LAUSR

Abstract We developed an artificial intelligence simulation method to predict adsorption process for removal of ions from water via an ordered nanostructure adsorbent. Separation of two ions including Ni and Hg were simulated using the developed artificial intelligence model. The studied adsorbent is a nanocomposite ordered structure made of layered double hydroxide/Metal organic framework structure. Some experimental adsorption data was obtained and employed in the artificial intelligence simulation of the adsorption. The model was built based on artificial neural network including two hidden layers, and different efficient activation functions. The training and validation processes of the neural network was conducted considering two inputs, i.e., type of ion (Hg or Ni) and initial concentration of ion in the solution. Moreover, two outputs were designed in the neural network structure including equilibrium concentration of ion in the solution as well as adsorption capacity of nanocomposite. The data of adsorption were used in the training/validation, and high accuracy with R2 > 0.999 was attained for the simulations of adsorption. It was revealed that the developed machine learning simulation can efficiently predict the adsorption behavior of nanocomposite in removal of ions from the solution, and the accuracy of machine learning model is higher than the well-known adsorption isotherm models.