Hybrid Modelling Approach of a Four-stroke Medium Speed Diesel Engine

Diesel engines will remain a fundamental component of propulsion systems due to their maturity, reliability, and power density. Building Digital Twins of the propulsion system is one feasible solution to pursue the optimal propulsions system operation, estimating system states and efficiency. This work will investigate a modelling approach that combines high accuracy while satisfying real-time prediction capabilities by coupling a physics-based model with a data-driven modelling approach. We will demonstrate that the proposed hybridisation framework can provide state-of-the-art prediction capabilities in real-time, utilising operational data from a turbocharged, four-stroke medium-speed diesel engine.