Conventional biological wastewater treatment plants (WWTPs) include one or more secondary settling tanks. Their primary function is to clarify wastewater and prevent solids to exit the WWTP and enter the receiving water body. In addition, settling tanks help to thicken sludge that is returned to the biological reactors, which ultimately helps to maintain a high biomass concentration. Thickening also increases the concentration of the excess sludge, thereby reducing excess sludge treatment costs. Conventionally, optimal settling tank design and operation is based on state-point analysis [1-3]. This requires the solids flux curve [4], which describes the sedimentation transport under zone settling conditions performance and which is the product of the sludge settling velocity (SSV) and the solids concentration. The SSV is a function of the sludge concentration. The SSV is often estimated with the Sludge Volume Index (SVI) measurements [5-6]. Improved SSV estimates are obtained with batch settling tests. Obtaining reliable data to estimate settling velocities remains challenging however [9] despite commercial availability of dedicated devices [7] and image-based developments [8]. It is hypothesized that this is due to high maintenance of the measurement devices and the sparse knowledge about image analysis in the wastewater modelling community [10]. If automated velocity estimation was possible, advanced control of settlers would enter the realm of possibilities [11-12]. In addition, obtaining information about settling characteristics could also assist in optimizing sequencing batch reactor systems or selecting a highly settleable sludge fraction (e.g., granular systems). With this work, we present a first step toward automated analysis of batch settling experiment image series.
Development of data- and knowledge-based modelling and optimization techniques constitutes the main focus of our efforts. This research is accommodated to three challenging goals: Sensor management - in which sensors are monitored and diagnosed and data quality is assessed and optimized on-line Process intelligence - in which computer-based representations and predictions of process states are used for process monitoring and operation Control - in which automation of aquatic processes is pursued The developed methods are applied to wastewater treatment processes primarily although the generality of our approach supports collaborative and inter-disciplinary projects outside of this application domain.
AET 30.4 Nov 2026
Product directory
.jpg)
.jpg)
.jpg)
.jpg)
