Comparative Analysis Study by Response Surface Methodology and Artificial Neural Network on Salicylic Acid Adsorption Optimization using Activated Carbon

Abstract In this wok, the performance of Activated Carbon (AC) for removal of Salicylic Acid (SA) from synthetic solution by adsorption process is studied. Along with this to minimize the number of experiments and obtain optimal conditions, a multivariate predictive model based on Response Surface Methodology (RSM) is developed. This is compared with data-driven modeling using the Artificial Neural Network (ANN) for prediction of the adsorption of SA. The interactive effects on SA removal efficiency with respect to independent process variables were investigated. A comparison between the model results and experimental data gave a high correlation coefficient (R2 (ANN) = 0.99, R2 (RSM) = 0.91). The results revealed that both models were able to predict removal of SA by AC, while the ANN models were slightly more accurate in predictions as compared to RSM models. The Sips, Toth, Langmuir-Freundlich isotherm equations were applied to the equilibrium data and the results revealed that sips isotherm (R2 = 0.998) had better correlation than the other isotherm (R2 = 0.9461).