The advancements in digital twins when combined with the use of the machine learning tools can facilitate the effective health assessment and diagnostics of safety critical systems. This study aims at developing a framework to address the health assessment of marine engines utilising digital twins based on first-principles. This framework follows four distinct stages, with the former two including the marine engine digital-twin set up by customising the required thermodynamic models, as well as its calibration using tests trials data representing the engine healthy conditions. In the third stage, measurements from actual operating conditions are corrected and subsequently employed to develop the digital twin representing the prevailing conditions. The fourth stage deals with the engine health assessment by assessing health metrics derived from the developed digital twins. This framework is demonstrated in a case study of a large marine four-stroke nine-cylinder propulsion engine. The results demonstrate that three cylinders are identified to be underperforming leading to an average increase of the engine Brake Specific Fuel Consumption (BSFC) by 2.1%, whereas an average decreases of 6.8% in Indicated Mean Effective Pressure (IMEP) and 6.1% in the Exhaust Gas Temperature (EGT) are exhibited for the underperforming cylinders across the entire operating envelope. The developed digital twins facilitate the effective mapping of the engine performance for the entire operating envelope under several health conditions, providing enhanced insights for the current engine health status. The advantages of the proposed framework include the use of easily obtained data, and its application to several engine types including two and four-stroke engines for both propulsion and auxiliary use.

First principle Digital Twins (DT) for marine engines are widely used to estimate in-cylinder pressure, which is a key parameter informing health of ship power plants. However, development and application of DT faces barriers, as they require exhaustive calibration and high computational power, which render their implementation for shipboard systems challenging. This study aims at developing a data-driven DT of low computational cost for predicting instantaneous pressure. Two different approaches using Artificial Neural Networks (ANN) with distinct input parameters are assessed. The first predicts in-cylinder pressure as a function of the phase angle, whereas the second predicts the discrete Fourier coefficients (FC) corresponding to the in-cylinder pressure variations. The case study of a conventional medium speed four-stroke diesel marine engine is employed, for which the first principle DT based on a thermodynamic, zero dimensional approach was setup and calibrated against shop trials measurements. The DT is subsequently employed to generate data for training and validating developed ANNs. The derived results demonstrate that the second approach exhibits mean square errors within ±2% and requires the lowest computations cost, rendering it appropriate for marine engines DTs. Sensitivity analysis results verify the amount of training data and number of Fourier coefficients required to achieve adequate accuracy.