An Excellent Relationship-Based Model Development in Deriving the Short-Term Probability Distribution of Offshore Structural Responses

Accurate calculation of the structural responses has been extensively used based on a time-domain approach. Due to its ability to compute the nonlinear association between wave forces and the structural responses, the most method regularly used for predicting the probability distribution of the extreme responses in such statistical random condition is the Monte Carlo time simulation (MCTS) method. However, this method suffers from excessive sampling variability, and hence, a large number of simulated response records are required to reduce the sampling variability to acceptable levels. As well, a simple technique for derivation of the probability distribution of extreme responses is still a limited source. Findings show that a more efficient technique by an efficient time simulation (ETS) method possesses an advantage of the correlation between the extreme values of surface elevation (input) and their corresponding responses (output). In this article, the extended adaptation of ETS procedures provides an excellent correlation between input–output variables for developing the ETS-regression (ETS-reg) models. This proposed ETS-reg model is then verified by the extensive MCTS method and compared to another association of relationship-based models, namely the modified finite-memory nonlinear system (MFMNS) and its improved version of MFMNS (eMFMNS) models. In conclusion, the different models-based relationships are examined to the high design waves. Thus, the most accurate regression-based models in forecasting the return period of 100-year extreme responses are discussed.