A predictive digital twin of an autonomous ship is a continuously updated virtual representation of the physical ship that can be used to predict and prevent potential failures. By utilizing sensor data, diagnosis algorithms, and time series prediction methods, the digital twin can predict component failure inside the ship, allowing for early detection and mitigation. In this paper, the predictive digital twin is developed in three steps: (i) by describing the dynamics behaviour of the ships based on mechanistic models, (ii) by continuously updating the mechanistic models using data from sensor systems, and (iii) by predicting the faults and visualizing the results. To this end, the mechanistic model is derived from the first principles, while the updating algorithm is developed based on an adaptive exogenous Kalman filter. Experimental validation using a small ship platform is conducted, for which the results show the method can potentially increase safety and reliability of the ship during operation.
Predictive Digital Twins for Autonomous Ships*
2023-08-16
4449949 byte
Conference paper
Electronic Resource
English
Digital Twins for Automotive Predictive Maintenance
IEEE | 2024
|DataCite | 2020
|DataCite | 2022
|