A Fuel-Flexible Quasi-Dimensional SI Engine Combustion Model for Burn Rate and Knock Prediction in Gasoline-Syngas Blends

This study presents a fuel-flexible quasi-dimensional Spark Ignition (SI) engine combustion model for predicting the burn rate and knock in gasoline-syngas blends. The framework utilizes a multizone structure and a dedicated knock combustion model, while introducing two key advancements for fuel flexibility of the entire engine model: (i) a novel machine-learning-based laminar flame speed model that captures the nonlinear dependence of the turbulent burn rate on fuel composition and thermodynamic properties, and (ii) an extended ignition delay model that incorporates syngas (H2/CO) components for knock prediction. In this study, gasoline is represented by an optimized toluene reference fuel (TRF) surrogate along with H2 and CO to enable simulation of gasoline-syngas blends. The model is calibrated and validated against single-cylinder SI engine experiments at 1500 rpm and 0.8–1.0 bar of intake pressures. The new SI engine model accurately reproduces in-cylinder pressure traces, apparent heat release rates, combustion phasing (CA05–CA90 deviations <2 °CA), indicated mean effective pressure (<2.5%), inducted mass flow (<2.5%), and knock-limited spark timing (<1 °CA), mostly within the experimental cycle-to-cycle variability. Crucially, the model successfully predicts the change in the detonation borderline resulting from syngas addition, demonstrating its ability to resolve the competing effects of a faster burn rate and a longer ignition delay time with satisfactory accuracy. Overall, this modular and composition-sensitive framework provides a scalable platform for predictive combustion and knock modeling of multifuel options in SI engines.