Hybrid experimental: ANN approach for rapid dibenzothiophene oxidation in oscillatory reactor using coated magnetic carbonaceous catalyst
摘要
The oxidative desulfurization (ODS) of aromatic sulfur compounds has emerged as a promising alternative to conventional hydrodesulfurization for producing ultra-low-sulfur diesel fuel under mild operating conditions. In this study, a coated magnetic carbonaceous (CMC) catalyst, composed of Al₂O₃-(MnO₂–Fe₂O₃)/activated carbon, was synthesized via wet impregnation and characterized using SEM, BET, XRF, and TEM analyses. The catalyst was employed in a newly developed oscillatory baffled reactor equipped with central basket baffles to achieve continuous oxidation of dibenzothiophene (DBT) using hydrogen peroxide as the oxidizing agent. The effects of operating temperature, desulfurization period, oscillation amplitude, and oscillation frequency on DBT removal were systematically investigated. The maximum desulfurization efficiency reached 95.98% under the optimum conditions of 383 K, 12 min reaction time, 12 mm oscillation amplitude, and 2 Hz oscillation frequency. In addition, an Artificial Neural Network (ANN) model based on a multilayer perceptron architecture was developed to predict the oxidation performance. The ANN model demonstrated excellent predictive capability with R² = 0.96, MAE = 0.2200, and MSE = 0.0829, showing close agreement between predicted and experimental data. The combination of the novel coated catalyst, intensified oscillatory reactor design, and ANN modeling provides an efficient and sustainable strategy for rapid oxidative desulfurization of diesel fuel and offers valuable insights for future industrial applications.