• digital twin
• operating conditions
• neural network
• AI
• yacht

The aim of this thesis is firstly to create a digital twin model of main parts of a 42 meters mega-yacht, called SeaCat42, using softwares such as MATLAB and Simulink, to summarise the behaviour of the ship in terms of electric power consumption, to monitor the operating conditions and to build an intelligence, called Blue AI, that gives advice to ship's users, guests or crew and then to create an Artificial Intelligence which predicts these variables in advance. Starting from the boat's layout, most of cabins' electric and electronic elements were modelled and also the navigation part, the power generation and propulsion part. The ultimate purpose of this work is to give to users some advice according to actual power consumption compared to ship's actual power and monitoring advice related to the operating conditions of the yacht. At the end of the project, Machine Learning techniques were used to train two neural networks, in order to give consumption's prediction and operating conditions in an amount of hours (e.g. 6 hours from now) according to many known inputs, such as the weather forecasts, the desired parameterers by users, the yacht's speed, etc. The studied NNs have been Feedforward NNs (FNNs); many datasets were built, then the NNs were modeled, trained and tested and the predictions were compared with the real model's outputs.