ABSTRACT

Jack-up offshore platforms have the advantages of convenient floating transportation and good stability during operation, so they are widely used in the fields of offshore oil resource extraction, wind power installation and offshore building installation. However, the offshore platform works in the harsh marine environment, and suffers from environmental loads such as wind load, wave load and current load for a long time, and once structural damage and failure occurs, it will lead to serious consequences such as casualties, property loss and environmental damage, so it is very important to ensure its structural safety. Wave loads are a very important part of the loads borne by ocean platforms, and the relationship between wave elements (wave height, wave period, flow velocity, etc.) and the size of wave loads can be established based on methods such as Stokes wave theory and Morison's formula. In this study, a numerical simulation method is used to construct a sample database by taking parameters such as wave height, period, and flow velocity as inputs and calculating wave loads using the finite element software ANSYS APDL. A Bayesian regularized neural network is used to construct a wave load prediction surrogate model. Changing the parameters such as wave height and period, the load values calculated by the finite element software were compared with those predicted by the neural network surrogate model, and the results showed that the surrogate model constructed in this study has high accuracy and can effectively predict the wave load values. This study provides a new idea for wave load forecasting, which has a positive effect on the intelligence and real-time wave load forecasting.

INTRODUCTION

With the increasing depletion of land resources and the need for environmental protection, the development of marine resources is becoming more and more important. Jack-up marine platforms are widely used in the development of marine oil and gas resources, which are the bases for offshore production operations and life. As the structure of the marine platform is in long-term service in the harsh maritime environment, it is subject to the interaction of wind, waves, currents and other loads, under the long-term action of these harsh environmental loads, it will cause various forms of damage to the structure, and even lead to the failure of the platform, resulting in property losses and casualties. Due to the complex structure of the marine platform and the huge amount of various data generated, the use of traditional statistical analysis methods to deal with these data is less efficient and less effective, and machine learning methods have advantages in this regard.

This content is only available via PDF.
You can access this article if you purchase or spend a download.