The sensors abnormalities, faults and failure detection and diagnosis for marine engines are considered crucial for ensuring the engine safe and smooth operation. The development of such system(s) is typically based on the manufacturers experience on sensors and actuators faults and failure events. This study aims to introduce a novel methodology for the sensors diagnostics and health management in marine dual fuel engines by employing a combination of thermodynamic, functional control and data-driven models. The concept of an Engine Diagnostics System (EDS) is developed to provide intelligent engine monitoring, advanced sensors' faults detection as well as timely and profound corrective actions. This system employs a neural networks (NN) Data-Driven (DD) model along with appropriate logic controls. The DD model is set up based on the derived steady state data from a thermodynamic model of high fidelity and is capable of real-time prediction of the engine health condition behaviour. The concept of a novel Unified Digital System (UDS) is proposed that combines the engine's existing control and diagnostic systems and the proposed EDS systems. The functionality of the UDS system is validated by employing a digital twin of the considered marine dual fuel engine by investigating scenarios for assessing the engine performance that entail abnormalities in the engine's speed and boost pressure sensors. The simulation results demonstrate that the developed UDS is capable of sufficiently capturing the engine's sensors abnormalities and applying appropriate corrective actions to restore the engine operation in its original state. This study benefits the development future systems facilitating the engines condition assessment and self-correction of the engine sensors' abnormalities, which will be required for smart and autonomous shipping.
- marine dual fuel (DF) engines
- 0D/1D modelling
- data-driven model
- digital twin
- engine control functional systems
- sensor abnormalities detection and correction