Optimization of adsorption performance of cerium-loaded intercalated bentonite by CCD-RSM and GA-BPNN and its application in simultaneous removal of phosphorus and ammonia nitrogen

Abstract

Excessive phosphorus (P) and ammonia nitrogen (NH3–N) in water bodies can lead to eutrophication of the aquatic environment. Therefore, it is important to develop a technology that can efficiently remove P and NH3–N from water. Here, the adsorption performance of cerium-loaded intercalated bentonite (Ce-bentonite) was optimized based on single-factor experiments using central composite design-response surface methodology (CCD-RSM) and genetic algorithm-back propagation neural network (GA-BPNN) models. Based on the determination coefficient (R2), mean absolute error (MAE), mean square error (MSE), mean absolute percentage error (MAPE), and root mean square error (RMSE), the GA-BPNN model was found to be more accurate in predicting adsorption conditions than the CCD-RSM model. The validation results showed that the removal efficiency of P and NH3–N by Ce-bentonite under optimal adsorption conditions (adsorbent dosage = 1.0 g, adsorption time = 60 min, pH = 8, initial concentration = 30 mg/L) reached 95.70% and 65.93%. Furthermore, based on the application of these optimal conditions in simultaneous removal of P and NH3–N by Ce-bentonite, pseudo-second order and Freundlich models were able to better analyze adsorption kinetics and isotherms. It is concluded that the optimization of experimental conditions by GA-BPNN has some guidance and provides a new approach to explore adsorption performance after optimizing the conditions.