OPTIMIZATION OF METHYLENE BLUE ADSORPTION PROCESS USING ARTIFICIAL NEURAL NETWORKS AND LEAST SQUARES SUPPORT VECTOR MACHINE METHOD

Abstract

The water pollution problem is becoming increasingly serious. However, the wastewater treatment process has not been carried out effectively causing a large amount of toxic substances being discharged directly into the water environment without treatment. Methylene Blue (MB) is an organic dye, widely used in many different fields such as medicine, biology, chemistry and industry. Using MB in large concentrations can cause significant health issues such as heart damage, vomiting, shock, and limb paralysis. This research aims to optimize the adsorption process for MB using a composite of Zinc Sulfate Nanoparticles and Activated Carbon (ZnS NPs/AC). We utilized Artificial Neural Networks (ANN) and Least Squares Support Vector Machine (LS-SVM) models to identify the optimal conditions for MB adsorption. The performance of the models was assessed by their determination coefficients (R²) and root mean square errors (RMSE). Results revealed that the LS-SVM model, with an R² of 0.99 and RMSE of 0.24, outperformed the ANN model, which had an R² of 0.98 and an RMSE of 0.74. The optimal adsorption conditions were achieved at a pH of 6.6, MB concentration of 8.8mg/L, adsorbent mass of 0.015g, and ultrasonication time of 4.9 minutes, yielding an adsorption efficiency exceeding 97%.