Bottlenecks and Breakthroughs: Developing AI Models for Optimizing Calera Creek’s Wastewater Treatment Operations

Presentation Description: Arup and the City of Pacific worked together to develop two AI/ML models as proof of concepts for sequential batch reactor (SBR) and auto-thermal thermophilic aerobic digestion (ATAD) operations at the Calera Creek Water Recycling plant. Arup and City of Pacifica first identified opportunities through visualizing historic plant data and then developed, trained, and tested chosen models. The project conclusions highlight barriers to implementation of AI/ML in plant operations, and identified necessary next steps to overcome these obstacles.

The Calera Creek Water Recycling Plant, located in Pacifica, California, is a critical facility that uses advanced SBR technology to efficiently combine aeration and clarification processes and ATAD technology to produce Class A sludge. These technical innovations allow the plant to meet the evolving challenges of wastewater treatment, nutrient removal, and environmental sustainability.

With the increasing complexity of its operations—especially during peak storm flows—there are a growing number of opportunities for advanced analytics, rooted in Artificial Intelligence and Machine Learning techniques, to optimize the plant’s operations. These include improving energy efficiency, enhancing process performance, and reducing the risk of unexpected faults or failures through managed predictive maintenance. This presentation will include insights from the Engineers, Operators, and Data Scientists’ perspective, as well as lessons learned from the project team when addressing challenges in developing optimization approaches towards SBR and ATAD operation.

The presentation will also cover how to effectively incorporate stakeholder input, pivot when presented with ambivalent results, and create roadmaps for effective integration of AI models in wastewater treatment plant operation.