| dc.description.abstract | This study examines the process of optimizing the amount of polymer added to the sludge that goes into the drum screens at Veas, which is the biggest wastewater treatment plant in Norway. The goal is to reduce the amount of polymer used while still achieving efficient separation of sludge and water. In order to address the absence of direct dry solids measurements at the drum screens, a comprehensive two-week sampling period was carried out to gather crucial data for subsequent analysis.
Three modeling methodologies were investigated in order to estimate the dry solids content and enable the implementation of Model Predictive Control (MPC) in polymer dosing. The approaches consisted of a linear model combining time delay and time constant, a non-linear model obtained through system identification, and a mechanistic model founded on mass balance principles. After a thorough evaluation, the non-linear model was chosen for both estimating dry solids and integrating with MPC.
The study highlights the difficulties and restrictions linked to the implementation of MPC, which requires continuous testing and improvement. Furthermore, a connection is established between the created model and MPC, which is implemented using Python programming language, in order to facilitate communication with the Programmable Logic Controller (PLC) through the Profibus protocol.
Future work will involve integrating a Kalman filter to enhance the accuracy of estimating dry solids and further enhance the efficiency of the process. This study makes a valuable contribution to the improvement of wastewater treatment methods, which has the potential to increase efficiency in operations and promote environmental sustainability. | |
