Integrating Activated Sludge Floc Size Information in Membrane Bioreactor (MBR) Fouling Modeling

In this dissertation, the mechanism of cake layer formation is described by an MBR fouling model. One of the important inputs of the model is the particle size distribution of activated sludge. This information is the result of an in-house developed image analysis software. The program has proven itself as an effective and reliable software in activated sludge image analysis. However, it used to experience some difficulties when dealing with some less common activated sludge morphologies and it resulted in less accurate information of the activated sludge characteristics. The three common issues which are (i) images with a lot of small filaments that are not attached to flocs, (ii) images with dark regions in flocs and (iii) images with bright filaments are solved by improving and adding new algorithms into the segmentation and recognition part. These adjustments and the implementation of an overall simplification of the algorithm significantly improve the image analysis procedure in terms of the quality of the resulting information and the computation time. The filtration model is developed based on the combination of the model proposed by Broeckmann et al. [15] with the particle size distribution of activated sludge and particle deposition rules inspired by Yoon et al. [134]. A three dimensional cake layer with multi-size particles is created and it can provide the required input information (e.g., the porosity profile of the cake layer and the cake layer thickness) for the fouling model. A very close resemblance between the model predictions and the experimental data is not obtained due to a lack of information on the particle size distribution of the activated sludge in the available data, but the correct trend of the transmembrane pressure (TMP) evolution is predicted: first a rather fast increase due to pore blocking, then a very slow increase and finally a TMP jump. The strong influences of particle size distribution, shear stress and compression on the characteristics of the formed cake layer are demonstrated by the model. Finally, in order to complete the operation of an MBR system, a physical cleaning model is developed and integrated in the filtration model. This model is basically based on the removal of the top layers during the relaxation phase and the washing out of the particles which were deposited above the pores during the backwashing phase. The latter takes into account a fraction of pores that will stay permanently blocked. The former