Experimental optimisation of the man overboard manoeuvre

In maritime operations, the safety of crew and passengers is paramount, making the man overboard manoeuvre a critical emergency procedure for all vessels. Among the different ways to perform this manoeuvre, the Williamson turn is used to come back on its own wake to retrieve the man overboard. This manoeuvre is especially useful if it is not precisely known when the person fell overboard. During this manoeuvre the rudder is switched to its extremes at well defined heading deviations. The IAMSAR guideline provides standard values for these, but these values do not guarantee that the vessel will sail back on its own wake. In this study we first estimated these switching angles with a numerical model and a data-driven model. Next, we used update rules to fine-tune these switching angles and tested those in a 1:24 model-scale experiment in one of MARIN’s basins. After one iteration in the basin the lateral error and heading error were within DNV GL’s guidelines. The predicted switching angles from the numerical model were closer to the experimentally found values. However, the course and speed were better predicted with the data-driven model. The approach lends itself well to minimise commissioning time of a new vessel, and the switching angles found could be included on the wheelhouse-chart for a better execution of the Williamson turn. • An iterative procedure is developed for optimising a man overboard manoeuvre. • The method determines heading angles for rudder switching. • The procedure was tested in simulations and validated in model-scale experiments. • The vessel returned to the start position within 20m, and 2° course deviation. • A detailed numerical model and a data-driven model provided initial angles.