Spatial and semantic enrichment of utility networks data

Abstract

Utility networks are critical components of urban infrastructure, providing essential services such as water supply, electricity, gas, and telecommunications. The traditional method for mapping these networks is typically two-dimensional (2D) schematic representations rather than topographically and geometrically correct maps. These representations lack the capacity to convey the complexity and vertical intricacies of urban infrastructures. This limitation hampers comprehensive planning, efficient management, and risk mitigation during construction activities because 2D maps do not effectively represent the multi-layered and interconnected nature of urban utilities, leading to potential oversights and inaccuracies. This thesis addresses the challenge of enriching utility network data by integrating detailed data from utility trench surveys. These surveys provide precise positional and attribute information about utilities that are often missing in standard maps, such as the exact depth, spatial configuration, and physical characteristics of the utility lines.

Data from utility trenches in three Dutch cities—Enschede, Rotterdam, and Amsterdam—was acquired and analyzed. Methodologies were developed to extract, standardize, and integrate this data into existing utility network maps, enhancing their semantic content and spatial accuracy. The research demonstrated that integrating trench data can reveal inaccuracies in traditional utility network maps at the utility trench locations.

Key findings include the development of algorithms for extracting and processing utility trench data, the identification of common challenges between cities such as cable/pipeline labeling inconsistencies, and the comparison of enriched utility data and that of traditional utility networks. The research also highlights the importance of standardizing data models and the potential of three-dimensional models to provide a more comprehensive understanding of utility networks. A resulting recommendation was to improve data collection by including all information found and providing properly geo-referenced data.