Experimental study of emissions and conversion efficiency analysis of hydrogen-enriched compressed natural gas engine before and after catalytic converter and predicted by improved particle swarm optimization in conjunction with back propagation neural network (IMPSO-BPNN)

Internal combustion engines emit harmful gasses that affect the environment adversely. The study aims to reduce harmful emissions from CNG-fueled spark ignition engines by using a three-way catalytic converter. This present work investigates the effect of different parameters on CNG-fueled spark ignition engine before and after the catalytic converter. The experiment was conducted to investigate the effects of hydrogen ratios (0%–20%), EGR ratios (0%–20%), spark timing (8o–59o CA bTDC), load (25%–75%), and speed (900 rpm–1500 rpm) under stoichiometric conditions. Emissions (CO, HC & NOx) and conversion efficiency of laboratory-based CNG SI engine before and after catalytic converter is investigated. CO emissions are 8.78 g/kWh, 8.62 g/kWh and 6.97 g/kWh at 16o CA bTDC before the catalytic converter, at same ignition timing by using the catalytic converter, CO emissions are 3.41 g/kWh, 3.34 g/kWh and 3.22 g/kWh with HCNG0, HCNG10 and HCNG20 respectively. Conversion efficiency of the HC emissions is increased by increasing the (900 rpm–1500 rpm) speed. The NOx emissions increased by increasing the load are present in the amount of 0.0551 g/kWh, 0.0557 g/kWh and 0.160 g/kWh approximately for loads 25%, 50% and 75% respectively at 25o CA bTDC and reduced after the catalytic converter. Additionally, an improved particle swarm optimization combined with back propagation neural network (IMPSO-BPNN) predicts emissions, achieving a higher correlation coefficient (R = 0.9970) and minimum MSE of 0.0057 for CO, (R = 0.9993) and minimum MSE of 0.0037 for HC, and (R = 0.9995) and minimum MSE of 0.0020 for NOx. The findings may enhance training for electronic control units and the development of HCNG-fueled engines.