Comparative Study of Response Surface Methodology and Adaptive Neuro-Fuzzy Inference System for Removal of 6-APA

The antibiotic-contaminated water treatment is an important step for pollutant reduction and the promotion of water environment quality. Uncertainty in wastewater treatment technology, fluctuations in effluent water quality, and operation costs cause an emerging issue to develop materials effective for the removal of antibiotics. The environment-friendly clay such as vermiculite could be potentially promising candidates for removing 6-APA (6-aminopenicillanic) from pharmaceutical effluent. Antibiotic removal was achieved by using an eco-friendly, time-saving, powerful, and easy applying synthesis method via tetraethoxysilane (Si). Expert systems are widely powerful tools for minimizing the complexities and complications in wastewater treatment. Response surface methodology (RSM) and adaptive neuro-fuzzy inference system (ANFIS) models were used to develop systematically predicting interactions of synthesis conditions on 6-APA adsorption capacity and optimize the best amount of compound. The three parameters of the amount of adsorbent (weight.), initial concentration (mg/mL), and reaction time (min) are selected as input and the adsorption capacity (mg/g) were computed as the output of the models. The effect of process variables investigated by RSM through central composite design matrix and the results compared with ANFIS model. The maximum amount of adsorption capacity predicted by RSM for VMT and VMT-Si were 162.5 and 179.8 mg/g, respectively. The suggested models were successfully validated with the acceptable confidence levels 0.99 and 0.97, for VMT and VMT-Si using RSM and 0.99 and 0.99 by ANFIS. ANFIS model demonstrated higher predictive capability than RSM model based on the good agreement in predictable dataset to experimental data.