Skip to Main Content
 

Global Search Box

 
 
 
 

Files

File List


37951.pdf (1.91 MB)

ETD Abstract Container

Abstract Header

Spatial and Temporal Modeling of Water Demands for Water Distribution Systems

Oliveira, Paulo Jose A
http://orcid.org/0000-0001-6352-6198
http://rave.ohiolink.edu/etdc/view?acc_num=ucin1613748818835557

Abstract Details

2020, PhD, University of Cincinnati, Engineering and Applied Science: Environmental Engineering.
Many benefits can be realized by real-time modeling and controlling of water distribution systems. However, such improvements can only be achieved if the driving conditions of the hydraulic model are continuously updated. Water demands are arguably the most challenging of those inputs due to complex spatial and temporal variations present over short times. The overall objective of this dissertation is to contribute to the spatial estimation and forecasting of water demands with probabilistic algorithms. Only after the water demands and associated uncertainty are accurately estimated, further demand prediction and uncertainty propagation can be properly estimated to support the real-time management of WDS. The spatial modeling of water demands, in the current dissertation, was improved by finding the set of demand clusters that best approximate the actual spatial distribution of water demands. The application results demonstrated that the hydraulic likelihood was the best and only necessary metric needed to identify water demand clusters. Additionally, the analysis of the tradeoff between hydraulic likelihood and the overall mean of the maximum demand pattern differences could be used to identify the best number of clusters. A high quality cluster solution was achieved with a relatively small number of additional sensors while utilizing a realistic amount of computational power. The current dissertation also studied the improvement of water demand estimation by evaluating several alternative priors under both normal operating and failure scenarios. Overall, the application results demonstrated the benefits of priors, such as the Seasonal Uncertain Autoregressive (SUAR) model, that incorporate a longer past-history of information to maintain accurate demand estimation performance under both normal operating conditions and failure modes. Additionally, two forecast methods, k nearest neighbors (knn) and a seasonal autoregressive model (SAR), were compared in terms of water demand prediction performance. The evaluation showed that the knn model applied over a differentiated water demand series was the most promising option with sharpness values much lower than any other model investigated while maintaining similar reliability. Additionally, the error metrics for the knn were much smaller than the SAR, which indicates the superiority of the knn method in comparison with an autoregressive strategy under the evaluated pattern conditions. From a broader perspective, the results illustrate that improved methods for the spatial estimation of water demand and forecasting are fundamental for promoting water demand accuracy and the associated uncertainty. In the future, the estimated spatial water demands and uncertainty information can be further propagated to more realistically predict WDS states, thereby improving the overall WDS management process in real-time.
Patrick Ray, Ph.D. (Committee Chair)
Dominic Boccelli, Ph.D. (Committee Member)
Xi Chen, Ph.D. (Committee Member)
Drew McAvoy, Ph.D. (Committee Member)
146 p.
Environmental Engineering
water demand; demand estimation; clustering; water distribution; time series; node grouping

Recommended Citations

Citations

  • Oliveira, P. J. A. (2020). Spatial and Temporal Modeling of Water Demands for Water Distribution Systems [Doctoral dissertation, University of Cincinnati]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=ucin1613748818835557

    APA Style (7th edition)

  • Oliveira, Paulo Jose. Spatial and Temporal Modeling of Water Demands for Water Distribution Systems. 2020. University of Cincinnati, Doctoral dissertation. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=ucin1613748818835557.

    MLA Style (8th edition)

  • Oliveira, Paulo Jose. "Spatial and Temporal Modeling of Water Demands for Water Distribution Systems." Doctoral dissertation, University of Cincinnati, 2020. http://rave.ohiolink.edu/etdc/view?acc_num=ucin1613748818835557

    Chicago Manual of Style (17th edition)

ucin1613748818835557
40
© 2020, some rights reserved.Creative Commons License Spatial and Temporal Modeling of Water Demands for Water Distribution Systems by Paulo Jose A Oliveira is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License. Based on a work at etd.ohiolink.edu.
This open access ETD is published by University of Cincinnati and OhioLINK.