Discuss
鈥�
AI Model in Fuel Consumption prediction of ships before voyages
Energy Efficiency Management (EEM) of ships is a keyword to approach the Sustainable Maritime Transportation by 2050, International Maritime Organization (IMO) through two phases: Phase 1- Prediction of ship fuel consumption; Phase 2- Saving energy consumption during ship operation activities. From this target, the METAPORT project is led by Prof. Alexander Micallef, and his colleagues at Department of Electrical Engineering has designed the fuel consumption prediction model based on Artificial Intelligence (AI).
This model has been established through Artificial Neural Network (ANN) with the preliminary element of neurons to be connected each other based on layers. The model includes three layers: Input layer, hidden layer, and output layer. The training algorithm is used in this model through Levenberg-Marquardt and it has been evaluated through mean squared error to determine the performance of this model. The input layer consists of the operational parameters impacting on fuel consumption of ships, including ship specifications: Ship dimension (L x B x D), Dead Weight Tonnage (DWT), Type of ships (Container ship, Bulk carrier, General cargo ship, etc); features of marine propulsion plant system: Main Engine (M/E), Diesel Generators (D/Gs), Boilers; mass of carried cargos; ship hull condition and marine propeller; navigation weather condition, etc. Finally, the performance of this model provided to be highly accurate during testing.
This model is valuable for the stakeholders to manage and determine the fuel consumption of ship before voyages during time at ports. This aim contributes on enhancement of ship energy efficiency management in maritime transportation nowadays.
