Smart detection and real-time learning in water distribution: an integrated data-model approach

Abstract

Water companies therefore rely on computer algorithms, such as pipe lifetime models, in order to predict pipe condition and chance of failure (Malm et al. 2012;Barton et al. 2019;Snider and McBean 2020). Hence, pipe lifetime models are an indispensable tool for realizing proactive leakage control. An estimate of remaining asset lifetime can be made based on pipe properties, pressure regime, and surroundings. Those assets that are predicted to be at the end of their lifecycle can be inspected in order to validate the model and determine actual condition of the conduits. In practice, assets subjected to equipmentbased inspection are always pre-selected using model-based techniques. The inspection results can also be used to validate and update the lifetime expectancy models.