Chlorine dioxide by-products predictive models for drinking water oxidation treatment

This study was aimed at developing statistically based model equations capable of predicting chlorite and chlorate formation during the primary oxidation/disinfection treatment using data collected on two real scale treatment plants. In this phase many operational parameters, raw water quality as well as the second oxidant dose applied, sodium hypochlorite, importantly influence the chlorite and chlorate formation. The operational parameters and the raw water quality correlated with chlorite and chlorate formation were: contact time, chemical doses, temperature, turbidity, organic substances (non-purgeable organic carbon and UV254nm absorbance), pH, iron and ammonium concentration. Given the high numbers of variables two models have been developed. The first one (complex) includes all the studied parameters; the other model (simplified) takes into account only the parameters that gave a significant correlation (>0.55) with the dependent variable. Nearly 60% of the predictions for chlorite formation were observed to be within ±40% of the measured levels and the found key parameters were the water organic content and the physico-chemical characteristics (pH and temperature). Nearly 85% of the predictions for chlorate formation were observed to be within ±40% of the measured levels and the found key parameters were again the chemical physical water characteristics (pH and temperature) and the applied chemicals doses.