Linear and Non-Linear Modelling of Bromate Formation during Ozonation of Surface Water in Drinking Water Production

Bromate formation is a complex process that depends on the properties of water and ozone used. Due to fluctuations in quality, surface waters require major adjustments treatment process. In this work, we investigated how time year, dose duration, ammonium affect bromides, bromates, absorbance at 254 nm (UV254), near-infrared (NIR) spectra, fluorescent components (humic-like tyrosine-like) during ozonation. Linear non-linear models were used determine predict relationships between input output variables. Season, ozonation correlated with variables, while affected only bromates. All coefficients determination (R2) for multiple linear regression >0.64, R2 piecewise was >0.89. The season had no effect bromate either model, bromides Three variables influenced UV254 both models. artificial neural network (ANN) model season, time, ammonium, NIR spectra an effective way describe results. multilayer perception 14-14-5 lowest errors best ANN values training, testing, validation 0.9916, 0.9826, 0.9732, respectively.