The electronic pdf versions of the documents found through http://www.dnv.com/ are the officially binding versions. Copyright Det Norske Veritas.
![[Table of Contents]](/dwicons/b_toc.gif)
|
DNV-OS-C105 Structural Design of TLPS (LRFD method)
|
SECTION 5
Global PerformanceSec.5
A. Introduction
Sec.5
A 100 General
Sec.5 A
101 The selected methods of response analysis are dependent on
the design conditions, dynamic characteristics, non-linearities
in loads and response and the required accuracy in the actual design
phase.Guidance note:
For a detailed discussion of the different applicable methods
for global analysis of tension leg platforms, see API RP 2T.---e-n-d---o-f---G-u-i-d-a-n-c-e---n-o-t-e---
Sec.5 A
102 The selected methods of analysis and models employed in the
analysis shall include relevant non-linearities and motion-coupling
effects. The approximations, simplifications and/or assumptions
made in the analysis shall be justified, and their possible effects
shall be quantified e.g. by means of simplified parametric studies.
Sec.5 A
103 During the design process, the methodology and software used
for analytical or numerical prediction of important system responses
shall be verified (calibrated) by appropriate model tests.
Sec.5 A
104 Model tests may also be used to determine specific responses
for which numerical or analytical procedures are not yet fully developed
and recognised.
Sec.5 A
105 Motion components shall be determined, by relevant analysis
techniques, for those applicable design conditions (design analyses
matrix) specified in DNV-OS-C101. The basic assumptions and limitations associated
with the different methods of analysis of global performance shall
be duly considered prior to the selection of the methods.
Typically
a combination of frequency domain and time domain analyses will
be applied by the designers.
Sec.5 A
106 The TLP should be analysed by methods as applicable to column-stabilised
units or deep draught floaters when the unit is free floating, respectively
see DNV-OS-C103 or DNV-OS-C106.
Sec.5 A
107 The method of global performance analysis as outlined in this
standard is one approximate method that may be applied. The designer
is encouraged also to consider and apply other methods in order
to discover the effects of possible inaccuracies etc. in the different
methods.Sec.5
B. Frequency Domain Analysis
Sec.5
B 100 General
Sec.5 B
101 Frequency domain high frequency (HF), wave frequency (WF)
and low frequency (LF) analyses techniques may be applied for a
TLP. Regarding load effects due to mean wind, current and mean wave
drift, see DNV-OS-C101.
Sec.5 B
102 For typical TLP geometries and tendon arrangements, the analysis
of the total dynamic load effects may be carried out as:
| — | a HF analysis of springing |
| — | a WF analysis in all six degrees of freedom |
| — | a LF analysis in surge, sway and yaw. |
Sec.5 B
103 The following assumptions are inherent in adopting such an
independent analysis approach:| — | the natural frequencies in heave,
roll and pitch are included in the wave frequency analysis |
| — | the natural frequencies in surge, sway and yaw are included
in the low frequency analysis |
| — | the high and low natural frequencies are sufficient
separate to allow independent dynamic analysis to be carried out |
| — | the low frequency excitation forces have negligible
effect on the wave frequency motions |
| — | the low frequency excitation forces have a negligible
dynamic effect in heave, roll and pitch |
| — | tendon lateral dynamics are unimportant for platform
surge or sway motions. |
Sec.5 B
104 Typical parameters to be considered for global performance
analyses are different TLP draughts, wave conditions and headings,
tidal effects, storm surges, set down, foundation settlement(s),
subsidence, mispositioning, tolerances, tendon flooding, tendon
removal and hull compartment(s) flooding. Possible variations in
vertical centre of gravity shall also be analysed (especially if
ringing responses are important). This may be relevant in case of:
| — | change in operation mode (e.g.
drilling/production) |
| — | changes in topside weights (e.g. future modules) |
| — | tendon system changes (altered utilisation) |
| — | changes in ballast weights or distributions |
| — | deviations from weight estimate |
| — | riser phasing scenarios |
| — | lateral positioning. |
Sec.5
B 200 High frequency analyses
Sec.5 B
201 Frequency domain springing analyses shall be performed to
evaluate tendon and TLP susceptibility to springing responses.
Sec.5 B
202 Recognised analytical methods exist for determination of springing
responses in tendons. These methods include calculation of Quadratic
Transfer Functions (QTF's) for axial tendon (due to sum
frequency loads on the hull) stresses which is the basis for determination
of tendon fatigue due to springing.
Sec.5 B
203 Total damping level applied in the springing response analyses
shall be duly considered and documented.Sec.5
B 300 Wave frequency analyses
Sec.5 B
301 A wave frequency dynamic analysis may normally be carried
out by using linear wave theory in order to determine first-order
platform motions and tendon response.
Sec.5 B
302 First order wave load analyses shall also serve as basis for
structural response analyses. Finite wave load effects shall be
evaluated and taken into account. This may e.g. be performed by
use of beam models and application of Morison load formulation and
finite amplitude waves.
Sec.5 B
303 In linear theory, the response in regular waves (transfer
functions) is combined with a wave spectrum to predict the response
in irregular seas.
Sec.5 B
304 The effect of low-frequency set-down variations on the WF
analysis is to be investigated by analysing at least two representative
mean offset positions determined from the low-frequency analysis.
Sec.5 B
305 Set-down or offset induced heave motion may be included in
the wave frequency response amplitude operators (RAOs).
Sec.5 B
306 A sufficient number of wave approach headings shall be selected
for analyses (e.g. with basis in global configuration, number of
columns, riser configuration etc.).
Sec.5 B
307 In determination of yaw induced fatigue responses (e.g. tendon
and flex element design) due account must be given to wave spreading
when calculating the long term responses.Sec.5
B 400 Low frequency analyses
Sec.5 B
401 A low frequency dynamic analysis could be performed to determine
the slow drift effects at early design stages due to fluctuating
wind and second order wave loads.
Sec.5 B
402 Appropriate methods of analysis shall be used with selection
of realistic damping levels. Damping coefficients for low frequency
motion analyses are important as the low frequency motion may be
dominated by resonant responses.Sec.5
C. Time Domain Analyses
Sec.5
C 100 General
Sec.5 C
101 For global motion response analyses, a time domain approach
will be beneficial. In this type of analyses it is possible to include
all environmental load effects and typical non-linear effects such
as:| — | hull drag forces (including
relative velocities) |
| — | finite wave amplitude effects |
| — | non-linear restoring (tendons, risers). |
Sec.5 C
102 Highly non-linear effects such as ringing may also require
a time domain analysis approach. Analytical methods exist for estimation
of ringing responses. These methods may be used for the early design
stage, but shall be correlated against model tests for the final
design. Ringing and springing responses of hull and deck may however
be analysed within the frequency domain with basis in model test
results, or equivalent analytical results.
Sec.5 C
103 For deep waters, a fully coupled time domain analysis of tendons,
risers and platform may be required. This may e.g. be relevant if:| — | model basin scale will not be
suitable to produce reliable design results or information |
| — | consistent global damping levels (e.g. in surge, sway
and yaw) due to the presence of slender structures (risers, tendons)
are needed |
| — | it is desirable to perform the slender structure response
analyses with basis in coupled motion analyses. |
Sec.5 C
104 A relevant wave spectrum shall be used to generate random
time series when simulating irregular wave elevations and kinematics.
Sec.5 C
105 The simulation length shall be long enough to obtain sufficient
number of LF maxima (surge, sway and yaw).
Sec.5 C
106 Statistical convergence shall be checked by performing sensitivity
analyses where parameters as input seed, simulation length, time
step, solution technique etc. are varied.
Sec.5 C
107 Determination of extreme responses from time domain analyses
shall be performed according to recognised principles.
Sec.5 C
108 Depending on selected TLP installation method, time domain
analyses will probably be required to simulate the situation when
the TLP is transferred from a free floating mode to the vertical
restrained mode. Model testing shall also be considered in this
context.Guidance note:
Combined loading
Common practice to determine extreme responses has been to
expose the dynamic system to multiple stationary design environmental
conditions. Each design condition is then described in terms of
a limited number of environmental parameters (e.g. Hs, Tp) and a
given seastate duration (3 to 6 hours). Different combinations of
wind, wave and current with nearly the same return period for the
combined environmental condition are typically applied.
The main problem related to design criteria based on environmental
statistics is that the return period for the characteristic load
effect is unknown for non-linear dynamic systems. This will in general
lead to an inconsistent safety level for different design concepts
and failure modes.
A more consistent approach (as required in API RP 2T March,
2010 edition) is to apply design based on response statistics. Consistent
assessment of the D-year
load effect will require a probabilistic response description due
to the long-term environmental loads on the system. The load effect
with a return period of D-year,
denoted xD, can formally be
found from the long-term load effect distribution as:
| ND | = | total number of load effect maxima during D years | FX(c) | = | long-term peak distribution of the (generalised)
load effect | |
The main challenge related to this approach is to establish
the long-term load effect distribution due to the non-linear behaviour.
Design based on response statistics is in general the recommended
procedure and should be considered whenever practicable for consistent
assessment of characteristic load effects.
Further details may be found in Appendices to DNV-OS-F201.
For guidance on coupled analysis, see DNV-RP-F205.---e-n-d---o-f---G-u-i-d-a-n-c-e---n-o-t-e---
Sec.5
D. Model Testing
Sec.5
D 100 General
Sec.5 D
101 Model testing will usually be required for final check of
TLP designs. The main reason for model testing is to check that
analytical results correlate with model tests.
Sec.5 D
102 The most important parameters to evaluate are:| — | air-gap |
| — | first order motions |
| — | total offset |
| — | set-down |
| — | WF motions versus LF motions |
| — | tendon responses (maximum and minimum) |
| — | accelerations |
| — | ringing |
| — | springing |
| — | susceptibility to hull VIM. |
Sec.5 D
103 The model scale applied in testing shall be appropriate such
that reliable results can be expected. A sufficient number of seastates
need to be calibrated covering the relevant limit states.
Sec.5 D
104 Wave headings, multidirectional sea, tests with wind, wave
and current, wave steepness and other variable parameters (water
levels, vertical centre of gravity, etc.) need to be varied and
tested as required.
Sec.5 D
105 If HF responses (ringing and springing) shows to be governing
for tendon extreme and fatigue design respectively, the amount of
testing may have to be increased to obtain confidence in results.Sec.5
E. Load Effects in the Tendons
Sec.5
E 100 General
Sec.5 E
101 Load effects in the tendons comprise mean and dynamic components.
Sec.5 E
102 The steady state loads may be determined from the equilibrium
condition of the platform, tendon and risers.
Sec.5 E
103 Tendon dynamic load effects arise from platform motions, any
ground motions and direct hydrodynamic loads on the tendon.
Sec.5 E
104 Dynamic analysis of tendon responses shall take into account
the possibility of platform heave, roll and pitch excitation (springing
and ringing effects).
Sec.5 E
105 Linearised dynamic analysis does not include some of the secondary
wave effects, and may not model accurately extreme wave responses.
A check of linear analysis results using non-linear methods may
be necessary. Model testing may also be used to confirm analytical
results. Care shall be exercised in interpreting model-test results
for resonant responses, particularly for loads due to platform heave,
roll and pitch, since damping may not be accurately modelled.
Sec.5 E
106 Lift and overturning moment generated on the TLP by wind loads
shall be included in the tendon response calculations.
Sec.5 E
107 Susceptibility to vortex induced vibrations shall be evaluated
in operational and non-operational phases.
Sec.5 E
108 Interference (tendon/riser, tendon/tendon,
tendon/hull, and tendon/foundation) shall be evaluated
for non-operational as well as the operational phase.
![[-]](/dwicons/i_colpse.gif)