Applying an artificial neural network to predict coagulation capacity of reactive dyeing wastewater by chitosan

Abstract

Chitosan derived from waste shrimp shells was experimentally evaluated to treat reactive dye (red 24) in aqueous solution. The research was conducted using one-factor-at-a-time experiment design (i.e., pH, reaction time, agitation speeds, initial dye, and chitosan concentration) to investigate dye removal efficiencies by reducing colour and COD parameters. The results obtained by performing Jar-tests indicated that prepared chitosan successfully removed the reactive dye in aqueous solution. In particular, 99.5% of colour and 72.7% COD removal efficiencies were recorded under neutral conditions (pH 7) with chitosan dosage 80 mg/L and agitation speed 60 rpm in 30 min. Fourier transform infrared spectra (FTIR) of the chitosan, dye, and formed sludge were also investigated to elucidate the removal mechanism of the dye by chitosan. These indicated the potential of using chitosan as a “green” coagulant to reduce pollutants of textile wastewater. Moreover, three-layer feed-forward Artificial neural network (ANN) models were applied to model the coagulation processes with the determination coefficient (R2) 0.986 and the root mean square error (RMSE) 2.951 between predicted and observed outputs. The ANN models were analysed with the connection weight method, neural interpretation diagram, and 3D/contour plots to study the influences of operation factors – both individually and combined – on both colour removal and COD removal process.