Predicting Propulsion Motor Power Demand for Weather Routing Using GAMs: A Case Study on the Oshima Maru Training Ship

The purpose of this thesis is to develop a Generalised Additive Model (GAM) in order to predict the propulsion motor energy consumption of the Oshima Maru training ship using exclusively data relating to weather and environmental conditions and location information, while also evaluating their effects on the energy consumption. This approach addresses the International Maritime Organisation’s Carbon Intensity Indicator (CII) framework [1] proposed tools to lower emissions and fuel demand caused by the maritime industry through more operational energy efficiency. The framework proposes as a tool weather routing in order to increase the efficiency. This thesis outcome will help this issue by providing a predictive model of energy consumption that can then be inserted in an optimisation framework to define the optimal route that minimises energy consumption. The thesis therefore targets a practical limitation of the energy efficiency interventions, such as the uncertainty regarding the energy consumption magnitude for a vessel given real operational conditions. Even when certain optimisations can be theoretically beneficial, the stakeholders lack sources of reliable, specific, customized estimates of how the conditions of the environment and weather translate into the propulsion motor energy consumption. Hence this thesis enables a step forward in operational efficiency by exploiting weather and location data into vessel-specific forecasts of the propulsion motor energy demand in an actionable way. It allows to give more strength to an evidence based weather routing that is relevant to the CII operational strategy and helps lower the uncertainty that is currently limiting a wider adoption of energy efficient practices.