Abstract:
This seminar shows the results of two models to improve water distribution systems management based on data-driven and model-based methods. The first application develops a model to forecast water demand using hourly data from smart meters and three machine learning methods. Demands are predicted using lagged demand, seasonality, weather, and household characteristics. Time-series clustering is applied to delineate data based on the time of day and day of the week, which improves model performance. Accurate predictions of hourly demands can be used to update operational decisions, identify post-meter leaks, and improve decisions to mitigate leaks and water quality incidents. The second application develops a new model using smart meter data about pressure and flow changes to locate leaks in water distribution systems. A simulation-optimization approach couples mathematical programming with a hydraulic model to locate potential leaks using unexpected changes in real-time pressure data. Water demand data collected through advanced metering infrastructure (AMI) is used to calibrate the hydraulic model and improve the performance of the leak localization approach. The applications of using smart metered data as part of a portfolio of modeling tools provide opportunities to researchers and practitioners to improve the management of water systems in the digital era.
Bio:
Dr. Jorge E. Pesantez is a postdoctoral research scholar in Center for Infrastructure Resilience in Cities as Livable Environments (CIRCLE), a collaboration between the University of Illinois Urbana-Champaign and Zhejiang University, leading efforts in the Energy Thrust around modeling and forecasting building-scale electricity consumption. Jorge’s expertise includes data-driven analysis of demand data from water and energy systems and model-based analysis of urban water systems. His research has focused on the development of demand-response strategies for the management of infrastructure systems including forecasting short-term water demand and building simulation-optimization models to improve the management and operation of water distribution systems. His research has covered the hydraulic and quality analyses of potable and non-potable water systems. He has also applied data analytics to build models for electricity demand reported by smart meters from single- and multi-family residential users.
