You are hereAirgap analysis of floating structures subject to random seas: Prediction of extremes using diffraction analysis versus model test results

Airgap analysis of floating structures subject to random seas: Prediction of extremes using diffraction analysis versus model test results

John Albert Sweetman
Principal Advisor: 
Steve Winterstein
Year Published: 
Mon, 01/01/2001 (All day)
BertSweetman.pdf2.72 MB

Statistical and dynamic non-linearities in ocean waves and wave-structure interaction are considered. Focus is on prediction of extreme airgap events for semisubmersibles. It is recognized that ocean waves are inherently nonlinear, and that this non-linearity affects the statistics of extreme crests. Two sources are assumed to account for all nonlinearity: incident waves, and wave-structure interaction. Various methods of predicting extremes based on post-processing hydrodynamic analysis and model test results are proposed. Methods are tested against model test results for the Veslefrikk semisubmersible.

First, methods using regression and fractile trend-lines to predict extreme airgap events from model test results are developed and confirmed.

Second, methods to predict extreme airgap events based on linear diffraction results are developed and confirmed. All of these new models include Stokes second-order incident waves.

Third, full second-order (WAMIT) hydrodynamic panel diffraction is applied. Two new methods of modifying quadratic transfer functions (QTF's) are developed: In one, QTF's predicted by WAMIT are replaced with those predicted by Stokes theory for short periods only. In the other, known on-diagonal QTF's predicted by multi-column analysis (WACYL) are extrapolated to estimate off-diagonal terms. WACYL QTF magnitudes are found to be reasonable for all period ranges, so no Stokes substitution is necessary.

Fourth, black-box system identification is used to extract first- and second-order transfer functions from measured data. The reasonableness of the Stokes substitution is confirmed, as is the capability of second-order modeling of airgap demand.

Finally, a system identification based on Stokes second-order theory is applied to incident waves. First- and second-order components of a specified wave history are identified, and results are used predict consistent first- and second-order wave time-histories at other spatial locations.

The most significant conclusions are: second-order effects are important to prediction of airgap demand; WAMIT over predicts these effects for high frequencies; incident waves are a meaningful source of second-order effects, and application of first-order diffraction with second-order incident waves is reasonable when estimating airgap extreme statistics.