Comparative analysis of RSM, ANN and ANFIS and the mechanistic modeling in eriochrome black-T dye adsorption using modified clay

Abstract

ABSTRACT: The application of artificial neural network (ANN), response surface methodology (RSM), and adaptive neuro-fuzzy inference system (ANFIS) in modeling the uptake of Eriochrome black-T (EBT) dye from aqueous solution using Nteje clay was the focus of this work. Acid activation with hydrochloric acid (HCl) was used to prepare the adsorbent while Fourier transform infrared spectroscopy (FTIR) and scanning electron microscopy (SEM) were utilized in the characterization of the adsorbent. The ANN, RSM, and ANFIS models were analyzed by considering the adsorbent dosage, contact time, solution temperature, and pH of the adsorption process. Sensitivity analyses involving six statistical error functions were further used to compare the acceptability of the models. Four mechanistic models (Weber and Morris, Film diffusion, Bangham, and Dummwald-Wagner models) were used to determine the mechanism of the EBT uptake. The result showed that the activation process enhanced the adsorption capacity of the clay. The ANFIS, ANN, and RSM models gave a high accuracy in predicting the adsorption of the EBT dye with correlation coefficients of 0.9920, 0.9910, and 0.9541, respectively. Further statistical indices lent credence to ANFIS as the best predictive model and RSM the least in adsorption of EBT dye. Process optimization using genetic algorithm gave optimum adsorption efficiency of 95.8%. Mechanistic modeling indicated film diffusion as the rate-limiting mechanism. The maximum amount of EBT adsorbed was 24.04 mg/g. The HCl-modified clay could be utilized as an efficient adsorbent in EBT uptake from wastewater.

Keywords

• Artificial neural network
• Response surface methodology
• Adaptive neuro-fuzzy inference system
• Adsorption
• Eriochrome black t
• Mechanistic modeling