Continuous silicic acid removal in a fixed-bed column using a modified resin: Experiment investigation and artificial neural network modeling

Abstract

In this study, the suitability of a novel modified resin (gallic acid modified resin: GA-type resin) for silicic acid removal was investigated using fixed-bed column adsorption. Laboratory dynamic experiments were conducted at different flow rates, bed heights and influent concentrations. With rising flow rate, the breakthrough time, exhaust time and absorption capacity of the column bed decreased while increasing with column bed height. With the increasing influent concentration from 20 mg/L to 60 mg/L, decrease in both breakthrough time and exhaust time, but an increase in adsorption capacity from 3.73 mg/g to 4.20 mg/g. To simulate the experimental breakthrough curves (BTCs) and predict the column dynamics, basic and empirical models (Bed Depth-Service Time (BDST), Thomas, and Yoon-Nelson models) as well as an artificial intelligence approach (Artificial Neural Network (ANN) model) were applied. All three conventional models were able to reflect well the behavior for silicic acid continuous adsorption with GA-type resin and estimated the characteristic model parameters for removal. ANN model had a higher accuracy (R2 = 0.9997) in predicting BTCs and predicting breakthrough times. Analyzing the parameters obtained from the simulation of the ANN model, the initial concentration was found to have the highest relative importance of 32.94 %, and the specific order of relative importance was as follows: initial concentration > column height > flow rate > total running time. Furthermore, due to its certain regeneration potential, GA-type resin can be recycled, making it a promising material for removing silicic acid from water.