N., SanthoshAfzal, AsifV., Srikanth H.Ağbulut, ÜmitAlahmadi, Ahmad AzizGowda, Ashwin C.Alwetaishi, Mamdooh2023-07-262023-07-2620230360-5442https://doi.org/10.1016/j.energy.2023.126826https://hdl.handle.net/20.500.12684/12360The purpose of the present study is to evolve an alternate non-edible source for the synthesis of biodiesel and use it as a fuel substitute in diesel-ethanol blends for DI-CI engines. The use of discarded poultry fat feedstocks for the sustainable production of biofuels in the current day scenario is a novel approach that is still in its embryonic stage. For the effective utilization of these processed biofuels, it is very much required to ascertain the characteristics and their performance attributes for different blends. In this regard, a set of experiments are planned to study the emission and performance attributes of a direct injection (DI) diesel engine operating on poultry fat biodiesel, and the three proportions of diesel-biodiesel-ethanol blends with varying vol. % over the wide load range on a diesel engine. The ethanol percentage in the blend is varied from 5 vol % to 15 vol % in increments of 5 vol % with the amount of poultry fat-based biodiesel kept constant at 10 vol %. The performance and emission characteristics, particularly, the CO, CO2, NOx, unused Oxygen, and hydrocarbon emissions are experimentally determined for different fuel blends. From the results, it is evident that the performance characteristics of the fuel blends improve with the addition of ethanol in the diesel-biodiesel blend. Further, regression modeling of the performance characteristics is carried out to optimize the blend and operating load conditions, and the regression model is evolved for developing a mathematical relation for predictions of the results for different operating conditions. Also, Artificial Neural Network (ANN) modeling of the performance characteristics is carried out at each stage to predict the outcomes for different blends and load conditions and provide a set of empirical relations for analyzing the performance characteristics of the engines operating on poultry fat-based biodiesel-diesel-ethanol blends. Excellent predictions are obtained using regression modeling and ANN with R-squared values above 0.9. Thus, the present work provides a newer model of effectively using the ANN for the systematic study of the performance characteristics of the biodiesel blends obtained from a set of experiments through various optimization methods for better performance and a significant reduction in emissions. © 2023 Elsevier Ltden10.1016/j.energy.2023.126826info:eu-repo/semantics/closedAccessDirect injectionEmissionEmulsifiersEthanolPerformancePoultry fatDiesel enginesDirect injectionEthanolForecastingNeural networksRegression analysisC.I. engineCI engineEmissionEmulsifierEthanol blendFuel substitutesPerformancePerformance characteristicsPoultry fatsRegression modellingBiodieseldiesel engineequipmentethanolexperimental studymodelingoptimizationperformance assessmentregression analysisArticle2702-s2.0-85148348141WOS:001013264100001N/AQ1