Integrating RSM and NSGA-III for Multiobjective Optimization of a Diesel/Methanol Dual-Fuel Engine Performance and Emissions.

Diesel/methanol dual-fuel engines are crucial for the decarbonization and sustainable development of transportation and industrial sectors in the carbon-neutral era. However, challenges such as low combustion efficiency, operational instability, and high emissions persist, particularly under low-load conditions. This study provides a comprehensive investigation into the combined effects of injection strategy, methanol substitution ratio (MSR), and exhaust gas recirculation (EGR) on the emission characteristics and fuel economy of a methanol-diesel dual-fuel engine under a 25% load. Multiobjective optimization was performed using response surface methodology (RSM) and the Nondominated Sorting Genetic Algorithm-III (NSGA-III). The results indicate that the interaction between the MSR and EGR has a considerable impact on both engine performance and exhaust emissions. Increasing MSR from 5 to 20% reduced nitrogen oxides (NO x ) by 11.2-11.7% and particle number (PN) by 12.8-16.1% but increased hydrocarbon (HC) by up to 231.8% and carbon monoxide (CO) by up to 231.7%, while a higher EGR rate (0-30%) effectively reduced NO x by 29.8-33.4% at the expense of 151.5-171.5% higher PN emissions, and slightly decreased HC by 2.6-2.8%, lowered equivalent brake-specific fuel consumption (ESFC) by 1.0-1.4%, and improved brake thermal efficiency (BTE) by 1.9-2.7%. Furthermore, advancing main injection timing (MIT) by 8° CA ATDC (from 5° CA ATDC to -3° CA ATDC) and increasing fuel injection pressure (FIP) from 80 to 100 MPa contributed to 27.2 to 27.9% lower NO x emissions and 1.6 to 3.5% higher BTE, although their impacts on PN, CO, and HC emissions were more complex. The RSM model constructed in this study exhibited high predictive accuracy, with the coefficient of determination (R 2) exceeding 96% and the error between predicted and experimental values of BTE, NO x , and PN being only 1.102, 1.046, and 5.465%, respectively. The implementation of the optimized parameters (average: MSR = 5.23%, EGR = 4.06%, MIT = -1.98° CA ATDC, FIP = 90.49 MPa) resulted in simultaneously minimized emissions and a BTE of over 33% (1.47% higher than that in the pure diesel mode). NO x and PN emissions were maintained below 250 ppm and 2.5 × 107 #/cm3, respectively, with an average reduction of 5.11 ppm in NO x and 0.14 × 107 #/cm3 in PN compared to the original engine operating in pure diesel mode. This study underscores the effectiveness of AI-driven optimization in advancing the performance and emission characteristics of diesel/methanol non-road dual-fuel engines.