Treffer: Efficiency Prediction for Emission Reduction in Highly Sour Diesel via Oxidative Desulfurization: A Python Neural Network Approach.

Machine learning algorithms have gained popularity recently as a method for predicting the efficiency of industrial processes. The fluctuation of desulfurization process operation conditions, especially sulfur content in diesel fuel, has an impact on the efficiency of the process. This study presents the development of an artificial neural network (ANN) model using Python to predict the efficiency of the oxidative desulfurization (ODS) process in highly sour diesel fuel. Experimental data from a trickle bed reactor were used to train and validate the model. The dataset included variables such as ODS temperature, pressure, liquid hourly space velocity (LHSV), and sulfur content in the feed. The ANN model demonstrated a high prediction accuracy with sulfur conversion results matching the experimental data with approximately 98% accuracy and a regression coefficient (R²) of 0.99. The model effectively captured the influence of the operating conditions, showing that higher temperatures and pressures significantly enhanced the desulfurization efficiency. Additionally, the optimization of LHSV contributed to achieving optimal sulfur removal. This work highlights the potential of machine learning techniques in enhancing the predictive capabilities and efficiency of industrial desulfurization processes. [ABSTRACT FROM AUTHOR]