An adaptive modeling of petroleum emulsion formation and stability by a heuristic multiobjective artificial neural network-genetic algorithm

Abstract

In this work, experimental and statistical modeling and optimization of process parameters for maximizing the o/w emulsion stability was carried out using the multiobjective artificial neural network-genetic algorithm (ANN-GA) coupled with response surface methodology. The independent model constrains were oil concentration (10–50% v/v), surfactant concentration (2–10% v/v), stirring intensity (2000–6000 rpm), stirring time (5–20 min), and pH (2–12). The responses were turbidity (τ) and emulsion stability index (ESI24). This fact that there is a reasonably good agreement between the experimental data and the predicted values was shown by the modeling results. The optimized conditions predicted by hybrid ANN-GA model to maximize ESI24 ( = 94.71) with 4.8% error were: oil concentration 50% v/v, surfactant concentration 5.571% v/v, stirring speed 6000 rpm, stirring time 5.97 min, and pH 12. The accuracy of the model is confirmed by the comparison between the predicted and experimental data. The proposed hybrid ANN-GA model was found to be useful for the modeling and optimization of process parameters for emulsion stability analysis and the other emulsification process.