Parametric optimization of microwave-assisted green fuel production from Caryota urens seed oil using response surface methodology and artificial neural networks

• Bio-oil productionfrom Caryota urens waste seed oil was optimized for industrial biodiesel applications. • An efficient Cu-BDC MOF heterogeneous catalyst was used for biodiesel production. • Integrated RSM and ANN studies revealed the optimized reaction conditions. • The study highlights Caryota urens as a sustainable biomass resource for the scalable production of biodiesel. Bio-oil, produced by non-edible seeds, is an emerging sustainable feedstock for biodiesel production in industrial applications. In this study, bio-oil obtained from Caryota urens seeds was used for biodiesel production using a heterogeneous catalyst (CuBDC MOFs). The reaction parameters were optimized using integrated RSM and ANN modeling approaches. RSM (Response Surface Methodology) using a Central Composite Design (CCD) identified reaction parameters: catalyst concentration, Reaction temperature, Reaction time, and Molar ratio of oil to alcohol, enhanced biodiesel yield and achieved high predictive accuracy with an R² of 0.9340. Similarly, the Artificial Neural Network (ANN), with a configuration of hidden neurons, outperforms RSM, yielding a correlation coefficient (R²) of 0.91663 and 0.94325 for training and validation, along with a lower Mean Absolute Error (MAE) and Root Mean Squared Error (RMSE). This integrated approach demonstrates the advantages of combining statistical (RSM) and Computational (ANN) tools for optimizing biodiesel production. The produced biodiesel was tested for its fuel properties and compared with ASTM standards to ensure it to met the required diesel properties. Based on the optimization study results, biodiesel produced from Caryota urens seed s oil can be tested in a diesel engine. The study highlights Caryota urens as a promising biomass resource for industrial-scale of sustainable biodiesel production..