Artificial neural network based optimization of engine performance using domestic biomass as a clean fuel

The growing demand for sustainable energy has encouraged the development of innovative approaches combining renewable biomass with advanced technologies. This work investigates poultry waste sourced from chicken shops as a potential raw material for oil generation. The optimal conditions, which resulted in the highest biodiesel yield and lowest Environmental Factor (E-factor), were achieved consuming a 9:1 methanol-to-chicken-oil ratio, 1% catalyst by weight, a 1-h reaction time, and a mixing speed of 500 rpm. Additionally, various nanofluid formulations containing Zinc sulfide (ZnS) nanopowder were prepared at concentrations ranging from 50 to 150 ppm, in increments of 50 ppm, and blended with a C20D80 fuel mixture. Among the tested blends, the C20D80 + 100ppm formulation demonstrated superior performance, reducing HC emissions by 29.23%, CO by 37%, and NOx by 5.7% while maintaining comparable engine performance to standard diesel. The experimental findings were substantiated through AI-based neural network, achieving a high R²-values ranging from 0.9197 to 0.9961, and low RMSE and MAPE. These findings highlight the significant potential of integrating waste biomass, E-factor assessment, and nanotechnology for the development of cleaner, sustainable fuel alternatives for future applications.