Development and Validation of a Multi-zone Predictive Combustion Model for Large-Bore Dual-Fuel Engines

Numerical simulation represents a fundamental tool to support the development process of new propulsion systems. In the field of large-bore dual-fuel (DF) engines, the engine simulation by means of fast running numerical models is nowadays essential to reduce the huge effort for testing activities and speed up the development of more efficient and low-emissions propulsion systems. However, the simulation of the DF combustion by means of a zero-dimensional/one-dimensional (0D/1D) approach is particularly challenging due to the combustion process evolution from spray autoignition to turbulent flame propagation and the complex interaction between the two fuels. In this regard, in this activity a 0D/1D multi-zone DF combustion model was developed for the simulation of the combustion process in large-bore DF engines. The model combines a multi-packet approach for tracking the evolution and the autoignition of the pilot fuel with an entrainment and burn-up approach for the simulation of the premixed air-gas mixture flame propagation. To properly consider the properties of the fuels involved in the combustion process and to capture the interaction between the two fuels, the DF combustion model was optimized by developing and implementing a refined ignition delay model and specific laminar and turbulent flame speed correlations optimized for high-pressure and lean air-gas mixture. In addition to this, a multi-zone Nitrogen Oxides (NOx) model was developed and integrated into the combustion model. Experimental measurements from a single-cylinder Wartsila research engine were used for the model development and validation. The proposed DF combustion model is able to properly capture the effect of the main engine settings (i.e., load, pilot fuel injection strategy, compression ratio (CR), and boost pressure), providing accurate predictions of the ignition timing, combustion duration, and NOx emissions. The developed numerical model can be therefore exploited to virtually assess the potential of different engine technologies and calibration strategies.