Search for a command to run...
• Eddy-Burn Up combustion model predictive capabilities assessed with 2 efuels. • A surrogate fuel algorithm was developed to replicate key fuel properties. • The combustion model was not recalibrated fuel-wise. • The combustion model is integrated with laminar flame speed neural networks. Synthetic fuels produced with renewable surplus electricity depict an interesting solution for decarbonising mobility and transportation applications, reducing greenhouse gas emissions and mitigating global warming. This work proposes a methodology to enable the possibility of replicating the combustion behaviour of 2 synthetic fuels: an e-gasoline fuel and MtG – E10. At first, a surrogate fuel algorithm was conceived to replicate real fuel target properties (density, C/H, RON, MON, oxygen volumetric percentage). The surrogate hydrocarbon palette was selected to represent each hydrocarbon type, and the CRECK chemical reaction mechanism was selected to define the laminar flame speeds of the surrogate composition. The methodology was developed in a 1D-CFD simulation environment, representing a single-cylinder research engine. The Eddy-Burn Up combustion model, previously calibrated for standard fossil gasoline, showed great accuracy in replicating key combustion metrics, highlighting its predictive capability without the need to recalibrate fuel-wise the turbulent flame seed parameters. The Root Mean Square Error for the maximum pressure is 1.7 bar, and the MFB50 maximum deviation is below 1°. Eventually, knock occurrence was evaluated by employing the Livengood-Wu induction time integral. The induction time integral reaches or overcomes the knocking threshold of 0.9 for most of the cases whose normalized Maximum Amplitude Pressure Oscillation 98.5 percentile value is above 1, thus showing the ability of the knock model to predict which engine operating points are knocking or not, leaving the possibility to develop control strategies based on such approach.