UNS: Data Assimilation and Forecasting for Real-Time Drinking Water Distribution System Modeling

[1511959] boccelli饮用水分配系统的实时建模有可能通过改善能源管理、水质维护和响应活动为水务公司提供一些好处。虽然大多数公用事业都有一个用于规划目的的网络模型，但这些模型是“静态的”，无法做出实时决策。因此，迫切需要开发一个实时建模框架，以估计和预测驱动网络动态的未观察到的需求。该项目的目标是发展一种复合需求-水力模型- -一种将需求模型与水力网络求解器结合起来的模型- -能够利用观察到的水力信息实时更新。中心假设是，通常通过公用事业计算机系统收集的观察到的水力数据可用于估计结构化需求模型的期望值和不确定性，该模型表征了消费需求的时空模式。开发实时复合需求-水力模型的基本原理是为与饮用水分配系统相关的实时分析和决策支持提供一个框架。总的来说，这些结果将通过开发复合需求-水力模型和监测放置算法，为实时需求估计和预测提供一种全面的方法。该项目的长期积极影响将来自于提供实时需求估计和预测的能力，这将导致实时分析和决策支持工具，如泵调度，以最大限度地减少能源使用，快速检测主断裂，以及污染事件的传播预测，以更好地确定响应和缓解活动。这些项目将加强学术研究与工业创业之间的联系，为学生提供以下机会：1)为研究生和本科生提供工业研究机会，以扩展我们的研究和发展活动；2)通过开源活动开发和分发我们的软件，以进一步增加我们领域内的机会；3)向更广泛的学生介绍技术转移以及与工业研究相关的机遇和挑战。

1511959BoccelliReal-time modeling of drinking water distribution systems has the potential to provide several benefits for water utilities by improving energy management, water quality maintenance, and response activities. While most utilities have a network model for planning purposes, these models are "static" and not amenable for making real-time decisions. Thus, there is a critical need to develop a real-time modeling framework that can estimate and forecast the unobserved demands that drive network dynamics. The objective of this project is to develop a composite demand-hydraulic model - one that couples a demand model with a network hydraulic solver - capable of being updated in real-time using observed hydraulic information. The central hypothesis is that the observed hydraulic data commonly collected via utility computer systems can be used to estimate the expected values and uncertainty of a structured demand model that characterizes the temporal and spatial patterns of consumptive demands. The rationale for developing the real-time composite demand-hydraulic model is to provide a framework to inform real-time analytics and decision support associated with our drinking water distribution systems. Collectively, these outcomes will produce a holistic approach for real-time demand estimation and forecasting through the development of the composite demand-hydraulic model and monitoring placement algorithm. The long-term positive impacts of this project will result from the ability provide real-time demand estimates and forecasts that will lead to real-time analysis and decision support tools such as pump scheduling to minimize energy usage, rapid main break detection, and spread forecasting of contamination events to better identify response and mitigation activities. The PIs will strengthen the link between academic research and industrial entrepreneurship to provide opportunities to the students through: 1) providing industrial research opportunities to both graduate and undergraduate students to extend our research and development activities; 2) developing and distributing our software through open-source activities to further enhance the opportunities within our field; and, 3) introducing a broader range of students to technology transfer and the opportunities and challenges associated with industrial research.