Utilizing GIS in Developing Realistic Demand Distributions to Support Modeling in Water Supply Master Planning

摘要

Computer models that predict the hydraulic performance of a distribution system are powerful tools that can be used to help utilities make decisions about future improvements to their water systems. In 1999, the City of Elkhart initiated a project to take the existing model of their distribution network and develop a planning-level model that would enable capital improvements planning. Developing a fully calibrated network model is a multi-step process; this process is described in the flow diagram presented on Figure 1. In this project, the City of Elkhart proceeded with development of a planning-level model; the work necessary to achieve a fully calibrated model has not been completed and is planned as a future project. The purpose of this paper is to present an innovative approach to allocating water demands within the model network. This demand allocation process, while imperfect, was successful enough to allow Elkhart to proceed with planning-level decisions prior to developing a fully calibrated model. Traditional approaches to projecting water consumption demands focus on population projections in conjunction with land use and zoning information derived from census bureau data, local housing authorities and municipal and regional planning agencies. Unfortunately, such data often requires subjective interpretation due to inconsistencies in the data, over-consolidation of the data, or the age of the data (i.e., outdated). In support of the hydraulic model used as a water master planning level tool only, the City of Elkhart has developed its water demand distributions using demand data derived from actual billing records over the past several years. Incorporating a state-of-the-art Geographical Information System (GIS) geocoding feature, actual usage rates corresponding to specific customer addresses were assigned to specific point locations throughout the distribution system at corresponding address points in Elkhart's GIS. Demands were then assigned to specific WaterCAD~(~R) model nodes through logical grouping assignments using GIS polygons. The result provided existing realistic demand distributions for rapid and reasonable model development for capital planning purposes.