SEDU Dynamic Analysis Procedure PDF

GUIDANCE NOTES ON
DYNAMIC ANALYSIS PROCEDURE FOR SELF-

ELEVATING DRILLING UNITS
JANUARY 2004
American Bureau of Shipping

Incorporated by Act of Legislature of
the State of New York 1862
Copyright  2004
American Bureau of Shipping
ABS Plaza
16855 Northchase Drive
Houston, TX 77060 USA
This Page Intentionally Left Blank
Foreword
The guidance contained herein should be used in conjunction with the ABS Rules for Building and
Classing Mobile Offshore Drilling Units for the purpose of ABS Classification of a Self-Elevating
Drilling Unit. The guidance indicates acceptable practice in a typical case for types of designs that
have been used successfully over many years of service. The guidance may need to be modified to
meet the needs of a particular case, especially when a novel design or application is being assessed.
The guidance should not be considered mandatory, and in no case is this guidance to be considered a
substitute for the professional judgment of the designer or analyst. In case of any doubt about the
application of this guidance ABS should be consulted.
The self-elevating drilling unit is referred to herein as “SEDU”, the ABS Rules for Building and
Classing Mobile Offshore Drilling Units, are referred to as the “MODU Rules”.
ABS welcomes comments about this guidance, especially suggestions for improvement. These can be
sent electronically to [email protected].
ABS GUIDANCE NOTES ON DYNAMIC ANALYSIS PROCEDURE FOR SELF-ELEVATING DRILLING UNITS . 2004 iii
GUIDANCE NOTES ON THE
DYNAMIC ANALYSIS PROCEDURE FOR SELF-
ELEVATING DRILLING UNITS
CONTENTS
SECTION 1 Introduction ............................................................................1
1 Background............................................................................1
3 Basic Concepts of the Inclusion of Dynamic Effects into
Structural Analysis .................................................................1
5 Exception ...............................................................................2
FIGURE 1 The Flowchart of the Two-Step Procedure ..................3
SECTION 2 Specification of Wave Parameters and Spudcan/Soil

Stiffness ..................................................................................5
1 Introduction ............................................................................5
3 Spectral Characterization of Wave Data for Dynamic
Analysis..................................................................................5
5 Spudcan-Soil Rotational Stiffness (SC-S RS) .......................6
SECTION 3 Dynamic Analysis Modeling..................................................7

1 Introduction ............................................................................7
3 Stiffness Modeling..................................................................7
3.1 Leg Stiffness ..................................................................... 7
3.3 Hull Stiffness ..................................................................... 8
3.5 Leg-to-Hull Connection ..................................................... 8
3.7 P-Delta Effect .................................................................... 8
3.9 Foundation Stiffness ......................................................... 9
5 Modeling the Mass.................................................................9
7 Hydrodynamic Loading ..........................................................9
9 Damping...............................................................................10
SECTION 4 Dynamic Response Analysis Methods ..............................11

1 General ................................................................................11
3 Time Domain Analysis .........................................................11
3.1 General ........................................................................... 11
3.3 Random Wave Generation.............................................. 12
ABS GUIDANCE NOTES ON DYNAMIC ANALYSIS PROCEDURE FOR SELF-ELEVATING DRILLING UNITS . 2004 v
3.5 Calculation of Structural Response .................................12
3.7 Prediction of Extreme Responses ...................................14
5 Other Dynamic Analysis Methods........................................17
5.1 Single-Degree-of-Freedom Approach .............................17
5.3 Frequency Domain Dynamic Analysis .............................19
FIGURE 1 Dynamic Amplification Factor (SDOF).......................18
SECTION 5 Application of the Inertial Load Set.................................... 21

1 Random Analysis Approach ................................................21
3 SDOF Approach...................................................................22
5 Evaluation of Structural Adequacy.......................................22
APPENDIX 1 Equivalent Section Stiffness Properties of Lattice

Legs....................................................................................... 23
1 Introduction ..........................................................................23
3 Equivalent Shear Area of 2D Lattice Structures ..................23
5 Equivalent Section Stiffness Properties of 3D Lattice
Legs .....................................................................................24
TABLE 1 Equivalent Shear Area of 2D Lattice Structures ........26

TABLE 2 Equivalent Moment of Inertia Properties of 3D
Lattice Legs................................................................27
FIGURE 1 Shear Force System for X Bracing and its

Equivalent Beam ........................................................23
APPENDIX 2 References............................................................................ 29
vi ABS GUIDANCE NOTES ON DYNAMIC ANALYSIS PROCEDURE FOR SELF-ELEVATING DRILLING UNITS . 2004
SECTION 1 Introduction
1 Background
These Guidance Notes present acceptable practice for an important aspect in the Classification of self-
elevating drilling units (SEDUs). The technical criteria contained in these Guidance Notes are based
on the results of a Joint Industry Project sponsored by Owners, Designers, Builders, Operators and
Classification Societies. The criteria were subsequently published as Reference 1. That reference is
specifically aimed at providing assessment criteria for the site-specific use of the SEDU.
The fundamental difference between site-specific evaluation and Classification is that the latter is not
site-specific in nature. Instead, the Owner specifies conditions for which the unit is to be reviewed for
Classification. The basic dimensions of the envelope of conditions that the Owner may specify for
Classification are:
i) Water depth (plus air gap and penetration depth into the seabed)
ii) Environmental conditions of wind, wave and current
iii) Total elevated load
iv) Spudcan-soil rotational stiffness
(The last item being a consideration introduced by ABS in 2003 when dynamic response is being
assessed for Classification.)
Therefore, a major theme of these Guidance Notes is to clarify the portions of the criteria in Reference
1 that should be applied without modification and the portions of the criteria that may need to be
adapted for Classification purposes.
3 Basic Concepts of the Inclusion of Dynamic Effects into

Structural Analysis
Because the natural period of an SEDU is typically in the range of 5 to 15 seconds, there may be a
concern that there will be dynamic amplification (resonance) with waves in this period range. It is
therefore often desirable to account for the dynamic effects of the SEDU in the elevated condition due
to waves (and waves with current).
The basic approach most commonly used to include dynamic effects into structural analysis is
characterized as a “quasi-static” method, which entails a two-step procedure. In the first step, a
Dynamic Analysis model of the structural system is analyzed. Then, the static response to the same
loads is obtained using the same model. A Dynamic Amplification Factor (DAF) is obtained as the
ratio of the extreme of a response when dynamics is considered to the extreme of the same response
statically considered. DAFs can be quantified for various structural responses, such as the global
overturning moment of the unit, base shear force or the lateral displacement of the elevated hull (i.e.,
surge and sway). From the DAFs, an “inertial load set” is established that simulates the dynamic
effects. The loads considered to produce the dynamic response are those induced by waves or waves
acting with current. Usually, it is sufficient that the level of structural system idealization used to
ABS GUIDANCE NOTES ON DYNAMIC ANALYSIS PROCEDURE FOR SELF-ELEVATING DRILLING UNITS . 2004 1
Section 1 Introduction
determine DAFs is, as often described, an “equivalent model”, which is an “equivalent 3-leg
idealization” coupled with an “equivalent hull structural model”. The need to appropriately account
for the stiffness of the leg-to-hull interaction and foundation/soil interaction adds some minor
complexity to this simplified modeling approach.
In the second step, the “inertial load set” is imposed, along with all of the other coexisting loads, onto
the usual, detailed static structural model that is used to perform the “unity checking” for structural
acceptance based on the Rules. Because this model now includes the “inertial load set” to simulate the
dynamic response, it is often also referred to as the Quasi-Static model.
The two-step procedure is summarized as:
i) Use an “equivalent” model to perform a random wave dynamic analysis deriving the inertial
load set caused by wave-induced structural dynamics and
ii) Use a “detailed” model to perform, with static gravity and wind loads and quasi-static wave
loads plus the derived inertial load set, a static structural analysis deriving the stresses for
unity checks in accordance with the ABS strength requirements in the MODU Rules for the
leg chords, braces and the jacking pinions.
The flowchart of the two-step procedure is shown in Section 1, Figure 1. More details on the
modeling procedure and the determination and application of the inertial load set are given in these
Guidance Notes as follows:
Specification of Wave Parameters and Spudcan/Soil Stiffness Section 2
Dynamic Analysis Modeling Section 3
Dynamic Analysis Methods Section 4
Application of the Inertial Load set Section 5
5 Exception
If the estimated DAF of the unit is smaller than 1.05, the dynamic magnification may be neglected.
The method that may be applied to estimate DAF for this purpose is the SDOF approach indicated in
4/5.1.
2 ABS GUIDANCE NOTES ON DYNAMIC ANALYSIS PROCEDURE FOR SELF-ELEVATING DRILLING UNITS . 2004
Section 1 Introduction
FIGURE 1
The Flowchart of the Two-Step Procedure
Dynamic Response Analysis

Loads Calculation
with "Equivalent" Model
Static Loads Including Dynamic Amplification

Static Wave Loads Factor (DAF)
Static Loads + Inertial Load Set
Inertial Load Set =

F(Static Wave Loads, DAF)
Quasi-static Strucutral Analysis with

Detailed Structural Model
Structural Assessment
SECTION 2 Specification of Wave Parameters
and Spudcan/Soil Stiffness
1 Introduction
Environmental and geotechnical data are inherent to site-specific design and analysis. In the
Classification of a MODU, the environmental actions producing load effects (such as wave, current
and wind) that are used in design are selected by the Owner and become a basis of the unit’s
Classification. It is an assumption of Classification that the Owner will not operate the unit in
environmental and other conditions that produce effects that are worse than those reviewed for
Classification. This principle carries over to the dynamic response assessment.
In Classification, it is usual that the design storm is expressed deterministically, via the parameters
(Hmax, Tass). However, procedures used to explicitly compute dynamic response mostly rely on a
spectral representation of the design-level sea states, so guidance is provided below in Subsection 2/3
on characterizing the design storm sea state in terms of (Hs, Tp) and the defining spectral formulation.
Also in Classification, the MODU Rules have specified that the bottoms of the legs should be assumed
to penetrate to a depth of at least 3 meters below the seabed, and that each leg end (i.e., spudcan) is
pinned (i.e., free to rotate about the axes normal to the leg’s longitudinal axes, but fixed against
displacements). Recently (2003), a change has been introduced in the MODU Rules that affects this
practice. When the Owner wishes to credit spudcan-soil rotational stiffness at the bottom of each leg,
this can be done in a manner as outlined in Subsection 2/5 below.
3 Spectral Characterization of Wave Data for Dynamic

Analysis
The wave conditions provided by the designers for Classification are normally regular waves. The
deterministically defined parameters (Hmax, Tass) of these regular waves need to be converted into
spectral parameters (Hs, Tp) to generate the random waves that will be employed in the dynamic
analysis.
The following procedures should be used to convert the regular waves:
Hsrp = Hmax/1.60 .................................................................................................................. (2.1)
Hs = (1.0 + 0.5 × e(-d/25)) × Hsrp .......................................................................................... (2.2)
Tp = 5.0 × Hs0.5 .................................................................................................................... (2.3)
Section 2 Specification of Wave Parameters and Spudcan/Soil Stiffness
where
Hsrp = significant wave height, in meters, of the three-hour storm for the assessment
return period
Hs = effective significant wave height, in meters
d = water depth, in meters (d > 25 m)
Tp = peak period associated with Hsrp (also used with Hs), in seconds
Equation (2.2) is the correction for the Wheeler stretching, which counts some nonlinear effects
around the free surface. The Pierson-Moskowitz spectrum with the above calculated Hs and Tp is to be
used to represent the wave energy. The short-crestedness of waves should not be considered.
5 Spudcan-Soil Rotational Stiffness (SC-S RS)

A 2003 Rule change to the ABS MODU Rules permits consideration of “spudcan-soil rotational
stiffness” for cases involving dynamic response. The maximum extent to which this rotational
stiffness can be applied to the system, Krs,fixed, is defined by the following equation.
Krs,fixed = E I /(L Cmin)
where
E = Young’s modulus, 209GPa for steel
I = moment of inertia, in m4
L = the sum of the distance, in m, from the underside of the hull to seabed plus the
seabed penetration (minimum 3 meters) ≥ 4.35(I/As)0.5
Cmin = (1.5 – J)/(J + F)
J = 1 + [7.8 I/(As L2)]
F = 12 I Fg/(A Y2)
A = axial area of the equivalent leg, in m2

As = shear area of the leg, in m2
Y = the distance, in m, between the centerline of one leg and a line joining the centers
of the other two legs for a 3-leg unit; the distance, in m, between the centers of
leeward and windward rows of legs; in the direction of being considered
Fg = 1.125 for a three leg unit and 1.0 for a four leg unit
The Owner may select values of SC-S RS ranging from zero (the pinned ended condition) up to the
maximum value indicated.
SECTION 3 Dynamic Analysis Modeling
1 Introduction
To determine a DAF, a simplified Dynamic Analysis model, as indicated below, may be used. The
usual level of modeling employed in this case is designated as an “equivalent model”. Inaccurate or
inappropriate modeling can have a major impact on the calculated structural responses, therefore,
special care should be exercised to assure that the modeling and application of the dynamic loading is
done correctly. The stiffness of the Dynamic Analysis model should also be consistent with that of
the separate Quasi-Static structural model used to check the adequacy of the structure by the
permissible stress unity check criteria of the MODU Rules.
3 Stiffness Modeling
The more significant local contributors to overall structural stiffness should be handled with special
care. These include:
• Leg stiffness
• Leg-to-hull connection (stiffness of jacking system, proper load transfer direction of guides,
pinions and clamps, etc.)
• Leg-to-seabed connection
3.1 Leg Stiffness

The stiffness of a leg is characterized by the following equivalent cross sectional properties:
• Cross sectional area
• Moment of inertia
• Shear area
• Torsional moment of inertia
The dominant factor affecting the system stiffness is leg bending, but other compliance should be
incorporated, such as the shear deflection of legs. The shear deflection of most members is small, but
it can be significant in a braced structure. Therefore, shear deflection of legs should be properly
incorporated in the analysis model.
In an equivalent model, a leg can be modeled by a series of collinear beams. The cross sectional
properties of the beams may be derived by employing the formulas given in Appendix 1 or by
applying various unit loads to the detailed leg model, if available. If the properties are calculated with
the formulas, they may change along the axis of the leg because the properties of the members
constituting the leg may vary along the axis of the leg. Although it is not required to model each bay
of the leg with a beam element, doing this will facilitate a more accurate mass distribution along the
leg.
A spudcan may usually be modeled as a rigid member.
Section 3 Dynamic Analysis Modeling
3.3 Hull Stiffness

Hull structure can be modeled as a grillage of beam members. The properties of the beam may be
calculated based on the depth of the bulkheads and side shell and the effective width of deck and
bottom plating.
The overall structural stiffness or, in turn, the natural period of a unit is less sensitive to hull stiffness.
Therefore, the grillage of beams can simply consist of five beam members, or more complicated, with
several beam members at each location of the bulkheads and side shell. When considering the
contribution from deck(s) and bottom plating, the effective width of deck(s) or bottom plating
assigned to a beam member is so determined that the overlapping plan area reaches minimum, i.e., to
minimize the areas whose contribution is either not included or included twice. This overlapping will
happen when the axes of adjacent beams are not parallel to each other.
The second moment of area of the hull is normally much higher than that of the leg. A common error
is to not make the rotational stiffness and “in plane” bending stiffness of equivalent hull members
high enough.
3.5 Leg-to-Hull Connection

The leg-to-hull connection is very important to the dynamic analysis. The compliance of the
connection is due to a number of factors:
• There may be a global rotation of the leg between the guides due to compliance of the
jacking/holding system.
• There may be a global rotation of the leg between the guides due to the local deflection of the
guide structure.
• Local deflection of the leg chords, induced by the guide reactions, may lead to an effective
rotation of the leg. Also, deformation of the chord wall itself will allow additional leg rotation.
Due to this compliance of the connection, the rotational, horizontal and vertical stiffness of the
connection should be modeled with adequate accuracy. A rigid connection is usually not considered
as acceptable unless the justification of this simplification is provided.
In an equivalent model, the rotational stiffness of the connection may be represented by rotational
springs and horizontal and vertical stiffness by linear springs. The stiffness of the springs may be
derived by employing the formulas given in Reference 1 or by applying various load sets to the
detailed leg-to-hull connection model, provided the detailed model properly simulates the stiffness of
the connection.
3.7 P-Delta Effect

The actual structure will be less stiff than estimated from a linear analysis because of displacement
dependent effects, P-δ or Euler amplification. This will tend to increase the deflection of the
structure, thereby reducing its effective stiffness. Therefore, the P-δ effect should be accounted for in
the Dynamic Analysis model.
A common way to account for the P-δ effect is the geometric stiffness method. In this method,
negative stiffness correction terms are introduced into the global stiffness matrix of the Dynamic
Analysis model. In order to do this, springs of negative stiffness are connected between each spring’s
fixed reaction point and a point on each leg where the hull intersects the leg. The negative stiffness for
horizontal displacements is given by:
Kpd = –Pg/L
where
Pg = total effective gravity load on each leg, including hull weight and weight of the
leg above the hull and leg joint point
L = distance from the spudcan reaction point to the hull vertical center of gravity
3.9 Foundation Stiffness

Additional stiffness to represent the Spudcan-Soil Rotational Stiffness may be included in the model
to the extent indicated in Section 5.
One way to implement this in the equivalent model is for each leg to have a spring of specified
stiffness connected to the reaction point on the leg and an “earth” point where all degrees of freedom
are fixed.
5 Modeling the Mass

The mass that will be dynamically excited and the distribution of that mass should be represented
accurately in the Dynamic Analysis model. Items that should be considered include:
• The elevated mass (arising from hull self-weight; mass of additional equipment, variable mass
from drilling equipment and consumables and other supplies)
• Leg mass, added mass and any entrained and entrapped (water) mass
• Spudcan mass and entrapped (water) mass
Usually, no mass from functional loads will need to be considered as participating in the dynamic
response.
Leg mass can be modeled as nodal mass along the legs. A mass for each bay is adequate for the
dynamic analysis. Added mass and any entrained/entrapped mass should be included. If more accurate
information about mass distribution is not available, elevated weight may be modeled as nodal masses
acting on the hull at its connection to legs.
Generally, the mass distribution of an SEDU, when assessing overall responses, is dominated by the
mass of the hull and its associated equipment. The contribution of the legs below the hull is usually
not great. But for units with free flooding legs, it is important to include both the entrained water and
the added mass.
7 Hydrodynamic Loading
The hydrodynamic loads to be considered in the dynamic analysis are those induced by waves and
waves acting with current. The basis of the hydrodynamic loading is Morison’s equation, as applied
to the Dynamic Analysis model. Equivalent drag and mass coefficients should be developed for the
“equivalent leg” idealization of the leg, and as applicable, the spudcan, etc. The current profile should
be as specified for Classification, with stretching and compression effects as specified in Reference 1.
The hydrodynamic load formulation should consider the relative velocities between the wave and the
structure.
When deriving the hydrodynamic properties, such as equivalent diameter, area, drag and mass
coefficients of a leg, it is important to account for all members, such as chords, horizontal members,
diagonal members, span breakers, etc., in a bay of the leg and their orientations. Some of the
properties, i.e., drag coefficient, are storm-heading-dependent.
Where the dynamic analysis is performed considering sea state simulation using random wave
generation procedures, as described in Section 4, Airy wave theory can be used to develop the
hydrodynamic forces.
9 Damping
Damping can have a significant impact on the response. The total damping ratio to be used in the
dynamic response analysis (expressed as a percentage of the critical damping) is defined as:
ζ = c/ccr · 100 %
where
c = system damping
ccr = critical damping = 2 m ⋅ k
m = effective mass of the system

k = effective stiffness of the system
The damping ratio should not be taken more than 4%. The three main sources of damping are:
• Structural, including holding system, normally taken as 2% maximum on an independent leg
SEDU.
• Foundation, normally taken as 2% maximum for an SEDU with independent legs.
• Hydrodynamic, but since the relative velocity term should be incorporated into either a time
domain or frequency domain dynamic analysis, additional damping to account for hydrodynamic
damping should not be considered.
SECTION 4 Dynamic Response Analysis
Methods
1 General
Fully detailed random dynamic analysis in the time domain or frequency domain may be pursued to
obtain dynamic response. However, the “inertial load set” approach is most often used in practical
design, and yields sufficiently good results in normal circumstances. In this approach, the dynamic
analysis is performed only for determining appropriate values for DAFs. There are several recognized
methods to calculate dynamic response. They range from the very simple to the more complex, and
the recognized methods can give suitable results in the correct circumstances.
There are three basic approaches. The simplest is referred to as the Single Degree of Freedom (SDOF)
Approach. The second and third approaches are collectively based on considering the wave (sea-state)
as a random quantity. Then, separate calculation methodologies are established based on whether the
dynamic response is solved in the time or frequency domain. Due to the limitations of the single-
degree-of-freedom (SDOF) approach and the frequency domain approach (refer to Subsection 4/5) the
random-wave-time-domain approach is the preferred one to be applied in dynamic response analysis,
which will be described in detail in the following Subsection.
Using either a frequency or time domain approach, the most probable maximum extreme (MPME)
value is obtained. The MPME is the mode, or highest point, of the probability density function (PDF)
for the extreme of the response being considered. This is a value with an approximately 63% chance
of exceedance, corresponding to the 1/1000 highest peak level in a sea-state with a 3-hour duration.
There are a number of methods to predict this extreme response, as will be addressed later in this
Section.
3 Time Domain Analysis
3.1 General
The equivalent model indicated in Section 3 is usually employed in time domain analysis. In time
domain simulation, a Gaussian random sea state is generated, and the time-step for the simulation is
required to be sufficiently small. The duration of the simulation(s) should also be sufficiently long for
the method being used to reliably determine the extreme values of the responses being sought.
The overall methodology is to determine the Most Probable Maximum Extreme (MPME) values of
the dynamic and static responses in the time domain. The ratio of these two values – defined as DAF
– represents the ratio by which the Quasi-Static response, obtained using a high order wave theory and
the maximum wave height, should be increased in order to account for dynamic effects. A DAF can
be calculated for each individual global response parameter, e.g., base shear, overturning moment or
hull sway. Usually, DAF of overturning moment is higher than the other two.
Section 4 Dynamic Response Analysis Methods
3.3 Random Wave Generation

The wave elevation may be modeled as a linear random superposition of regular wave components,
using information from the wave spectrum. The statistics of the underlying random process are
Gaussian and fully known theoretically. An empirical modification around the free surface may be
needed to account for free surface effects (Wheeler stretching, Equation 2.2). The following criteria
are to be satisfied for the generated random waves.
3.3.1 Wave Components

The random wave generation should use at least 200 wave components with divisions of equal
wave energy. It is recommended that smaller energy divisions be used in high frequency
regions of the spectrum, where the enforcement and cancellation frequencies are located.
3.3.2 Validity of Generated Sea State

The generated random sea state must be Gaussian and should be checked for validity, as
follows:
• Correct mean wave elevation
• Standard deviation = (Hs/4) ± 1%
• -0.03 < skewness < 0.03
• 2.9 < kurtosis < 3.1
• Maximum crest elevation = (Hs/4) 2 ln( N ) (error within -5% to +7.5%),
where
N = number of wave cycles in the time series being qualified, N ≈ Simulation
Duration/Tz
Tz = zero up-crossing period of the wave
3.3.3 Random Seed Effect

Depending on the method used to predict extreme responses and DAF, the random seed effect
can be significant. Care should be taken to ensure that the predicted results are not affected by
the selection of random seeds.
3.5 Calculation of Structural Response

The structural response should be obtained using the Dynamic Analysis model discussed in Section 3.
The analysis model, i.e., the equivalent model with proper loading and boundary conditions, is to be
solved using a reliable solver having the capability to do time domain calculations and response
statistics calculations. Special attention is to be paid to the topics listed below.
3.5.1 Validity of the Natural Periods of Equivalent Model

The natural periods of a structure are the most important indicators of the dynamic
characteristics of the structure. If the computed natural periods are not reasonable, there must
be something wrong with the established equivalent model, either its stiffness distribution or
its mass distribution, or both. Therefore, the check of natural periods is an indispensable step
in the dynamic analysis.
The natural periods of the established equivalent model can be found by solving the eigen-
value problem, and the fundamental natural period should be checked against that estimated
from the SDOF approach in 4/5.1.
3.5.2 Number of Simulations and Simulation Duration

There are four prevalent methods, as listed in 4/3.7, which can be used to establish the needed
MPME values of the response from the time domain analysis. Each of these extreme value
prediction methods has specific needs regarding the recommended number and duration of the
simulations that should be performed to establish a sufficient statistical basis on which to
obtain the MPME value. Therefore, the recommended number and duration of the simulations
given below should be followed in the calculation of structural response.
i) Drag-Inertia Parameter Method: Simulation time of at least 60 minutes; four
simulations with different control parameters, i.e., fully dynamic, quasi-static, quasi-
static with Cd (drag coefficient) = 0 and quasi-static with CMe (inertia coefficient) = 0.
ii) Weibull Fitting method: Simulation time of at least 60 minutes; number of simulation
≥ 5.
iii) Gumbel Fitting method: Simulation time of at least 180 minutes; number of simulation
≥ 10.
iv) Winterstein/Jensen method: Simulation time of at least 180 minutes; number of
simulation = 1.
More detailed descriptions of these four methods are provided in 4/3.7.
It should be noted that the “quasi-static response analysis” described here and in 4/3.7.1 is
performed using the Dynamic Analysis model, but with the mass and damping terms set to
zero. This analysis is performed to establish DAFs. It should not be confused with the analysis
that is described later with the more detailed, Quasi-Static model that is used to obtain the
“unity- checks”, as described in Section 5.
3.5.3 Time Step of the Simulations

The integration time-step should be less than, or equal to, the smaller of the following
equations, unless it can be shown that a larger time-step leads to no significant change in
results.
Tz/20 or Tn/20
where
Tz = zero up-crossing period of the wave
= Tp/1.406 for the Pierson-Moskowitz spectrum
Tn = first mode natural period of the SEDU
3.5.4 Transients
Transient response is to be discarded by removing the first 100 seconds of the response time
series before predicting the extreme responses.
3.5.5 Relative Velocity

It is expected that the relative velocity between the wave particle and structural velocities will
be included in the hydrodynamic force formulations used in the time domain analysis.
3.7 Prediction of Extreme Responses

Although the waves are considered linear and statistically Gaussian, the structural response of an
SEDU is likely to be non-Gaussian due to non-linear drag force and free surface effects which are
included in the wave kinematics calculations. The statistics of such a non-Gaussian process are
generally not known theoretically, but the extremes are generally larger than the extremes of a
corresponding Gaussian random process. For a detailed investigation of the dynamic behavior of an
SEDU, the non-Gaussian effects should be included. The four prevalent methods elaborated below are
considered acceptable for this purpose.
3.7.1 Method I – Drag/Inertia Parameter Method

The drag/inertia parameter method is based on the assumption that the extreme value of a
standardized process can be calculated by splitting the process into two parts, evaluating the
extreme values of each and the correlation coefficient between the two, then combining as:
(mpmR)2 = (mpmR1)2 + (mpmR2)2 + 2ρR12(mpmR1)(mpmR2)......................................(4.1)
The extreme values of the dynamic response can therefore be estimated from the extreme
values of the quasi-static response, which is obtained by solving the dynamic equation with
both mass term and damping term equal to zero, and the so-called “inertia” response, which is
in fact the difference between the dynamic response and the quasi-static response. The
correlation coefficient of the quasi-static and “inertia” responses is calculated as:
2 2 2
σ Rd − σ Rs − σ Ri
ρR = ................................................................................................(4.2)
2σ Rsσ Ri
The extreme value of the “inertia” response can be reasonably expressed as:
mpmRi = 3.7 σRi .........................................................................................................(4.3)
Reference 1 recommends that the extreme value of the quasi-static response be calculated
using one of the three approaches, as follows:
Approach 1: Static extreme can be estimated by combining the extreme of quasi-static
response to the drag term of Morison’s equation and the extreme of quasi-static response to the
inertia term of Morison’s equation, using Equation (4.1), above. It is also assumed that the
quasi-static responses to the two terms are fully uncorrelated, and hence, the correlation
coefficient ρR12 = 0 when applying Equation (4.1).
Approach 2: Static extreme may be estimated by using a non-Gaussian measure. The structural
responses are nonlinear and non-Gaussian. The degree of non-linearity and the deviation from
a Gaussian process may be measured by the so-called drag-inertia parameter, K, which is a
function of the member hydrodynamic properties and sea state. This parameter is defined as
the ratio of the drag force to inertia force acting on a structural member of unit length.
K = (2CDσV2)/(πCMDσA) ..........................................................................................(4.4)
As an engineering postulate, the probability density function of force per unit length may be
used to predict other structural responses by obtaining an appropriate value of K from time-
domain simulations. K can be estimated from the standard deviation of response due to drag
force only and inertia force only.
π σ R (C M = 0)
K= ..............................................................................................(4.5)
8 σ R (C D = 0)
Approach 3: Alternatively, K can be estimated from the kurtosis of structural response.

1/ 2
  26(κ − 3)  
1/ 2
 (κ − 3) +   
  3  
K = ..............................................................................(4.6)

(35− 3κ ) 

 
3.7.2 Method II – Weibull Fitting

Weibull fitting is based on the assumption that for a drag dominated structure, the cumulative
distribution of the maxima of the structural response can be fitted to a Weibull class of
distribution:
  R − γ β 
FR = 1 − exp −    ......................................................................................(4.7)
  α  
The extreme value for a specified exceedance probability (e.g., 1/N) can therefore be
calculated as:
R = γ + α[–ln(1 – FR)]1/β ...........................................................................................(4.8)
Using a uniform level of exceedance probability of 1/N, Equation (4.8) leads to
RMPME = γ + α[–ln(1/N)]1/β .......................................................................................(4.9)
The key issue for using this method is therefore to calculate the parameters α, β and γ, which
regression analysis, maximum likelihood estimation or static moment fitting can estimate. For
a 3-hour storm simulation, N is approximately 1000. The time series record is first
standardized (R* = (R – µ)/σ), and all positive peaks are then sorted in ascending order.
As recommended in Reference 1, only peaks corresponding to a probability of non-exceedance
greater than 0.2 are to be used in the curve fitting, and least square regression analysis is used
for estimating Weibull parameters.
3.7.3 Method III – Gumbel Fitting

The Gumbel fitting method is based on the assumption that the three-hour extreme values
follow the Gumbel distribution:
  1 
F ( x extreme ≤ X MPME ) = exp − exp − ( X MPME − ψ )  ......................................(4.11)
  κ 
The most probable maximum extreme discussed here corresponds to an exceedance
probability of 1/1000 in a distribution function of individual peaks or to 0.63 in an extreme
probability distribution function. The MPME of the response can therefore be calculated as:
XMPME = ψ –κ ln{–ln[F(XMPME)]}
= ψ – κ ln[–ln(0.37)] ≈ ψ ...........................................................................(4.12)
Now, the key issue is to estimate the parameters ψ and κ based on the response obtained from
time-domain simulations. Reference 1 recommends that the maximum simulated value be
extracted for each of the ten 3-hour response simulations, and that the parameters be computed
by maximum likelihood estimation. Similar calculations should also be performed using the
ten 3-hour minimum values. Although it is always possible to apply the maximum likelihood
fit numerically, the method of moments may be preferred.
For the Gumbel distribution, the mean and variance are given by
Mean: µ = ψ + γ⋅κ, γ = Euler constant (0.5772…)
Variance: σ2 = π2κ2 / 6
By which means, the parameters ψ and κ can be directly obtained using the moment fitting
method:
6σ
κ= , ψ = µ – 0.57722κ .................................................................................(4.13)
π
3.7.4 Method IV – Winterstein/Jensen Method

The basic premise of Winterstein/Jensen method is that a non-Gaussian process can be
expressed as a polynomial (e.g., a power series or an orthogonal polynomial) of a zero mean,
narrow-banded Gaussian process (represented here by the symbol U), that is
R(U) = C0 + C1U + C2U2 +C3U3 ............................................................................(4.14)
The same relationship exists between the MPMEs of the two processes. Since the MPME of
Gaussian process U is theoretically known, the MPME of the non-Gaussian process can be
calculated if the coefficients C0, C1, C2, C3 are determined.
3.7.4(a) Determination of Um.. Calculate the following statistical quantities of the time series
for the response parameter R under consideration:
µR = mean of the process
σR = standard deviation
α3 = skewness
α4 = kurtosis
Then construct a standardized response process, z = (R − µR)/σR. Using this standardized
process, calculate the number of zero-upcrossings, N. In lieu of an actual cycle count from the
simulated time series, N = 1000 may be assumed for a 3-hour simulation.
The most probable value, Um, of the transformed process is computed by the following
equation:
 3 hours 
Um = 2 log e  N ⋅  ......................................................(4.15)
 simulation time (in hours ) 
where Um is the most probable value of a Gaussian process of zero mean, unit variance.
3.7.4(b) Determination of C coefficients Using Equation (4.14), one can establish the
following equations for C1, C2 and C3:
σR2 = C12 + 6C1C3 + 2C22 + 15C32
σR3α3 = C2(6C12 + 8C22 + 72C1C3 + 270C32)
σR4α4 = 60C24 + 3C14 + 10395C34 + 60C12C22 + 4500C22C32 + 630C12C32 +

936C1C22C3 + 3780C1C33 + 60C13C3
Solve the equations with the initial guesses as:

C1 = σRK(1 – 3h4)
C2 = σRKh3
C3 = σRKh4
where
h3 = α 3 /[ 4 + 2 {1 + 1.5(α 4 − 3)}]
h4 = [ {1 + 1.5(α 4 ]
− 3)} − 1 / 18
K = [1 + 2h32 + 6h42]-1/2
Obtain
C0 = µR – σRKh3
3.7.4(b) Determination of RMPME. The most probable maximum extreme in a 3-hour storm,
for the response under consideration, can be computed from the following equation:
RMPME = C0 + C1Um1 + C2Um2 + C3Um3 .................................................................(4.16)
5 Other Dynamic Analysis Methods

The random wave time domain method is the recommended approach for the dynamic analysis of an
SEDU. However, the analysis procedure is relatively complicated and under some circumstances,
other methods can also generate results of sufficient accuracy. Besides, some results obtained from the
simpler methods, e.g., natural period of the structure determined by SDOF approach, can be used to
check the results of time domain analysis. For these reasons, the single degree of freedom approach
and frequency domain method are briefly discussed below.
5.1 Single-Degree-of-Freedom Approach

In a single-degree-of-freedom (SDOF) approach, the SEDU is modeled as a simple mass/spring/
damper system. Due to its simplicity, this approach is recommended for an initial evaluation of the
dynamic amplification. If the dynamic amplification determined with this approach is relatively small
(i.e., DAF < 1.05), then random dynamic analysis is not required.
5.1.1 Natural Period

The natural period of an SEDU is an important indicator of the degree of dynamic response to
be expected. The first and second vibratory modes are usually surge and sway (i.e. lateral
displacements at the deck level). The natural periods of these two modes are usually close to
each other. Which of the two is higher depends on which direction of the structure is less stiff.
The third vibratory mode is normally a torsional mode. Since the period varies with the
environmental load direction, care should be taken that the period used in analysis is consistent
with the environmental load being considered.
An estimate of the first mode (fundamental) natural period, Tn, is obtained for a single-degree-
of-freedom (SDOF) system, as follows:
1 Me
Tn = = 2π
f Ke
where
f = natural frequency
Me = effective mass associated with one leg
Ke = effective stiffness associated with one leg, which suitably accounts for the
bending, shear and axial stiffness of each leg, the stiffness of the hull-to-
leg connection and the degree of spudcan-soil rotational restraint that is to
be considered
The detailed information for the calculations of Me and Ke can be found in Reference 1.
5.1.2 Calculation of the SDOF DAF

The Dynamic Amplification Factor (DAF) of a SDOF system under the influence of a
sinusoidal (monotonic) forcing function is given by the following formula:
1
DAF =
(1 − Ω )
2 2
+ (2ζΩ )2
where
Tn Natural period of the jackup
Ω = =
T Period of the applied load (wave period)
T = 0.9 Tp
ζ = damping ratio
As illustrated in Section 4, Figure 1, if the natural period of the SEDU is equal to the period of
the applied load (i.e., Ω is equal to 1.0), the DAF becomes just over 7 (when a damping ratio
of 7% is used). Conversely, if there is a very large separation between the natural period and
the load period, the DAF could be underestimated. An actual sea state can have a significant
spread of energy over the period range, and the curve of DAF against Ω is likely to be much
more shallow than that predicted by the SDOF model. This is also illustrated in Section 4,
Figure 1.
FIGURE 1
Dynamic Amplification Factor (SDOF)
ζ = 0.07
6 SDOF DAF
DAF
4 DAF in Random Seas
0
0 long period wave 1 short period wave 2
Omega (T n /T )
Care should be taken when determining the appropriate wave period to be used in an SDOF
analysis. A range of wave periods should be investigated, along with a range of associated
wave heights. The applicable sea states that result in maximum responses should be identified
and used in the assessment of the adequacy of the structure’s strength.
5.1.3 Dynamic Load Application

The dynamic effect can be applied to the Quasi-Static model by applying an extra force
representing the dynamically-induced inertial load set at the center of gravity of the hull
structure. The procedure is presented in Section 5.
5.1.4 Limitations
The greatest problems with the SDOF approach are that it may grossly over-estimate the
response when the natural period of the unit is close to the monotonic period of the applied
load and may possibly underestimate the response when there are large differences in periods.
However, this method can give reasonable results when Ω is not too close to unity (outside the
range of 0.6 to ~1.3).
5.3 Frequency Domain Dynamic Analysis

A frequency domain analysis combines a spectral representation of the wave energy with a transfer
function that defines how the SEDU responds to unit wave heights of different periods, and arrives at
the statistical responses in an irregular sea state. In the mathematics of “combining”, it is inherently
assumed that the wave force (or moment, etc.) on the unit is monotonic at each wave period, and that
the responses are linear with wave height. Special care is then needed to determine the appropriate
transfer functions to use, and how to produce reasonable results for what is usually a non-monotonic
and non-linear loading.
When a frequency domain analysis is pursued, it should be demonstrated that the potential errors
arising from the following items are suitably accounted for.
• The wave forces are evaluated using linear kinematics up to the mean water level only, therefore,
the wave forces in wave crests may be underestimated.
• Assuming that the responses are linear with wave height. On a drag dominant unit, the wave
forces are more closely proportional to the square of the wave height rather than being linear with
wave height.
• Failing to account for the higher frequency harmonics in the wave force due to its non-linearity.
For example, the force due to a 12 second wave on a drag dominant structure will have a
significant component of force at 6 and 4 seconds.
• Improperly interpreting the result and predicting realistic maximum responses from the mean and
standard deviation responses that are produced by the analysis.
Since the statistics, e.g., mean and standard deviation, of a response can be obtained from the
frequency domain analysis, Method I – Drag/Inertia Parameter Method, indicated in 4/3.7.1, may be
employed to predict the MPME, and therefore, DAF of the response.
SECTION 5 Application of the Inertial Load Set
1 Random Analysis Approach

Once the DAFs are obtained, from either the time domain method or frequency domain method, for
the response quantities of interest, they can be used to establish an inertial load set that is imposed on
the detailed model of Quasi-Static structural analysis to simulate the dynamic response. Usually, the
DAFs for Overturning Moment (OTM) and Base Shear (BS) force are obtained. The inertial load set
is combined with all of the other statically considered loads, such as those from wind, currents,
deterministically considered storm wave, weights, functional loads, etc. After the strength analysis of
the Quasi-Static model is performed, the permissible stress checks, in accordance with the MODU
Rules unity check equations, can be performed.
A commonly accepted way that the inertial load set is included in the Quasi-Static model is as a
concentrated load applied to the elevated hull structure. This idealization is most suitable for the case
where the preponderance of the structural system’s total mass is in the hull, which is usually
considered to be the case. If it were not the case, the complexity of the inertial load set would increase
so that instead of a single concentrated load, the inertial load set should be distributed in accordance
with the mass distribution and vibratory mode shapes.
The magnitude of the concentrated load representing the dynamic response from waves (or waves
acting with current) in the wave loading direction can be obtained from the following quantities:
d = vertical distance from the base of a leg to a location in the elevated hull structure
where the concentrated inertial load is to be imposed.
DAFOTM = dynamic amplification factor for overturning moment obtained from the Dynamic
Response analysis using the MPME values for the dynamic and statically
considered simulated hydrodynamic loads on the unit
DAFBS = dynamic amplification factor for the base shear force obtained from the Dynamic
Response analysis using the MPME values for the dynamic and statically
considered simulated hydrodynamic loads on the unit
OTMQS = maximum, deterministic overturning moment from the considered wave (or wave
acting with current) on the Quasi-Static structural model before the imposition of
the inertial load set.
BSQS = maximum, deterministic shear force from the considered wave (or wave acting
with current) on the Quasi-Static structural model before the imposition of the
inertial load set.
The magnitude of the concentrated inertial force, FI, is then found from the following equation:
FI = (DAFOTM – 1) OTMQS /d
Section 5 Application of the Inertial Load Set
It should also be checked that the increased total base shear force from the imposition of the inertial
load set is approximately equal (±5%) to the dynamically amplified base shear force, BSQS*, from the
deterministic, Quasi-Static model. This check is expressed as follows:
The increased base shear from the addition of inertial force FI is:
BSQS + FI
and the dynamically amplified base shear is:
BSQS* = BSQS (DAFBS – 1)
then
0.95 BSQS* < (BSQS + FI) <1.05 BSQS*
If this check is not satisfied, various strategies may be pursued to better balance the dynamic effects on
the wave-induced overturning moment and base shear force. These may include judiciously adjusting
the distance d; distributing the inertial load set according to the mass distribution and vibratory mode
shapes, etc. However, when (BSQS + FI) > BSQS* and no adjustment is made, the structural assessment
will tend to be conservative.
3 SDOF Approach
When the approach presented in 4/5.1 is applied, the procedure that may be followed to establish the
inertial load set is as follows.
The magnitude of the force is determined from:
Fi = (DAF − 1) × Fwave amp
where
Fi = inertial load set to be applied at the center of gravity of the hull and associated leg
and in the direction of waves (i.e., that contained in, and above, the lower guide)
DAF = SDOF dynamic amplification factor
Fwave amp = static amplitude wave force = 0.5(Fmax − Fmin)
Fmax, Fmin = maximum/minimum total combined wave and current force (or wave/current base
shear) obtained from quasi-static structural analysis, using the appropriate sea
state
5 Evaluation of Structural Adequacy

The inertial load set is combined with all of the other loads that should be included in the detailed
Quasi-Static structural analysis model that is used to obtain the stresses and deflections that are
evaluated with respect to the acceptance criteria given in the MODU Rules.
APPENDIX 1 Equivalent Section Stiffness
Properties of Lattice Legs
1 Introduction
In order to evaluate the equivalent section stiffness properties of 3D lattice legs, it is necessary first to
identify the equivalent shear area of 2D lattice structures, which composes each wall of the 3D lattice
legs and the equivalent polar moment of inertia of the 3D lattice leg’s cross-section. The equivalent
shear area uses the equivalent 2D lattice shear area of the structure.
3 Equivalent Shear Area of 2D Lattice Structures

The equivalent shear area of 2D lattice structures is evaluated by the principle of virtual work, as
indicated in Reference 4. For example, Appendix 1, Figure 1 shows that the strain energy of the shear
beam deformation is made equivalent to the complimentary virtual work in the X bracing system.
FIGURE 1
Shear Force System for X Bracing and its Equivalent Beam
s V s V
( ) ( )
h 2 h 2 Vs
0
V V V
2 2
0 −C s s
C 0
V V
2 2 V
Vs
s V s V
( ) ( )
h 2 h 2
h
Appendix 1 Equivalent Section Stiffness Properties of Lattice Legs
The forces in the diagonals are ±C = ±(V/2)d/h, where d is the diagonal length, and the corresponding
complementary energy for the 2D lattice truss is:
1 Fi 2 Li 1 C 2d 1 V 2d 3
W* =
2
∑ EAi
= 2(
2 EAD
)=
4 h 2 EAD
.......................................................................(A1.1)
i
where Fi , Li, Ai are the force, length and area of the i-th member, and E is the modulus of elasticity.
According to the principle of virtual forces, one obtains
V ∂W * 1 Vd 3
s = = ................................................................................................(A1.2)
GAQ ∂V 2 h 2 EAD
where
G = E/[2(1 + ν)]
AQ = equivalent shear area of the structure
Then
(1 + ν )h 2 s
AQ = .................................................................................................................(A1.3)
d3
4 AD
The formulae for four types of 2D lattice structures commonly employed in constructing the legs of an
SEDU are derived and listed in Appendix 1, Table 1. The shear areas calculated by these formulae are
very close to the results formulae presented in Reference 1 for typical SEDUs in operation.
5 Equivalent Section Stiffness Properties of 3D Lattice Legs

The equivalent section stiffness properties of 3D lattice legs are obtained as follows:
i) The cross-sectional area of a leg is the summation of the cross-sectional areas of all of the
chords in the leg. The contribution from the braces is neglected.
ii) The shear area of a leg’s cross-section in k direction (i.e., y or z direction) can be expressed as
N
AQk = ∑ AQ sin 2 β i
i =1
where
AQ = equivalent shear area of 2D lattice structure
βi = angle between k direction and the normal direction of the i-th 2D lattice
structure
N = total number of the 2D lattice structures in the leg (i.e., 3 or 4)
iii) The moment of inertia of the leg’s cross-section for k direction (i.e., y or z direction) is the
summation of the cross-sectional area of a chord times the square of the distance from the
chord center to the neutral axis of the leg’s cross-section in k direction for all chords. The
contribution from the braces is neglected.
iv) The polar moment of inertia of the leg’s cross-section is

N
IT = ∑ AQ l i 2
i =1
where
AQ = equivalent shear area of 2D lattice structure
li = distance from the i-th 2D lattice structure to the geometry center of the
leg’s cross-section
N = total number of the 2D lattice structures in the leg (i.e., 3 or 4)
Appendix 1, Table 2 presents the equivalent beam moment of inertial, which when multiplied by the
modulus of elasticity provides the section stiffness properties of three types of leg configurations.
TABLE 1
Equivalent Shear Area of 2D Lattice Structures
Structure Equivalent Shear Area
s AC
AV
(1 + ν )h 2 s
AQ =
d AD h d3 h3 s3
+ +
AD 8 AV 4 AC
AC s
AD
(1 + ν )h 2 s
AQ =
AV h d3
d 4 AD
s AC
AV
AD (1 + ν )h 2 s
AQ =
h d3 s3
d +
2 AD AC
s AC
AD (1 + ν )h 2 s
AQ =
h d3 h3 s3
d + +
AV 2 AD 2 AV 4 AC
Note:
ν = Poisson ratio
Ak = cross sectional area of the corresponding member (k = C, D or V)
TABLE 2
Equivalent Moment of Inertia Properties of 3D Lattice Legs
Leg Configuration Equivalent Section Stiffness Properties
z
AC
AQ
A = 3 AC
h AQy = AQz = 3 AQ / 2
I y = I z = AC h 2 / 2
o
y
I T = AQ h 2 / 4
AQ
A = 4 AC
h AQy = AQz = 2 AQ
y
o
I y = I z = AC h 2
AC
I T = AQ h 2
AQ
h
A = 4 AC
AQy = AQz = 2 AQ
y
o I y = I z = AC h 2
AC I T = AQ h 2
h
APPENDIX 2 References
1. Guidelines for Site Specific Assessment of Mobile Jack-Up Units, Technical & Research
Bulletin 5-5A, SNAME, Rev. 2, Jan. 2002
2. Rules for Building and Classing Mobile Offshore Drilling Units, Part 3, ABS, 2001
3. Dynamics of Structures, Second Edition, R. Clough and J. Penzien, 1993
4. Stochastic Response of Fixed Offshore Structures in Random Sea, Research Report R78-37,
Department of Civil Engineering, MIT, 1978
5. Continuum Models for Beam and Platelike Lattice Structures, Anne Teughels, IASS-IACM
2000, Fourth International Colloquium on Computation of Shell and Spatial Structures,
Chania – Crete, Greece, June 5-7, 2000

SEDU Dynamic Analysis Procedure PDF

Uploaded by

Copyright:

Available Formats

SEDU Dynamic Analysis Procedure PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

SEDU Dynamic Analysis Procedure PDF

Uploaded by

Copyright:

Available Formats

GUIDANCE NOTES ON

DYNAMIC ANALYSIS PROCEDURE FOR SELF-

American Bureau of Shipping

FIGURE 1 The Flowchart of the Two-Step Procedure ..................3

SECTION 2 Specification of Wave Parameters and Spudcan/Soil

SECTION 3 Dynamic Analysis Modeling..................................................7

SECTION 4 Dynamic Response Analysis Methods ..............................11

FIGURE 1 Dynamic Amplification Factor (SDOF).......................18

SECTION 5 Application of the Inertial Load Set.................................... 21

APPENDIX 1 Equivalent Section Stiffness Properties of Lattice

TABLE 1 Equivalent Shear Area of 2D Lattice Structures ........26

FIGURE 1 Shear Force System for X Bracing and its

3 Basic Concepts of the Inclusion of Dynamic Effects into

Dynamic Response Analysis

Static Loads Including Dynamic Amplification

Static Loads + Inertial Load Set

Inertial Load Set =

Quasi-static Strucutral Analysis with

3 Spectral Characterization of Wave Data for Dynamic

Hs = (1.0 + 0.5 × e(-d/25)) × Hsrp .......................................................................................... (2.2)

Tp = 5.0 × Hs0.5 .................................................................................................................... (2.3)

5 Spudcan-Soil Rotational Stiffness (SC-S RS)

J = 1 + [7.8 I/(As L2)]

A = axial area of the equivalent leg, in m2

3.1 Leg Stiffness

3.3 Hull Stiffness

3.5 Leg-to-Hull Connection

3.7 P-Delta Effect

3.9 Foundation Stiffness

5 Modeling the Mass

m = effective mass of the system

3 Time Domain Analysis

3.3 Random Wave Generation

3.3.1 Wave Components

3.3.2 Validity of Generated Sea State

3.3.3 Random Seed Effect

3.5 Calculation of Structural Response

3.5.1 Validity of the Natural Periods of Equivalent Model

3.5.2 Number of Simulations and Simulation Duration

3.5.3 Time Step of the Simulations

3.5.5 Relative Velocity

3.7 Prediction of Extreme Responses

3.7.1 Method I – Drag/Inertia Parameter Method

Approach 3: Alternatively, K can be estimated from the kurtosis of structural response.

3.7.2 Method II – Weibull Fitting

3.7.3 Method III – Gumbel Fitting

3.7.4 Method IV – Winterstein/Jensen Method

σR2 = C12 + 6C1C3 + 2C22 + 15C32

σR3α3 = C2(6C12 + 8C22 + 72C1C3 + 270C32)

σR4α4 = 60C24 + 3C14 + 10395C34 + 60C12C22 + 4500C22C32 + 630C12C32 +

Solve the equations with the initial guesses as:

5 Other Dynamic Analysis Methods

5.1 Single-Degree-of-Freedom Approach

5.1.1 Natural Period

5.1.2 Calculation of the SDOF DAF

4 DAF in Random Seas

5.1.3 Dynamic Load Application

5.3 Frequency Domain Dynamic Analysis

1 Random Analysis Approach