Optical sensors in drinking water production : Towards automated process control in relation to natural organic matter

Access to safe and clean drinking water is a basic human right (A/RES/64/292). In Sweden, large drinking water treatment plants use mainly surface water as water source. The long-term trend of increasing natural organic matter (NOM) in boreal and north European surface waters negatively affects the overall performance of the treatment processes. To address this issue, sensors are increasingly used as a tool for real-time analysis of water quality providing early warning of potential contamination and decision support for process control. In this thesis, absorbance- and fluorescence-based sensors were used to estimate dissolved organic matter (DOM) concentrations in two Swedish rivers prior to managed aquifer recharge (MAR) and their accuracy was compared (Paper I and IV). The possibility of coupling a coagulation treatment with MAR was explored at laboratory-scale. Two pilot-scale experiments using granular activated carbon filtration were carried out to optimise DOM removal (Paper II). A recent method for molecular DOM analysis was tested to investigate the effect of ozone on low molecular weight compounds. An open-source Python toolbox called “AbspectroscoPY” was developed to pre-process the large amount of absorbance-based sensor data and compute a range of spectral metrics from the time-series data. This allowed a preliminary identification of variability in the spectrophotometric profiles of treated water as a step forward toward automated early warning systems (Paper III). An algorithm for turbidity compensation of the raw absorbance spectra was added (Paper IV). This thesis contributes to an increased knowledge on NOM removal in water treatment using high frequency sensor data from optical sensors