Prediction of Co(II) and Ni(II) ions removal from wastewater using artificial neural network and multiple regression models

In this research, carboxymethyl chitosan-bounded Fe3O4 nanoparticles were synthesized and used for removal of Co(II) and Ni(II) ion metals from wastewater. The capability of magnetic nanoparticles for metal ions removal was investigated under different conditions namely pH, initial concentration of metal ions and adsorbent mass. The assessment of adsorbent performance for metal ions removal under different conditions requires cost and time spending. In this regard, the capability of artificial neural network (ANN) and nonlinear multi-variable regression (MNLR) models were investigated for predicting metal ions removal. The values of operational parameters such as pH, contact time, initial concentration of metal ions and adsorbent mass were applied for simulation by means of ANN and MNLR. A back propagation feed forward neural network, with one hidden layer (4:8:2), was proposed. Two criteria, including mean square error (MSE) and coefficient of determination (R2) were used to evaluate the performance of models. The results showed that two models satisfactorily predicted the adsorbed amount of metal ions from wastewater. However, the ANN model with higher R2 and lower MSE than the MNLR model had better performance for predicting the adsorbed amount of metal ions from wastewater.