CISE-MSI: DP: CNS: AI-powered Diagnosis Augmented by Self-sustaining Sensing System for Intelligent Wastewater Infrastructure Management

污水基础设施对现代城市至关重要，但老化的下水道系统往往存在管道破裂和检修孔损坏等缺陷，导致渗透和流入问题。这一问题导致过多的地表径流和地下水流入下水道系统，造成下水道溢出，对公众健康和环境构成威胁。目前的管理方法需要大量的时间和昂贵的资源。该研究项目通过将图神经网络和现场水压监测相结合来应对这一挑战。图形神经网络代理模型将城市污水系统表示为图形，允许对其时间，空间和拓扑属性进行有效建模。此外，通过整合物理传感系统，该研究能够准确地检测渗透和流入异常，并利用图形神经网络骨干预测级联影响，从而及时采取主动和纠正措施。该研究通过利用水文水力科学，嵌入式系统和人工智能的跨学科知识和专业知识，彻底改变了城市污水系统的管理。这项研究的技术成果可以通过提高城市废水系统对气候变化影响的适应能力和促进自然灾害后的快速恢复，大大加强沿海地区的公共安全和健康。通过探索德克萨斯州南部目标废水系统的完全感知数字孪生模型的开发，该项目朝着提高对废水系统的理解和管理能力的方向发展。研究成果与学生的教育和培养紧密结合。此外，该项目还促进了代表性不足的群体和K-12学生参与STEM学习。所有的设计都是公开的，以确保公平地获得人工智能驱动的决策工具，以实现广泛的影响和未来的研究进展。该奖项反映了NSF的法定使命，并通过使用基金会的智力价值和更广泛的影响审查标准进行评估，被认为值得支持。

Wastewater infrastructures are critical for modern cities, but aging sanitary sewer systems often suffer from defects like cracked pipes and damaged manholes, leading to the infiltration and inflow problem. This problem results in excessive surface runoff and groundwater flowing into the sewer system, causing sewer overflows and posing risks to public health and the environment. Current management approaches require significant time and expensive resources. This research project addresses this challenge by combining Graph Neural Networks and in-situ water pressure monitoring. The Graph Neural Networks surrogate model represents the urban wastewater system as a graph, allowing for efficient modeling of its temporal, spatial, and topological properties. Further by integrating a physical sensing system, this research enables accurate infiltration and inflow anomaly detection and predictions of cascading impacts with the Graph Neural Networks backbone allowing proactive and corrective actions to be taken in time.This research revolutionizes the management of urban wastewater systems by leveraging the interdisciplinary knowledge and expertise from hydrological & hydraulic sciences, embedded systems, and artificial intelligence. The technical outcomes of this research can significantly enhance public safety and health in coastal regions by improving the resilience of urban wastewater systems to climate change effects and facilitating quick recovery after natural hazards. By exploring the development of a fully sensed digital twin of the targeted wastewater system in south Texas, the project advances towards increased understanding and management capabilities of wastewater systems. The research results are closely integrated into the education and training of students. Besides, this project promotes the participation of underrepresented groups and K-12 students to pursue STEM studies. All designs are made publicly available to ensure equitable access to Artificial Intelligence-powered decision-making tools for broad implications and future research advances.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.