Ship digital twin architecture for optimizing sailing automation.

A fully autonomous ship should be monitored in a digital environment to track its real-time response to the changes in its surroundings. It is commanded by a controller algorithm whose optimum parameters is subject to change with respect to the changes in the environment. The ship's response with respect to these variations in the controller parameters need to be tracked to develop a full understanding of the ship's autonomy. As such, it is believed that the digital twin of a ship is an essential component of the path to fully autonomous ships. Despite maritime institutions not giving the credit to it yet, the ship digital twin concept is expected to be one of the leading topics in marine engineering soon. In this study, we try to establish the ship digital twin concept in terms of navigation autonomy. The paper starts by introducing the path to full autonomy in seas and the need for the digital twin by defining this concept for ships. A maneuvering mathematical model is used to represent the physical ship. The details of the model and the controller algorithms are given next. The propeller and rudder models are first validated by free-running self-propulsion and turning circle tests. The ship considered in this study has a twisted rudder and does not possess course-keeping ability; therefore, it requires autopilot to move straight ahead. Three types of simulation cases are identified with one being the autopilot mode. In the other two cases, the ship is forced to conduct two hard maneuvers while accelerating/decelerating. Investigation of the generated results reveals the effectiveness of the digital twin architecture used in this study. It has also allowed us to conduct a controller stability analysis, which has shown the behavior of the controller gains to detect the most optimum values for the considered model ship. • Ship digital twin (SDT) is a misunderstood concept in marine engineering: it is not a digital model. • SDT is an essential component of the path to fully autonomous ships. • It can be used to enhance navigation autonomy and this study presents a digital twin architecture. • SDT offers an environment to optimize controller gains before implementation. • The environment is open for improvement when coupled with machine learning algorithms. [ABSTRACT FROM AUTHOR]