Performance, emissions and combustion analysis of hydrogen-enriched compressed natural gas spark ignition engine by optimized Gaussian process regression and neural network at low speed on different loads

The transportation industry is increasingly focused on hydrogen based fuel as a promising alternative due to its potential for reduced emissions and enhanced performance. The purpose of this study is to improve efficiency and reduce emissions on low speed on different loads for heavy duty vehicles. This study can be impactful to train the electronic control unit (ECU) for heavy duty vehicles working on aforementioned conditions. This study investigates the effect of hydrogen ratios (0%–40 %) in HCNG, (0%–15 %) exhaust gas recirculation (EGR) ratios, and spark timing (8°-34o CA bTDC) at low and high loads (15 % & 75 %) under stoichiometric conditions at low speed (700 rpm). Performance, emissions and combustion parameters were thoroughly analysed across these conditions. Brake thermal efficiency increases by 20.7 % & 19.4 % by the addition of (0 %–40 %) hydrogen at low and high load at 14o CA bTDC running on 5 % and 0 % EGR respectively. NOx emissions reduces by 11.9 % & 17.9 % by the addition of (0 %–15 %) EGR and increases by 46.1 % & 46.4 % by increasing the amount of hydrogen in HCNG at low and high load at 14o CA bTDC at 0 % EGR respectively. Coefficient of variation reduces 13.8 % by (0 %–40 %) hydrogen addition at 11 % EGR at 16o CA bTDC. Optimized Gaussian process regression (GPR) and neural network (NN) machine learning techniques were applied to the dataset, and found GPR matern 5/2 is best one. The findings can be utilized in the development of HCNG engine.