Application of an artificial neural network-genetic algorithm methodology for modelling and optimization of the improved biosorption of a chemically modified peat moss: kinetic studies

摘要

In this study, peat moss was chemically modified using a hot-alkali treatment technique and subsequently used to remove Cu(II) and Pb(II) ions from synthetic solutions. Batch kinetic studies were carried out to elucidate the mechanisms of biosorption. Process operational parameters such as agitation, particle size, conductivity and pH were varied. This method of hot-alkali treatment was successful in reducing the occlusion of pores and resulted in greater adsorptive performance. The kinetic behaviour of the treated peat moss was best simulated by the diffusion-chemisorption model. Film diffusion and intraparticle diffusion were the dominant transport mechanisms. An artificial neural network (ANN) was used to construct a predictive model built-in with the joint effect of the operating parameters. A comparison with the experimental data revealed a significantly high coefficient of determination of 0.9965. The Garson connection weight method showed reaction time as the most influential parameter. Artificial neural network-genetic algorithm (ANN-GA) optimization revealed that maximum biosorption could be obtained using pH 5.5, particle size 0.21 mm, agitation 690 rpm, conductivity 290 mu S/cm and contact time 50 min. The ANN-GA prediction was verified through subsequent laboratory experiments which revealed an excellent prediction with 2.8% residual error. The findings of this study serve to improve the performance of peat biosorption as well as presents a predictive model which can aid in process scale-up.