Prediction of safe sea-state using finite element method and artificial neural networks

Abstract: This article proposes a predictive method for identifying the range of sea-states considered safe for the installation of offshore structures. A finite element dynamic analysis of the system for various sea-states characterized by significant wave heights and mean zero-up-crossing wave periods and modeled as a combination of several wave components has been performed. Using this procedure a table of safe and unsafe sea-states is generated. The significant wave height (H s ) and mean zero-up-crossing wave period (T z ) of a future sea-state in a location in the north east Pacific were predicted from the distributions whose parameters were estimated using the artificial neural networks (ANNs) trained for this purpose. The location of US National Oceanographic Data Center (NODC) Buoy 46005 is used in this study. The H s and T z of some future sea-states were predicted from their corresponding conditional 7-parameter distribution given some information including a number of previously measured H s ’s and T z ’s. This gives a predicted sea-state for a specific time in future. The parameters of the distributions have been estimated from the outputs of two different 7-network sets of trained ANNs. A pile-driving operation is used as a case study in which the pile configuration, including the non-linear foundation and the gap between the pile and the pile sleeve shims, has been modeled by the finite elements method and the range of sea-states suitable for safe pile-driving operation was identified. [Copyright &y& Elsevier]