Hybrid ANN-MOGA-II-AHP Based Performance-Emission Optimization of Diesel-Ethanol Fueled CRDI Engine

The present investigation explores the pivotal significance of artificial intelligence (AI) coupled multi-objective genetic algorithm (MOGA)-II in scrutinizing the favorable operating conditions of a common rail direct injection (CRDI) engine fueled with Diesel-Ethanol. In the beginning, with the intention of trading off the performance-exhaust emission parameters of the CRDI engine, four artificial neural network (ANN) models are developed for the four engine output parameters, namely, brake thermal efficiency (Bth), carbon monoxide (CO), cumulated oxides of nitrogen and hydrocarbon (NOHC) and particulate matter (PM) wherein the engine load, Ethanol share and split injection strategy are preferred as input parameters. With the consolidated impacts of the four ANN models, Uniform Latin Hypercube sampling design of experiment (DOE) and MOGA-II scheduler, a total of 2120 Pareto optimal solutions (POSs) are obtained for the Diesel-Ethanol blends. To this end, a multi-attribute decision-making (MADM) based analytic hierarchy process (AHP) is used to select an operating condition from the obtained POSs. The AHP points out that the 72.19% load, 2.85% (by vol.) Ethanol share and 14.53% split injection strategy are the favorable CRDI engine operating conditions for obtaining minimal emissions along with higher performance. The experimental validation and noteworthy performances in desirability and composite desirability have elevated the hybrid method of ANN-MOGA-II-AHP in screening and selecting the performance emissions of Diesel-Ethanol fueled CRDI engine under environmental protection agency (EPA) Tier 4 mandates.