SHIP PREDICTIVE CONTROL SYSTEM DESIGN BASED ON THE VESSEL DIGITAL TWIN

Model Predictive Control (MPC) is defined as a model-based approach, thus necessitating the development of an internal predictive model, that is both reliable and realistic.It is not feasible to carry out this identification process on a real ship, due to the trials number and duration of each experiment.Moreover they are affected by the environmental disturbances that cannot be eliminated.Consequently, there is an imperative to identify a methodology for generating input-output data sets for the purpose of predictive model identification and validation.This paper details the development of a digital twin of a training ship, utilizing the mathematical model of the MMG (Mathematical Model Group).Following validation, the digital twin is employed as a source of input-output data for the model identification process.The paper goes on to describe the design of a ship predictive control system, starting from the initial idea and considering requirements, modelling and prediction, control signals computation, simulations and hardware-in-the-loop tests.The entirety of the research is grounded in the utilization of the Matlab software.The digital twin operates in real time as a component of the hardware-in-the-loop system, thereby addressing a current research gap and enhancing potential for the control quality in autonomous shipping predictive systems.