PREDICTION AND OPTIMIZATION OF BINARY AND TERNARY ISO-STOICHIOMETRIC GEM BLENDS ON PERFORMANCE AND EMISSSION CHARACTERISTICS OF SI ENGINE USING ANN AND RSM

The present work aimed at optimizing the performance and emission characteristics of a Port Fuel Injection (PFI) SI engine
fueled with Gasoline-Ethanol-Methanol (GEM) blends using Response Surface Methodology (RSM). Test fuels used in the
study are pure gasoline (E0), E10, E10 equivalent iso-stoichiometric GEM blend (E10_Eq), E20, E20 equivalent isostoichiometric GEM blend (E20_Eq). Formulated E10 and E20 equivalent blends have identical air-fuel ratios, lower heating
values, density, and octane number as target binary blends (E10, E20). The test engine was operated with different fuel blends
by varying the engine speed from 1700 to 3300 rpm at a constant engine load of 5 kg. For optimization of the engine, speed
and fuel blends were considered as input parameters and brake thermal efficiency (B_The), brake specific fuel consumption
(BSFC) and, nitrogen oxide (NOx) emissions as responses. Optimization was carried out using the desirability approach with a target of maximizing the B_The and minimizing the BSFC and NOx. From the results, it was observed that the E10_Eq GEM
blend operation of the test engine has optimized values of B_The, BSFC, and NOx emissions with values of 33.17%, 251
g/kW-hr, and 1389.8 ppm respectively at an engine speed of 2416 rpm. A composite desirability value of 0.64 obtained from
the regression model shows that RSM can be conveniently employed to determine the significant factors that could impact
engine performance and emissions.