← All published articles · Journal site
Muhammad Alhamdani Amin, S.Tr.T.
Journal of Engineering, Infrastructure and Geosciences (2025)
Infrastructure decisions are made about systems that have not been measured. A pipe network is renewed without a failure database, a source developed without knowing what share of production reaches a customer, an abstraction limit set without a model of the aquifer. This study argues that measurement capacity is a precondition of infrastructure investment rather than an accompaniment to it, and that the customary reform sequence inverts the dependency. The method is a structured conceptual review of research published between 2017 and 2024 across water loss and asset management, groundwater and subsidence modelling, and access measurement. Model accuracy is bounded by data drawn from operator records, so a renewal programme without a failure database cannot be prioritised whatever technique is applied. Measurement pays measurable returns: hyperparameter tuning on network data improved one failure model's mean absolute error by 20.51 per cent, and a classifier using full features achieved 88.5 per cent accuracy with an area under the curve of 0.91, while a model using asset type alone performed poorly. The instruments are now inexpensive. Governance rather than technique is the binding constraint. Measurement's absence is invisible: a system that is not measured generates no evidence of its own decline.