Influent Characterization and Primary Clarifier Modelling for Digital Twin Enhancement at Viikinmäki WWTP

This thesis presents the results of a sampling and analysis campaign conducted at the Viikinmäki wastewater treatment plant (Helsinki, Finland). The collected data were made available to the DIGICARBA project team, which used them to calibrate the plant digital twin currently under development. For this reason, the sampling strategy, laboratory analyses, and data processing were designed and carried out in accordance with guidelines, standards, and methods specifically developed for wastewater treatment plant modelling. During six sampling days, a 24-hour sampling programme was implemented, consisting of eight 3-hour composite samples per day. These samples were analysed to assess the diurnal variation of organic matter and suspended solids loads at three strategic points within the plant. These measurements enabled a detailed influent characterisation consistent with the requirements of digital twin modelling tools. In parallel, the availability of dimensional parameters and operational information of the primary clarifier enabled the development of prototype models of this unit. The modelling work was performed using SUMO simulation software, testing three different approaches: the Volumeless Point Separator (VPS), the Layered Flux Model (LFM), and the Three Compartment Model (TCM). Their performances were evaluated using the Mean Absolute Percentage Error (MAPE). The models were calibrated using the data from the present campaign and validated against those from a previous one. This comparison allowed the identification of modelling strategies that could serve as a foundation for extending the digital twin model to include the primary clarifier.