Water Resources Research Act Program

Details for Project ID 2009OH135B

Exploring Spatial and Temporal Demand Aggregation on Transport Characteristics in Distribution System Modeling

Institute: Ohio
Year Established: 2009 Start Date: 2009-07-01 End Date: 2010-06-30
Total Federal Funds: $25,540 Total Non-Federal Funds: $51,968

Principal Investigators: Dominic Boccelli

Project Summary: Drinking water distribution system network models have traditionally been used to evaluate the ability of the network to provide adequate water quantity to the consumers. More recent applications of network models (e.g., selecting regulatory sampling locations, and protecting public health) have resulted in an increase in the use of these models for water quality analysis. In addition to performing more detailed analysis, utilities have also been moving towards the development of "all-pipe" network models (due to the proliferation of geographic information systems). Unfortunately, as the industry has begun to use these models more progressively, the assumptions that most of these models are based have not been evaluated. Typical modeling approaches assume that consumptive demands are constant over a 1-hour time frame, which was sufficient when utilities were only modeling transmission or large diameter distribution pipes. The use of "all-pipe" models have resulted in the same modeling assumptions being applied to areas with only 8 consumers. While the greater spatial detail can provide additional information, the usage of water at such a scale begins to behave in a stochastic manner that cannot be adequately represented by current modeling techniques. The objectives of this study are to explicitly evaluate the impacts of demand variability, aggregated at different temporal and spatial scales, on the underlying hydraulic and transport characteristics. The study will be performed by integrating a common distribution system network hydraulic and water quality solver (EPANET) with a computational framework capable of representing stochastic demand behavior (PRPsym). Two different size "all-pipe" network models will be utilized to evaluate the impacts of spatial and temporal aggregation of demand on the underlying hydraulic and transport characteristics. Demands will be temporally aggregated at 1-min, 10-min, and 1-hour time intervals, and spatially aggregated through the use of commercially available algorithms capable of reducing the network model size and redistributing demands. Monte Carlo simulation will be utilized to generate multiple realizations of stochastic demands and used to simulate both the hydraulics and water quality aspects associated with the two models. The analysis of the hydraulic data will be focused on exploring the impacts of the different scales of demand aggregation on flow rate variability. This information will ultimately be linked with water quality metrics to determine if there is a specific level of underlying flow variability that has a deleterious impact on water quality variability. The water quality analysis will first be focused on evaluating the impact of demand aggregation on the simulated hydraulic residence time, which can be used as a surrogate for water quality. Additional water quality simulations will be performed to assess the potential variability associated with exposure to 1) disinfectant by-products, which are ubiquitous in distribution systems, and 2) a short-duration contamination event, which would be more susceptible to hydraulic uncertainty. All of these simulations will be utilized to understand the potential impacts of various spatial and temporal aggregation scales on different distribution system network modeling objectives. These results will be compiled into a guidance document for the industry, and result in guidance for future research efforts aimed at improving distribution system network modeling.