Chemometric approaches to data assessment for a long-term case study of MIEX pretreatment performance

Abstract

A specifically designed pilot plant facility located at the Mount Pleasant Water Treatment Plant in South Australia was used to investigate the efficiencies of several drinking water treatment processes, including (1) magnetic ion exchange (MIEX) as a pretreatment followed by conventional coagulation treatment, (2) conventional coagulation at pilot the plant alone, (3) MIEX followed by microfiltration (MF), and (4) MF alone. Dissolved organic matter (DOM) removal efficiency of the treatment processes and more importantly the treated water quality was assessed using dissolved organic carbon (DOC) measurement and together with the changes in the chromophoric organics character based on their molecular weight profiles determined by high-performance size exclusion chromatography (HPSEC) were also reported. Cluster analyses (CAs) were performed using the HPSEC peak area (after separated by peak fitting), and supported by DOC, UV absorbance at 254 nm, and SUVA in order to compare treatment efficiencies of these technologies. It was a novel way of applying CA as a data mining tool to interpret and assess DOM removal results. From the results of these CA, it was found that the MIEX process provided consistent treatment performance and highest removal of DOM, as well as removal of a broad range of molecular weight organics. In comparison coagulation with alum tended to remove the high molecular weight (>1000 Da) compounds. This statistical approach provided improved understanding of the performances of the treatment processes, investigated at a molecular level, for the removal of DOM.