▎ 摘　　要

In this work, the heavy metal ions (lead, Pb) adsorption process were studied using an artificial intelligence simulation based model for prediction of the adsorption process by using magnetic ash/graphene oxide (GO) nanocomposite. Also, the adsorption mechanism of Pb ions on the adsorbent were investigated using molecular dynamics (MD) calculations in aqueous solution. Reactivity of structures, ionization energy (I), electron affinity (A), chemical hardness (eta), chemical softness (sigma), and energy gap (Delta E-gap) of all compounds were obtained from the HOMO-LUMO energy levels. The outcomes demonstrated that the adsorption of Pb ions on the adsorbent occurred through electrostatic interactions and van der waals bonding and the lead-water-GO configuration had the highest adsorption affinity according the Delta Egap calculations. The artificial neural network (ANN) with two hidden layers was used for developing the model with a mixture of linear and non-linear transfer functions. The equilibrium (Eq.) concentration of the Pb ion as an important factor in predicting the adsorption capacity of adsorbent was considered for the model output and initial Pb ion concentration as well as solution temperature were assumed as the model inputs. The training and validation procedure of ANN indicated great agreement between the experimental and predicted data according to the high coefficient of determination and low root mean square error (R-2 > 0.999, RMSE = 0.086). Based on the simulation results increasing the initial concentration of Pb ion significantly affect the Eq. concentration while the solution temperature had a lower effect on Eq. concentration. The results of this study provide valuable model for pollutants removal. MD calculations and artificial intelligence simulation methods could be an appropriate combined technique for predicting the adsorption behavior of nanocomposite in heavy metal ions removal from the aqueous solution with high accuracy. (C) 2021 Elsevier B.V. All rights reserved.