Phosphorus (P) recovery through struvite crystallization in wastewater treatment plants (WWTPs) is essential for sustainable nutrient management. However, facilities like the Nine Springs Wastewater Treatment Plant (NS WWTP) face inefficiencies in P removal, excessive chemical usage, and inconsistent nutrient recovery, leading to higher costs and operational challenges. Optimizing this process can improve sustainability, reduce expenses, and enhance fertilizer production. This study develops a machine learning (ML)-based predictive model to enhance struvite recovery by analyzing historical operational data. The model accurately predicts orthophosphate removal, conversion, and yield, achieving high predictive accuracy (R² > 0.8). Key factors such as influent phosphorus levels, magnesium dosing, and particulate phosphorus significantly impact reactor performance. This data-driven approach offers a scalable solution for WWTPs, reducing reliance on complex models while improving process efficiency. Future integration into control systems can enable real-time decision-making, optimizing nutrient recovery while lowering operational costs and environmental impact.
Speaker Bio: Rishabh Puri is a second-year Ph.D. student in Professor Roland Cusick’s research group, where he is advancing the application of data-driven machine learning prediction models to evaluate the performance of struvite reactors across the United States. He holds a master’s degree in environmental engineering from the Indian Institute of Technology (IIT), India, and a bachelor’s degree in civil engineering from VIT, India. He recently earned 2nd place in the prestigious Intelligent Water Systems Challenge at IWPC 2025 hosted by the Illinois Water Environment Association. Beyond his research, he enjoys traveling to new places and expressing his creativity through dance.