This paper addresses a new procedure for spatial and statistical analysis of water mains failure. We use this methodology to analyze and predict the spatial behavior of failure on the water mains network. The analysis is based on data from Sanandaj city in Iran over a 10 year period. It includes 395 mains failure which they were geocoded as discrete points. Prediction maps are constructed using different spatial interpolation methods, namely; the Thiessen polygons, the Density estimation, and the Triangulated Irregular Network model (TIN). Thiessen polygons are used for proximity and area allocation analysis of failure points. Density of failure points is calculated for each cell by summing the number of failures in the cell. TIN model generates a surface form interpolation of failure density. The study indicates that the construction of continuous surface from point dataset offers an interesting tool for the prediction of water main failures. It also provides comprehensive spatial information system that can be easily used by utility managers as a decision-support tool to set up and compare strategies for the renewal of water mains.

This content is only available via PDF.
You do not currently have access to this content.