Abstract
Detection of mooring line failure of a floating vessel without any inputs of environmental conditions and/or directions can be made possible by measuring and analyzing vessel positions and headings, and also drafts in the case of an FPSO. This approach simplifies the required monitoring equipment and certainly helps in reducing cost and required maintenance.
In addition to changing a vessel’s mean positions and headings, mooring line failure changes natural periods of the system. Therefore, to determine the condition of mooring lines and to detect mooring line failure, a dry monitoring system can (or needs to) detect subtle changes of natural periods of the system and variations in vessel headings as a function of vessel position and draft (or mass) for an FPSO. This task that can be categorized as pattern recognition and classification is very difficult to carry out with human visualization or conventional methods/tools to get to a reasonable detection rate. Although challenging, solving this task can be rewarding. This is an area that Artificial Intelligence (AI) is very good at, and AI can be an essential component of a reliable dry monitoring system.
This paper presents several AI models, including Artificial Neural Networks (ANN), that can be used to determine the condition of mooring lines and to detect mooring line failure. This paper compares the performance of these AI models for a vessel’s loading condition (draft) that has not been included in the training of the models. This paper also discusses a method to enable an ANN model to be more tolerant and adaptive to new conditions and its applicability for other AI models.
AI models can be used in combination with conventional digital methods (numerical algorithms) to determine the condition of mooring lines and to detect mooring line failure. This approach of combining AI models and conventional digital methods for a dry monitoring system requires knowing how and when to use AI models and conventional digital methods. This important aspect is illustrated in the presented examples on how the monitoring system detects mooring line failure in real time. These examples show the importance of AI and conventional digital methods and the significant contribution from each of them.
This paper demonstrates that AI can be an important tool for a dry monitoring system of mooring lines and can be used in conjunction with conventional digital methods to increase the robustness of the solution. This application of AI is an example of the potential of AI in solving an engineering problem, which is otherwise difficult to solve using conventional methods/tools.
