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

VARIABLE

Vex - Pw Hex Mex Pw Vex - Pw Hex Mex Pw Table 5.4

Distribution Mean Unconsolidated Soil Condition Gumbel 5.0 MN Gumbel 4.9 MN Gumbel 169.5 MNm Lognormal 71.6 MN Consolidated Soil Condition Gumbel 5.3 MN Gumbel 6.4 MN Gumbel 323.9 MNm Lognormal 78.7 MN

CoV

111% 54% 82% 14% 118% 49% 51% 14%

: Extreme Loads on Most Utilised Footing

When the effects of load redistribution among the footings were neglected, the computed foundation safety indices were respectively 1.85 and 1.45 for the unconsolidated and consolidated soil conditions. The reason for these low values was that when the possibility of load redistribution among the jack-up legs was not taken into account, the failure mode of the most utilised leg was governed by the large overturning moment for both soil conditions. This failure mode, however, is not realistic for a 3-leg jack-up structure because for the whole foundation system consisting of the 3 footings, it is more optimal to resist the external overturning moment by axial forces, rather than by local moments at each footing. With traditional spud cans, the moment fixity is completely lost when the bearing capacity is reached. However, with skirted spud cans, the moment acting on the most utilised footing at failure may be 60 to 80% of the moment capacity. The main results from the foundation reliability analyses, after accounting for the redistribution of reactions among the 3 footings and reduction of fixity of the most utilised footing at large loads, are summarised in Table 5.5.

FORM Reliability index - Unconsolidated Soil : = 4.11 - Consolidated Soil : = 4.22 Variable

Static Sliding Capacity, HSmax Cyclic Loading Factor, F2 Extreme Base Shear, Hex All other parameters Table 5.5

Unconsolidated Soil Importance Factor 11% 1% 88% <1%

Consolidated Soil Importance Factor 13% 1% 86% <1%

: Results for Most Utilised Footing with Load Redistribution


Page No. 40

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

There is a lack of documentation concerning the reliability of jack-up foundation ultimate limit state conditions. For the example application, it was considered appropriate to compare the computed safety indices with those in Table 2.7 of the Reliability Guidelines DNV (1996a). This table presents target annual failure probability and corresponding reliability indices. Once the effects of optimal utilisation of the foundation 'system' (i.e. redistribution of reactions among the 3 footings when the loads approach the foundation capacity) are considered, the foundation failure development may be considered as being 'ductile with no reserve capacity'. The failure consequence is considered as being somewhere between 'not serious' and 'serious'. Therefore an annual target failure probability of 10-4 to 10-5 (b = 3.71 to 4.26) is appropriate. The safety indices of b = 4.11 for the unconsolidated soil condition and b = 4.22 for the consolidated soil condition would therefore appear to be satisfactory.


Page No. 41

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

6.0

RECOMMENDATIONS FOR FURTHER WORK

6.1

General

From the literature review study undertaken and documented in section 4.3, it is clear that there have been very few reliability analyses undertaken for jack-up structures. ·

6.2

This section discusses recommendations for further workings in connection with identification of the reliability of jack-up structures.

Elevated Condition

Jack-up structures have traditionally been used in shallow waters. There is a current tendency to utilise jack-ups in deeper waters, in harsh environment conditions, for extended periods of operations. ·

The uncertainty weightings for these two scenarios are different and the safety implicit in current jack-up design procedures may not necessarily be appropriate, Gudmestad (1990), Dalane (1993).

Jack-up structures, as compared to jacket structures, have a number of unique characteristics, which add to the complexity of the problem being considered. (e.g., See sections 1.4 and 2.3). With respect to limit state formulation, the most important of these characteristics may be considered as ; (i) The non-linearities in the system generally preclude the use of linear analytical procedures (see section 2.3). (ii) Jack-up chord sections normally include a rack construction. This means that traditional formulation for stress concentration factors and joint static capacity (e.g. punching shear) are generally not appropriate. ·

Results from reliability studies undertaken for traditional jacket type offshore structures are, generally, not ‘transferable’ to jack-up structures.

The stiffness characteristics (fixity) of spudcan footings are complicated and strongly nonlinear. Jostad et al. (1994) show that while spudcans might have significant moment fixity under operational loads, the moment fixity disappears as the loads approach foundation capacity. The footing stiffness affects the dynamic characteristics of the jack-up, which in turn influence the loads on the spudcans. So far, there have been no systematic studies of the effects of the uncertainties in the spudcan stiffness characteristics on the jack-up response. Conclusions : -1-

The implicit probability of failure of jack-ups by use of dedicated jack-up codes and standards should be evaluated for their applicability to deep water, harsh environment operations for extended periods.

-2-

Jack-up system capacity due to accidental damage load events should be evaluated. The robustness of the jack-up structure should then be compared to that of a jacket structure. (The U.K., H.S.E. is currently engaged in such a project and the findings from these workings should be considered in this connection.)


Page No. 42

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

-3-

Traditional, frequency domain (linear analysis) based fatigue reliability should be compared with that reliability achieved utilising time domain (non-linear) analysis in order to identify, for a jack-up structures, the importance of the non-linear effects for the fatigue limit state.

-4-

Reliability considering the following foundation related criteria is recommended to be investigated ; -system effects -response as related to the uncertainty and non-linearity in foundational support.

6.3

Floating / Installation Phase Conditions

Of the 250 jack-up casualties reported during the period 1979 to 1991, some 50% of the total losses, or major incidents occurred during towage, Standing and Rowe (1993). Standing and Rowe (1993) document the following listed items as being the major source of accident in respect to a jack-up in the transit condition ; (i) Wave damage to the unit structure leading to penetration of watertight boundaries. (ii) Damage to the structure as a result of shifting cargo (usually caused by direct wave impact, excessive motions and/or inadequate seafastenings). (ii) Structural damage in the vicinity of the leg support structures. ·

There does not appear to have been any reliability studies undertaken for jack-ups in the transit condition.

During the installation phase, there are normally two main areas of concern, these being; impact loadings upon contact with the seabed, and, foundation failure (i.e. punch-through) during preloading. Sharples et al (1989) summarised the causes for jack-up mishaps in a 10 year period. Out of 226 “accidents", over 50 were attributed to “soils”. The causes for unsatisfactory foundation performance were distributed as follows: Punch-through of footings 70% Failure due to storm loading 16% Scour around footings 5% Other causes 9% Based on a survey of major accidents between 1980 and 1987, Arnesen et al. (1988) came to similar conclusions. ·

It is evident from the above statistics that punch-through during preloading is the most frequently-encountered foundation problem for jack-ups.

·

The physics of the impact loading problem are extremely complicated and the uncertainties in the process are not well documented. Additionally, regulation requirements for the installation condition are considered to be vague and incomplete.


Page No. 43

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

Conclusions : -1-

Reliability analysis for the transit condition would appear to be necessary, not least, in order to understand the importance of uncertainties associated with the process and to identify areas where further workings are required.

-2-

Reliability investigations in the installation phase should be considered for the following listed loading conditions ; -preloading -impact loading.


Page No. 44

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

7.0

REFERENCES

Ahilan, R.V. et.al.(1992), ‘Reliability Based Development of Jackup Assessment Criteria’, Tenth Structures Congress (ASCE), San Antonio, 1992. Ahilan, R.V., Baker, M.J., and Snell, R.O.(1993), ‘Development of Jackup assessment Criteria using Probabilistic Methods’, OTC 7305, Houston, 1993 AISC(1984),’Specification for the Design, Fabrication and Erection of Structural Steel for Buildings’, American Institute of Steel Construction, Eighth Edition, Oct.1984. Andersen, K.H. and Lauritzsen, R.(1988), ‘Bearing Capacity for foundation with Cyclic Loads,’, ASCE. Jorn. of Geotechnical Engineering. V 114, No 5, pp. 516-555 Andersen, K H., Lauritzsen, R., Dyvik, R., and Aas, P.M.(1988), ‘Cyclic Bearing Capacity Analysis for Gravity Platforms; Calculation Procedure, Verification by Model Tests, and Application for the Gullfaks C Platform.’, Proc. BOSS'88 Conf. Trondheim, Norway. V 1, pp. 311-325 Andersen, K.H., Dyvik, R., Lauritzsen, R., Heien, D., Hårvik L., and Amundsen, T., (1989), ‘Model Tests of Gravity Platforms. II: Interpretation.’ ASCE. Jorn. of Geotechnical Engineering. V 115, No 11, pp. 1550-l568. Andersen, K.H., Dyvik, R., and Schrøder, K.(1992), ‘Pull-Out Capacity Analyses of Suction Anchors for Tension Leg Platforms.’, Proc. BOSS'92 Conf. London, U.K. V 2, pp. 13111322. Andersen, K.H., Dyvik, R., Schrøder, K., Hansteen, O.E., and Bysvecn, S.(1993). ‘Field Tests of Anchors in clay II: Predictions and Interpretation.’, ASCE Jorn. of Geotechnical Engineering. V 119, No 10, pp. 1532-l549. Andersen, K.H, Allard, A. and Hermstad J.(1994), ‘Centrifuge Model Tests of A Gravity Platform on Very Dense Sand; II. Interpretation.’, Proc. BOSS'94 Conf. Cambridge, Mass. USA. Vol. 1, pp. 255-252. Ang, A.H.S. and Tang, W.H.,(1975), ‘Probability Concepts in Engineering Planning and Design. Volume I - Basic Principles.’, John Wiley and Sons, Inc., New York, 409p. Arnesen, K., Dahlberg, R., Kjeøy, H., and Carlsen, C.A.,(1988), ‘Soil -Structural Interaction Aspects for Jackup Platforms’, BOSS’88 Conf. Trondheim, Norway, June 1988. API(1993), ‘Recommended Practice for Planning, Design and Constructing Fixed Offshore Platforms -Load and Resistance Factor Design’, API Recommended Practice 2A-LRFD (RP 2A-LRFD), First Edition, July 1993. Bjerager, P., Løseth, R., Winterstein, S., and Cornell, A., (1988) ‘Reliability Method for Marine Structures Under Multiple Environmental Load Processes’, Proceeding of 5th International Conf. on Behaviour of Offshore Structures, Vol.3, Trondheim, Norway, June 1988, pp1239-1253. Boswell(1986), ‘The Jackup Drilling Platform’, Edited by L.F.Boswell, City University, London. Collins Publication, 1986


Page No. 45

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

Bærheim M.(1993), ‘Structural Effects of Foundation Fixity on a Large Jackup’, Proc. The Jackup Platform, 4th Int.Conf., 1993 Dalane, J.I.(1993), ‘System reliability in Design and Maintenance of Fixed Offshore Structures’, Dr.Ing. Thesis, NTH, May 1993. DNV(1992),‘Structural Reliability Analysis of Marine Structures’, DNV Classification Note No. 30.6, Example 4.5, July 1992 DNV (Feb. 1992), ‘Strength Analysis of Main Structures of Self-elevating Units’, Classification Note no. 31.5, Feb. 1992. DNV(1995), ‘Buckling Strength Analysis’, Det Norske Veritas Classification Note no. 30.1, July 1995. DNV(1996), ‘Rules for Classification of Mobile Offshore Units’, Det Norske Veritas, Part 3 Chapter 1, ‘Structural Design General’, January 1996 DNV(1996a) “Guideline for Offshore Structural Reliability Analysis - General”, DNV Technical Report no.95-2018, Dated: May 1996 DNV(1996b),‘Guidelines for Offshore Structural Reliability - Examples for Jackups’, DNV Technical Report no.95-0072, Dated: February 1996 DNV(1996c),‘Guidelines for Offshore Structural Reliability Applications to Jacket Platforms’, DNVI Technical Report no.95-3203, Dated: Draft Format May 1996. Dyvik, R., Andersen, K.H., Madshus, C., and Amundsen, T., ( 1989). ‘Model Tests of Gravity Platforms I: Description.’, ASCE. Jorn. of Geotechnical Engineering. V 115, No 10, pp. 1532-1549. Dyvik, R., Andersen, K.H., Hansen, S.B., and Christophersen, H.P. (1993). ‘Field Tests of Anchors in Clay I: Description.’, ASCE Jorn. of Geotechnical Engineering. V 119, No 10 pp. 1515-1531. ECCS(1976), ‘Manual on Stability of Steel Structures’, Second Edition, June 1976 Endley, S.N., Rapoport, V., Thompson, V.J., and Baglioni, V.P.(1981). ‘Predictions of JackUp Rig Footing Penetration’, 13th Offshore Technology Conference, Houston, Texas, USA, Paper OTC 4144, Vol. 4, pp.285-296 Eurocode 3(1992): ‘Design of Steel Structures -Part 1.1: General Rules and Rules for Buildings’, CEN, April 1992. Fernandes, A.C.(1985), ‘Analysis of a Jackup Platform by Model Testing’, Proc. of the 5th Int. Sym. on Offshore Engineering, Vol.5, 1985 Fernandes, AC, et.al.(1986), ‘Dynamic Behaviour of a Jackup Platform in Waves’, Proc. of the 21st American Towing Tank Conf., 1986


Page No. 46

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

Frieze, P.A., et al.(1991), ‘Report of ISSC Committee V.1 Applied Design’, 11th ISSC, Wuxi, China, Elsevier Applied Science, London 1991 Galambos, V. (1988), Guide to Stability Design Criteria for Metal Structures, Fourth Edition, John Wiley & Sons Gudmestad, O.T.(1990), ‘Refined Modelling of Hydrodynamic Loads on Dynamically Sensitive Structures’, Integrity of Offshore Structures-4, Elsevier Applied Science Publication, pp19-37, July 1990. Gudmestad, O.T., et.al.(1992), ‘Nonlinear Dynamic Response Analysis of Dynamically Sensitive Offshore Structures’, OMEA, 1992. Gudmestad, O.T., and Karunakaran, D.(1994), ‘Wave Kinematics Models for Calculation of wave Loads on Truss Structures’, OTC 7421, Houston 1994. ISO(1995), International Standard ISO/DIS 13819-1, ‘Petroleum and Natural Gas Industries Offshore Structures’, Part 1 : General Requirements, 1995 Jones, D.E., Hoyle, M.J.R., and Bennett, W.T.(1993), ‘The Joint Industry Development of a Recommended Practice for the Site-Specific Assessment of Mobile Jackup Units’ OTC 7306, Houston, 1993 Jostad, H.P., Nadim, F., and Andersen, K.H.,(1994). ‘A Computational Model for Fixity of Spud Cans on Stiff Clay.’, Proc. BOSS’94, Conf. Cambridge, Mass., USA, Vol.1 pp 151-171. Karunakaran, D.N.(1993), ‘Nonlinear Dynamic response and Reliability Analysis of Dragdominated Offshore Platforms’, Dr.Ing. Thesis, NTH, Nov. 1993 Karunakaran, D., et.al.(1993), ‘Prediction of Extreme Dynamic Response of a Jackup using Nonlinear Time Domain Simulations’, OMEA, 1993. Keaveny, J., Nadim, F., and Lacasse, S.(1989). ‘Autocorrelation Functions for Offshore Geotechnical Data.’, Proc 5th ICOSSAR. San Francisco, Cal. USA. pp. 263-270 Langen, I.; and Sigbjørnsson, R.(1979), ‘Dynamisk Analyse av Konstruksjoner’, Tapir Publications, 1979. Leira, B.J., and Karunakaran, D.(1991), ‘Site Dependent Reliability of a Mobile Jackup Platform’, OMAE, 1991. Lotsberg, I., et. al.(1991), ‘Probabilistic Design of a Ship Type Floating Production Vessel’, OMAE Conf., Stavanger, ASME, New York 1991 Løseth, R., Bjerager, P.(1989), ‘ Reliability of Offshore Structures with Uncertain Properties under Multiple Load Processes’, OTC 5969, Houston 1989 Løseth, R., Mo, O., and Lotsberg, I.(1990), ‘Probabilistic Analysis of a Jackup Platform with Respect to the Ultimate Limit State’, Euroms-90, EOMS, Trondheim 1990. Madsen , H.O., Krenk, S., and Lind, N.C.(1986), ‘Methods of Structural Safety’, PrenticeHall Inc., Englewood Cliffs, NJ, 1986.


Page No. 47

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Report No. 95-0072

Manuel, L., Cornell, C.A. (1993), ‘Sensitivity of the Dynamic Response of a Jack-up Rig to Support Modelling and Morison Force Assumptions’, Proc. of the 12th Int. Conf. on Offshore Mech. and Arctic Eng., ASME, Vol2, Jan. 1993. Matheron, G. (1985), ‘Principles of Geostatistics.’, Economic Geology, Vol. 58, pp. 12461266. McClelland, B., Young, A.G., and Remmes B.D.,(1982). ‘Avoiding Jack-Up Rig Foundation Failures.’, Geotechnical Engineering, V 13, No 2, pp. 151-188. Mo, O., Lotsberg, I., Løseth, R.M., (1991) ‘Response Analysis of Jackup Platforms’, Inter. Society of Offshore and Polar Eng. (ISOPE), 1991, Vol.1. Nadim, F., and Lacasse, S. (1992). ‘Probabilistic Bearing Capacity Analysis of Jack-Up Structures.’, Canadian Geotechnical Journal. v 29. No 4. pp. 580-588. Nadim, F., Haver, S., and Mo, O.,(1994). ‘Effects of Load uncertainty on Performance of Jack-Up Foundation.’, Proc. 6th ICOSSAR. Innsbruck, Austria. NPD(1990), ‘Guidelines on Design and Analysis of Steel Structures’, Norwegian Petroleum Directorate, 3rd January 1990. NPD(1992), ‘Guidelines concerning Loads and Load Effects to Regulations concerning Loadbearing Structures in the Petroleum Activities’, Issued by the Norwegian Directorate, 7th Feb. 1992. NS 3472(1984), Norwegian Standard NS 3472 E, ‘Steel Structures, Design Rules’, June 1984. PROBAN (1989) Proban Theory Manual, Det Norske Veritas Research A.S. Report no. 89-2023, 22nd December 1989. PROTIM (1989) ‘Theoretical and Users Manual for PROTIM -Probabilistic Analysis of Time Domain Simulation Results’, Det Norske Veritas Research A.S. Report no. 89-2038, 21st December 1989. Ronold, K.O.(1990), ‘Long Term Reliability of a Jackup Foundation’, Proc. 3rd IFIP Working Conf. on Reliability and Optimisation of Structural Systems, Berkeley, California, 1990. Sarpkaya, T.; and Isaacson, M.(1981), ‘Mechanics of Wave Forces on Offshore Structures’ Van Norstrand Reinhold Publication, 1981. Sesam(1993), ‘Integrated System for Structural Design and Analysis’, Sesam User’s Manual, Sesam System DNV Sesam A.S., 1st January 1993.

Scot Kobus L.C., et al.(1989), ‘Jackup Conversion for Production’, Marine Structures Design, Construction and Safety, Vol.2, 1989.

Jack-up

Recommend Documents