Comprehensive optimization study for the methanolysis of Linum usitatissimum oil using response surface methodology and artificial neural network

Abstract

This study investigated the methanolysis of Linum usitatissimum oil using a homogeneous catalyst. Advanced optimization techniques, such as response surface methodology (RSM) and artificial neural networks (ANNs), have been employed to examine the relationship between the reaction parameters and biodiesel yield. The study investigated four reaction variables: the methanol-to-oil molar ratio, the catalyst concentration, the reaction temperature, and the methanolysis reaction completion time using RSM and its authentication was conducted using an ANN. The ANN model, consisting of 14 neurons and a well-trained Levernberg–Marquardt backpropagation method, showed a mean square error (MSE) of 0.027 at the best validation performance of 1.52 epoch-6. The coefficients of determination, R² for the RSM-CCD were 0.99 for the observed and 0.98 for the predicted values, proving the significance of the overall model (p-value < 0.001). Furthermore, the ANN had an R² value of 0.97, confirming the reliability of the test and complementing the RSM results. Analysis of variance (ANOVA) and regression models identified the interactions among these variables, and temperature × time (CD), catalyst concentration × temperature (BC), and catalyst concentration × time (BD) were identified as significant factors for enhancing the biodiesel yield. The RSM-CCD gave a highest possible yield of 98.7% biodiesel that was attained in only 50.4 min using a 12 : 1 methanol-to-oil ratio, 1.25% catalyst concentration, and a reaction temperature of 65 °C, whereas 97.52% yield was predicted using the ANN. A biodiesel confirmation test was performed using infrared spectroscopy and gas chromatography (GC), while adhering to ASTM D6751 biodiesel specifications to evaluate its fuel properties such as the flash point (175 °C), kinematic viscosity (5.72 °C), cloud point (−4 °C), pour point (−9 °C), acid value (0.39 mg KOH per g), higher heating value (43 MJ kg⁻¹), water content (0.019%) and density (897 kg m⁻³), thus emphasizing that the Linum usitatissimum oil has significant potential for use in biodiesel production.