Monitoring the microparticle transfer process in wastewater treatment systems is crucial for improving treatment performance. Supervised deep learning methods show high performance to automatically detect particles, but they rely on vast amounts of labeled data for training. To overcome this issue, we proposed a semi-supervised learning (SSL) method based on the Simple framework for Contrastive Learning of visual Representations (SimCLR), to detect microparticles free from sludge and attached to sludge. First, we pre-trained a ResNet50 backbone by SimCLR, to extract features from much unlabeled data (1,000 images). Then, we constructed a Mask R-CNN architecture based on the pre-trained ResNet50, and fine-tuned it on a small quantity of labeled data (≈200 images with ≈600 annotated particles) in supervised learning fashion. We showcased its performance and practical applicability for microscopy images obtained from the water lab of TU Delft. The results demonstrate that the SSL methods obtain a significant improvement in mean average precision of up to 5% compared to the conventional supervised learning method, when a limited amount of labeled data is available (e.g., 91 labeled images). Furthermore, these methods improve the average precision for detecting attached particles by over 12%. With the detection results from the SSL methods, we measured the attachment efficiency of microparticles to sludge under varying mixed liquor suspended solids concentration and aeration intensity. The precise measurements demonstrate the effectiveness and practical applicability of the SSL method in facilitating long-term monitoring of particle transfer processes in biological wastewater treatment systems.

The aerobic granular sludge (AGS) is an emerging technology widely spread, since most organic matters in actual domestic sewage were particulate matters, this study aims to determine whether the attachment between micro particles and different sized AGS was influenced by granule surface area. The attachment of micro particles by different sized AGS (2.0–5.0 mm) were investigated. Furthermore, to simulate the attachment by broken fragments of AGS, complete 4.0–5.0 mm AGS were cut into 2,4, and 8 pieces, and the attachment performance between the broken pieces and similar sized complete AGS were compared. Fourier transform infrared (FTIR) and fluorescence staining were applied to analyze the chemical bonds and amyloid-glucan like structure of AGS from outside to inside. The results showed the 3.1–4.0 mm AGS had the best surface area attachment of micro particles, followed by the 2.5–3.1 mm AGS. The attachment performance of micro particles was not determined by specific surface area, but was closely related to the surface roughness caused by the amyloid-glucan like structure. The distribution density of amyloid-glucan like structure decreased from outside to inside, and if an granule was broken into pieces during aeration, micro particles were preferential to be attached by the outer layer of the broken pieces from the initial granule. The micro particles attachment showed little relationship with the hydrophilicity of AGS surface, either the outer layer or the inner layer. This study highlighted the crucial role of AGS outer layer in micro particle attachment, particularly the broken pieces from the original AGS outer layer, which facilitate to attach micro particles and contribute to form new granules.

>50 % of the organic matter in sewage consist of particulate chemical oxygen demand (pCOD). This study used 250 μm fluorescent microbeads, 130±58 μm microparticles and 100 nm nanobeads to simulate sewage particles, and investigated the fate of these particles under both plug flow feeding and aeration phases in an aerobic granular sludge (AGS) system. Filtration performance was dominantly influenced by the particle size rather than the upflow velocity (V_upflow). The microbeads exhibited 95±3 % filtration efficiency with obvious accumulation around the AGS bed bottom, even as slight fluidization started at the V_upflow of 5.0 m·h^-1. In contrast, the nanobeads filtration efficiency was significantly lower (43±6 %). During the aeration phase, the attachment efficiency increased with the decrease of particle size. The microbeads attachment efficiency variated between 39–49 %, whereas the microparticles and nanobeads achieved better attachment of 89.4–95.2 % and 98.8–99.3 %, respectively. Furthermore, aeration batch tests showed both nanobeads and the irregular microparticles attachment by AGS was strong, and the detach-attach of nanobeads/microparticles between different sized AGS was very limited duration aeration. This work provides insight into the fate of particles in AGS system. The optimal sludge treatment was also evaluated in the scope of this removal of non-biodegradable, and potentially harmful particles.

Investigating the interaction between influent particles and biomass is basic and important for the biological wastewater treatment. The micro-level methods allow for this, such as the microscope image analysis method with the conventional ImageJ processing software. However, these methods are cost and time-consuming, and require a large amount of work on manual parameter tuning. To deal with this problem, we proposed a deep learning (DL) method to automatically detect and quantify microparticles free from biomass and entrapped in biomass from microscope images. Firstly, we introduced a “TU Delft-Interaction between Particles and Biomass” dataset containing labeled microscope images. Then, we built DL models using this dataset with seven state-of-the-art model architectures for a instance segmentation task, such as Mask R-CNN, Cascade Mask R-CNN, Yolact and YOLOv8. The results show that the Cascade Mask R-CNN with ResNet50 backbone achieves promising detection accuracy, with a mAP50_box and mAP50_mask of 90.6 % on the test set. Then, we benchmarked our results against the conventional ImageJ processing method. The results show that the DL method significantly outperforms the ImageJ processing method in terms of detection accuracy and processing cost. The DL method shows a 13.8 % improvement in micro-average precision, and a 21.7 % improvement in micro-average recall, compared to the ImageJ method. Moreover, the DL method can process 70 images within 1 min, while the ImageJ method costs at least 6 h. The promising performance of our method allows it to offer a potential alternative to examine the interaction between microparticles and biomass in biological wastewater treatment process in an affordable manner. This approach offers more useful insights into the treatment process, enabling further reveal the microparticles transfer in biological treatment systems.

To investigate energy-saving approaches in wastewater treatment plants and decrease aeration energy consumption, this study successfully established a floc-granule coexistence system in a sequencing batch airlift reactor (SBAR) employing micro-bubble aeration. The analysis focused on granule formation and pollutant removal under various aeration intensities, and compared its performance with a traditional floc-based coarse-bubble aeration system. The results showed that granulation efficiency was positively associated with aeration intensity, which enhanced the secretion of extracellular polymeric substances (EPSs) and facilitated granule formation. The SBAR with the micro-aeration intensity of 30 mL·min-1 showed the best granulation performance (granulation efficiency 52.6%). In contrast to the floc-based system, the floc-granule coexistence system showed better treatment performance, and the best removal efficiencies of NH4+-N, TN, and TP were 100.0, 77.0, and 89.5%, respectively. The floc-granule coexistence system also enriched higher abundance of nutrients removal microbial species, such as Nitrosomonas (0.05-0.14%), Nitrospira (0.14-2.32%), Azoarcus (2.95-12.17%), Thauera (0.43-1.95%), and Paracoccus (0.76-2.89%). The energy-saving potential was evaluated, which indicated it is feasible for the micro-aeration floc-granule coexistence system to decrease the aeration consumption by 14.4% as well as improve the effluent.