Understanding the leakage process for multi-scale water infrastructure asset management: necessity for a dialogue between sociological and data sciences
Abstract
Reducing water losses is one of the most pressing issues for modern water utilities. To that end, improving the efficiency of the pipe leakage and repair process and aiding the selection of the pipes that are to be renewed or rehabilitated are essential. To help addressing these tasks, in this work, we develop a model predicting the probability of a pipe to be leaking. This work is set the context of a multidisciplinary project with Soci{\'e}t{\'e} Wallone des Eaux and it is aligned with their goal to improve their Infrastructure Asset Management in the short and the long terms. Developing and feeding this leakage probability model relies on an intense data processing phase, mobilizing data and water engineering sciences, since the raw data from SWDE is not ready to be used in the model. Complementarily, we thus employ techniques from sociology (e.g., interviews, analyses of the human/non-human actors and of the tools, sociotechnical translations) in order to complete the data, to improve our understanding of its production, and to increase its value and its availability for the prediction of the pipe leakage probability. This model will be implemented in SWDE's information system and used for strategies to reduce water losses.
Keywords
Cite
@article{arxiv.2311.17777,
title = {Understanding the leakage process for multi-scale water infrastructure asset management: necessity for a dialogue between sociological and data sciences},
author = {Marie Collet and Nicolas Rodriguez and Selma Baati and Alain Husson and Eddy Renaud and Kevin Caillaud and Yves Le Gat},
journal= {arXiv preprint arXiv:2311.17777},
year = {2023}
}