Abstract
Probabilistic models are important to be able to predict the often non-linear behavior of offshore structures. A measured time history of a specified motion captures the non-linearities and often it is more useful to use statistical methods to interpret the data. From the time history useful statistical parameters may be obtained, including the covariance nature of the data. Sometimes the data set is to short to adequately characterize the covariance nature of the process. This study makes use of a new technique to predict the expected mean behavior of a stationary random process, above a level crossing. The technique is based on the assumption that the covariance function behaves in the same manner as a damped harmonic oscillator. Based upon this observation the covariance function can be represented as an analytical function that can be adjusted to best fit the data. Data obtained from tests conducted at the Offshore Technology Research Center are used as the basis for a comparison between a statistical model and a model based upon the use of an analytical covariance function. The expected mean behavior of surge, sway, heave and yaw are analyzed and the numerical predictions compared with the measured data. The results for surge and sway motions were reasonable, but for heave and yaw the predictions were less promising.
Tysse, Leiv Andre (2000). Response of mini-TLP after a level crossing. Master's thesis, Texas A&M University. Available electronically from
https : / /hdl .handle .net /1969 .1 /ETD -TAMU -2000 -THESIS -T77.