Modeling and Optimization of Oil-in-Water Emulsions Treatment Using Ultrafiltration and Neural Networks

In this study, the separation of oil from water in oil-in-water (O/W) emulsions was investigated using the ultrafiltration process and artificial neural network (ANN) modeling. The goal of this research was to assess the capabilities of the ANN model in predicting the performance of the ultrafiltration system and to assist in optimizing the separation process. The results indicate that the ANN model effectively predicted the system's behavior, leading to improved efficiency and enhanced adaptability of the oil-water separation system. The combination of ANN modeling with the ultrafiltration process holds significant potential for reducing operational costs and improving the quality of industrial wastewater treatment. Specifically, the accuracy of different ANN models used in this study, based on simulations and data evaluations under various conditions, was as follows: for a single hidden layer with 80 neurons, the prediction accuracy was 0.82; for a single hidden layer with 120 neurons, the accuracy was 0.80; for a network with two hidden layers [80, 40] neurons, the accuracy was 0.90; and for a network with two hidden layers [120, 50] neurons, the accuracy was 0.89. These results demonstrate the high accuracy of ANN models in predicting the behavior of the ultrafiltration system for oil-water separation, highlighting their effective application in optimizing industrial processes and reducing costs in wastewater treatment industries.