Stochastic Dynamic Analysis of Offshore Bottom-Fixed Structures
Master thesis
Permanent lenke
http://hdl.handle.net/11250/2614989Utgivelsesdato
2015Metadata
Vis full innførselSamlinger
- Institutt for marin teknikk [3437]
Sammendrag
Paper 1
The first part of this master thesis presented in paper one, is evaluating several methods to estimate dynamic amplification of drag dominated structures. This is done primarily by 3 hour time-domain analysis in the computer program USFOS with linear wave theory, Wheeler stretching and the Morison load equation in irregular, extreme seas.
The dynamic amplification in irregular seas is estimated by performing dynamic and quasi-static analysis in the exact same conditions. Then, the most extreme quasi-static and dynamic extremes are found from each seed, and fitted to a Gumbel distribution. The dynamic and quasi-static extremes used to determine the equivalent dynamic amplification (EDAF), are found at the 0.9 fractile of the Gumbel fit. In other words, the 30 hour characteristic value is used. A convergence test is made with respect to the number of 3 hours analysis, or seeds, required to do a confident EDAF estimation, and it is recommended to use 30 or more seeds of full 3 hour simulations.
Further, an analytical expression of DAF, accounting for multi-harmonic excitation due to the drag term in the Morison equation, is fitted to the results. Several runs for each structure with different natural period is obtained by varying the topside mass. By estimating the total damping ratios for the different natural periods, the analytical expression is able to reflect the results well. A test case is performed on a jack-up, resulting in quite accurately estimated EDAF values in the natural period range of 6-9 seconds. Accounting for statistical errors in the Gumbel method, the analytical solution may give equally reliable results. Finally, the SDOF expression for DAF is not recommended for further use.
Paper 2
The second part of this master thesis is presented in the paper "Hydro-Elastic Contributions to Fatigue Damage on a Large Monopile". The objected has been to use USFOS as a time-domain simulation tool for an offshore wind turbine with a large monopile foundation. There is no doubt that USFOS can handle the elastic problem, but the hydrodynamics for a large-volume structure (D=9m) is limited to the built-in MacCamy and Fuchs correction for first order waves. However, wave kinematics can be given as a time variant three-dimensional grid, which is done in this study.
Typical fatigue limit sea-states for Dogger Bank are used for 30 minute simulations where the baseline moment and fatigue damage is calculated and compared between the applied wave theories. The hydrodynamic load cases include first order theory, second order theory, third order FNV formulation and second order forces calculated with the panel code Wadam from DNV. Results show that very high fatigue damage estimates are found when the undisturbed second order kinematics are used, even in sea-states outside the diffraction regime. For small sea-states, the first order diffraction theory corresponds quite well to the higher order loads.