Predicting propeller emergence severity in adverse sea conditions: An integrated approach using stacked machine learning models

Abstract

In adverse sea conditions, variations in the propeller submergence depth can induce ventilation and out-of-water effects. These effects result in propeller-torque and thrust losses, impacting the shipboard power network's reliability and accelerate wear on propulsion shaft system. Ensuring the stability and safety of sailing ships necessitates predicting the occurrence and severity of ventilation and out-of-water effects. This, in turn, enables the implementation of a switching strategy with a fast response for propulsion control. Nonetheless, existing methods can only identify current propeller ventilation states, mainly categorised as ventilation and non-ventilation, or more elaborately, as full ventilation, partial ventilation and non-ventilation. These methods, however, face a critical limitation—they cannot predict the severity of ventilation or the out-of-water effect, leading to ineffective propulsion-control-strategy switching. This study proposes a method for predicting propeller emergence severity (that is, non-emergence, light emergence and severe emergence) under wave conditions. This method introduces a novel approach for establishing propeller emergence states and a sophisticated stacked model. The new method categorises different propeller emergence states for diverse classifications based on the maximum emergence severity that the propeller can develop. This to overcome the challenges in predicting the development of the ventilation severity and the out-of-water effect. The innovative stacked model helps improve the overall prediction performance of a single intelligent algorithm. Finally, using simulation experimental data, it is verified that the propeller emergence severity method has satisfactory effectiveness and accuracy. The proposed method provides crucial technical support for the efficient switching of propulsion control strategies in electric propulsion systems when ships are navigating under adverse sea conditions.