Machine learning based simulation of water treatment using LDH/MOF nanocomposites

Abstract

In this work we reported a novel methodology in prediction of molecular separation using adsorption process, and understanding the effect of underlying adsorption process on removal of pollutants from water. The data for adsorption of thallium (I) from water was collected and the model was developed based on machine learning (ML) approach. The type of adsorbent studied in this work is Ni50Co50-Layered double hydroxide/UiO-66-NH2 which is a metal organic framework-based nanocomposite material. The adsorbent was selected due to its high capacity in separation and removal of thallium (I) from aqueous solutions with surface area of around 900 m2/g and pore volume of 0.9 cc/g. The modeling and computations were performed using artificial neural network which is a machine learning technique considering the equilibrium concentration of ion in the liquid solution at equilibrium as the main output. Two inputs were considered including temperature and the initial concentration of the adsorbate. The training and validation of the model indicated very high accuracy of the model compared to other modeling approaches with high determination coefficient (R2) more than 0.99 for both training and testing the model stages.