Impact of combustion chamber wall temperature on knock in HCNG-fueled SI engines: A regression-based knock intensity correlation

Abstract

The efficiency of spark ignition internal combustion engines can be enhanced by utilizing hydrogen-based fuels. Hydrogen-enriched compressed natural gas is a promising alternative fuel, but its efficiency is limited by knock. This study investigates the knock limit of a spark ignition internal combustion engine fueled with hydrogen-enriched compressed natural gas by analyzing the effect of combustion chamber wall temperature and the role of exhaust gas recirculation. A numerical simulation is performed using a detailed chemical kinetics software package, CHEMKIN Pro-2021 R2, incorporating GRIMECH3.0 reaction mechanisms for hydrocarbon and hydrogen combustion to model knock behavior. The hydrogen content is varied from 0 % to 50 % in hydrogen-enriched compressed natural gas, while the exhaust gas recirculation rate is adjusted between 0 % and 10 %. Results show that increasing the hydrogen content to 50 % reduces the combustion chamber wall temperature at knock onset by 51.67 % without exhaust gas recirculation and 42.2 % with exhaust gas recirculation, significantly improving resistance to knock. Furthermore, a regression correlation is developed using multiple linear regression in MATLAB to predict knock intensity. The proposed model is validated through both literature comparison and an advanced statistical technique known as the least absolute shrinkage and selection operator regression, which selects the most relevant predictors for regression analysis. The developed model achieves a mean squared error of 2.59 % for the regression correlation, 7.68 % for combustion parameters, 12.24 % for working medium properties, 3.11 % for controllable factors, and 2.46 % for mixed parameters, demonstrating high accuracy. The findings of this study provide valuable insights for optimizing hydrogen-enriched compressed natural gas engines by improving knock resistance and predictive modeling.