Catalytic Co-pyrolysis of soybean husk and high-density polyethylene: Artificial neural network modeling and synergistic approach for enhanced gasoline-range biofuel production using zeolite HZSM-5

This study investigates catalytic co-pyrolysis of soybean husk (SBH) and high-density polyethylene (HDPE) using HZSM-5 to enhance gasoline-range bio-oil yield. Thermogravimetric analysis (TGA) and artificial neural network (ANN) models predicted thermal degradation at 10, 20, and 30 °C/min. Analytical pyrolysis was performed at 450 °C, 550 °C, and 650 °C, under non-catalytic conditions, with varying SBH-to-HDPE ratios (1:3, 1:1, and 3:1). Catalytic pyrolysis used 550 °C, with residual waste-to-catalyst ratios (RW:C) from 1:1 to 1:5. HZSM-5 reduced activation energy by 30 %, enhanced conversion, and increased hydrocarbon production by 25 % over non-catalytic runs for both models. Optimal conditions (550 °C, 1:1 SBH:HDPE) yielded over 90 % gasoline-range hydrocarbons. Unlike previous studies, this work employs SBH, an abundant but underutilized agro-industrial waste. The ANN model incorporates catalyst loading and feedstock ratio, an advancement over existing pyrolysis models, allowing accurate prediction and reduced experimental demands. ANN models (R2 = 1, MSE = 0.0023) predicted co-pyrolysis behavior at untested heating rates (15 and 25 °C/min). This approach offers technical advancements and also supports sustainable waste management by converting biomass and plastic residues into high-value fuels. The combination of experimental and machine learning provides a scalable strategy for optimizing biofuel production from mixed waste streams.