Marine structural mooring systems can suffer a lot of damage during their service life, and if not detected in time, these damages can lead to serious consequences. Traditionally, damage identification in mooring systems has focused on measuring the tension of mooring lines. However, physical tension sensors are difficult and costly to maintain. In recent years, machine learning and deep learning methods have developed rapidly, and if these more accurate, lower cost, and smarter methods can be applied to the field of marine engineering, they will greatly improve the efficiency and accuracy of mooring system damage identification. In this paper, based on the Autogluon framework in automated machine learning (AutoML), we determine whether the mooring system of drilling semi-submersible is damaged by feature extraction and damage identification of the six-degree-of-freedom motion response under different sea conditions. The method can provide safety assessment information for marine structures and has some practical value.
With the gradual development of the field of offshore engineering to deep offshore areas, the research and application of floating structures are increasing. In the mooring system of offshore engineering, the conventional method to monitor the breakage or not of mooring cable is to arrange a certain number of sensors to evaluate whether the mooring cable is intact or not by measuring the physical tension. However, the accuracy and reliability of this method are often not guaranteed. Kai-Tung Ma(2013), of Chevron Energy Technology Co. Ltd, said that there were 21 mooring failures between 2001 and 2011, with an average of more than two per year. Nine of them were multiple-line failures. Maslin(2013) estimated that 150 mooring lines were repaired or replaced during that period, which had integrity issues. The sensors used to monitor physical tension on mooring lines are expensive and difficult to maintain. According to DNV(2019), field experience shows that they are prone to failure within a few years after installation.
