Performance and emissions optimization of an ethanol-gasoline fueled SI engine with oxygen enrichment using artificial bee colony algorithm

Abstract

Abstract. The use of arti cial neural network in conjunction with arti cial bee colony algorithm is proposed as a method for performance and emissions optimization of an SI engine. The case study here involves the oxygen enriched combustion of an SI engine fueled with hydrous ethanol and gasoline. In this study, the engine was considered as a black box and its performance and emissions were extracted experimentally at di erent intake air oxygen concentrations, hydrous ethanol injection rates, and ethanol concentration in the hydrous ethanol mixture. Then, the simultaneous injection of hydrous ethanol and oxygen enriched combustion was investigated to maximize the fuel conversion e ciency and minimize the CO and NOx emissions. Therefore, an objective function consisting of both the emission and performance parameters was optimized using the Arti cial Bee Colony algorithm. The engine model used in this optimization process was obtained from an Arti cial Neural Network trained with experimental engine data. For operating speed of 3000 rpm, the optimization results indicated 1.21% improvement in fuel conversion e ciency and 31.11% and 13.94% reduction in CO and NOx emissions, respectively. At the speed of 2000 rpm, fuel conversion e ciency improved by 4.11% and CO emission decreased by 18.73%, while NOx concentration increased by 28.35%.