With the abundance of data availability across treatment facilities, artificial intelligence is paving the way for unique optimization opportunities. The examples provided encompass the Agua Nueva Water Reclamation Facility (WRF), which concentrates on minimizing aeration energy expenses and enhancing nutrient management, as well as the Wilmington WWTF, emphasizing the optimization of disinfection chemicals. Several artificial intelligence (AI) and Bayesian modeling frameworks were developed with both sites using AI algorithms with mean absolute percentage errors < 10% for both forecasting of future conditions at a facility, as well as causal inferencing for predicting water quality. Post deployment, both sites are regularly achieving anywhere from 10-30% savings in energy or chemical usage. These case-study demonstrates significant progress in the successful implementation of AI at a treatment facility and the ability to aid in the empowerment of treatment plant operators while improving efficiencies in wastewater treatment.

Blue Plains Advanced wastewater treatment plant (AWWTP) was the first utility in the USA to install Cambi thermal hydrolysis process (THP) together with anaerobic digester (AD) to reduce solids production, increase energy resource recovery and produce class A biosolids. The new facilities were commissioned at the Blue Plains Advanced Wastewater Treatment Plant (AWTP) in 2015. The technology choice provided a major benefit on dewatering process by improving the cake TS, which has a large economic impact on treatment operations by decreasing the hauling costs, and by improving the quality of the "Bloom” biosolids product that is regionally marketed (Li, 2021). Dynamic changes in operational condition of the THP and AD process, sludge characteristics, and seasonal impacts (Li, 2021) result in high variety of cake TS (%) from 24 to 38% TS and filtrate TSS from 50 to 1530 mg TSS/L at the final dewatering belt filter presses. Poor cake TS causes higher hauling costs, while unreliable filtrate TSS significantly impacts the filtrate treatment facility. In addition, the polymer cost also may increase year by year depending on the market price, so picking the right optimal polymer dose range can optimize operational cost. In practice, the polymer dose is set manually by the operator and requires constant attention to minimize polymer dose while maintaining good operational throughput. Polymer dose decision making is thus mainly driven by human factors. Therefore, this study focused on investigating if the polymer dosing decision made by operators is adequate or if there is room to improve dewatering performance, which helps to define research development needs for helping optimizing polymer costs at the plant.

The primary source of nitrogen and phosphorus in municipal wastewater treatment plants is human urine. One of the primary concerns associated with the use of urine-based fertilizers in agriculture is the presence of micropollutants, including pharmaceutical residues, which are co-excreted with nutrients in urine. The present study experimented sewage sludge biochar pyrolyzed at a temperature of 700°C and its modified form using dimethyl sulfoxide (DMSO) for selective adsorption of antibiotics in human urine. The adsorption isotherm and kinetics can be adequately described by the Langmuir and pseudo-second order, respectively. The maximum adsorption capacities of azithromycin, sulfamethoxazole, trimethoprim, and tetracycline were found to be 385, 270, 357, and 1,000 μg/g, respectively.

The City of Sunnyvale Water Pollution Control Plant (WPCP) is a 12 MGD Average Dry Weather Flow tertiary wastewater treatment facility (1). As part of the secondary system, the plant operates oxidation ponds (2) for removal of organic matter and some ammonia. The ponds are rich in nutrients and as such propagates algae. Prior to 2021, the dominant algae species in the ponds were nano-, micro- and mesoplankton species that consisted of some cyanobacteria, protozoa, small algae, and some large algae ranging in size from 2 micron to 2 mm. These species served the plant well with photosynthesis providing needed oxygen for the oxidation reactions and nutrient uptake by algae. During the early part of 2021, the ponds became dominated by picoplankton species, namely a unicellular cyanobacterium (also known as bluegreen algae(3) ) named Synechocystis with a size range of 0.2 to 2 micron. Granted Synechocystis continued to serve the plant well with photosynthesis and nutrient uptake, however, it upended our plant by not allowing itself to be removed in our Dissolved Air Flotation (DAF) tanks where the pond effluent is treated for solids/liquid separations. As a result, the plant violated its NPDES permit limit on multiple occasions for solids discharging to the lower San Francisco Bay(4), resulting in possible fines to the City. This paper discusses the research work conducted this year in terms of water chemistry at the DAF in determining the most optimal condition to remove Synechocystis from the pond effluent. This effort not only prevented the plant from violating the NPDES permit limit for solids again but also provided the City with a cost-effective solution.

With the abundance of data availability across treatment facilities, artificial intelligence is paving the way for unique optimization opportunities. The examples provided encompass the Agua Nueva Water Reclamation Facility (WRF), which concentrates on minimizing aeration energy expenses and enhancing nutrient management, as well as the Wilmington WWTF, emphasizing the optimization of disinfection chemicals. Several artificial intelligence (AI) and Bayesian modeling frameworks were developed with both sites using AI algorithms with mean absolute percentage errors < 10% for both forecasting of future conditions at a facility, as well as causal inferencing for predicting water quality. Post deployment, both sites are regularly achieving anywhere from 10-30% savings in energy or chemical usage. These case-study demonstrates significant progress in the successful implementation of AI at a treatment facility and the ability to aid in the empowerment of treatment plant operators while improving efficiencies in wastewater treatment.

Solids thickening is an indispensable unit process for process intensification of biosolid s treatment. The fluctuations of feed sludge characteristics often impact the stability of the thickening process, reduce process capability, and increase operation and maintenance costs. Often, the optimization of sludge thickening is focused on meeting performance requirements as detailed in specifications. This study expands the assessment of sludge thickening by including both thickening equipment and its up- and downstream infrastructure from a process capability perspective. The presentation will benefit engineers and utility operators in better design and troubleshooting of sludge thickening processes with designed thickened sludge solids total equal to or higher than 6%. The presentation will provide practical suggestions for improving the process capabilities of sludge thickening, including pumping improvement, pipe flushing using digested sludge, and improvements on sensors.

Water utilities are under increasing pressure to maintain plant performance while reducing operational cost and carbon, due to population density, ageing infrastructure and climate change. Digital technology can address these challenges through the application of data science and advanced analytics. The Rosedale Wastewater Treatment Plant in Auckland, New Zealand, consists of primary, secondary, and tertiary treatment and serves a population equivalent of 240,000. After creating a digital twin to support commissioning of a fourth reactor in 2020, it was identified that the digital twin could provide new insights into operational cost and greenhouse gas and support decisionsmaking, to help Watercare achieve their objective to reduce operational greenhouse gas by 50%. Live dashboards were developed to provide new insights into operational greenhouse gas emissions and cost which leveraged real-time biological modelling and emissions monitoring data. Soft sensors were developed to improve the instrumentation data quality through early identification of sensor drift.