DOC signal-based alum dose control for drinking water treatment plants

This paper describes a model, based on dissolved organic carbon (DOC) input data, developed for prediction of alum dosage for targeted removal of DOC from raw surface waters. DOC, UV absorbance at 254 nm/cm (A254), colour and turbidity data were acquired for raw and treated water samples. The results show that for 17 samples jar tested, only-57 % & PLUSMN; 13 % (mean & PLUSMN; SD) of the DOC content was coagulable. For the range of samples studied, it was found that 85 % (& PLUSMN;8 %) removal of coagulable DOC fraction was achieved at enhanced coagu-lation (EnC) based on & UDelta;DOC/& UDelta;Alum = -0.015 mg/mg. The 85 % value was then used as a basis to develop an alum dose prediction model for application over the wide range of water qualities of the samples studied. The model outputs were compared with EnC doses (EnD) obtained through jar testing, as well as predicted EnD values obtained through mEnCo & COPY; software, that uses input data of A254 and colour. Trials of the model pre-dictions were conducted for five drinking water treatment plants (DWTPs) over a four-month period. The results show the potential of the model for alum dose predictions with correlation coefficients of predicted alum doses versus 1) actual alum doses and 2) estimated EnD from the mEnCo & COPY; software ranging from 0.72 to 0.98 and 0.92 to 0.99, respectively. It is suggested that the DOC model has potential to be deployed for online feedforward alum dose control.

Automated summary

This paper presents a model for predicting alum dosage for targeted removal of dissolved organic carbon (DOC) from raw surface waters using DOC input data. The model was developed based on jar testing results and showed a 85% removal rate of coagulable DOC fraction at enhanced coagulation. It was further validated through trials at five drinking water treatment plants, showing a high correlation between predicted alum doses and actual alum doses.