Performance Metrics for Leak Detection and Pressure Control in Urban Water Networks: Assessing Latency, Drift, and Response Efficacy

Abstract

This article presents an engineering-oriented framework that treats leak detection and pressure management as an end-to-end decision system, modeling uncertainty propagation from instrumentation through inference and threshold governance into outcomes that utilities actually manage, including probability of missed leaks, false alarm rate, time-to-detection and time-to-intervention distributions, expected volume loss before containment, and probability of customer pressure noncompliance under mitigation actions. A scenario-based quantitative study is developed for a generic but realistic distribution network with district metered areas and pressure zones, using flow and pressure telemetry augmented by periodic acoustic surveys, and four operational architectures are compared: baseline fixed-threshold minimum night flow, increased sensing without governance, model-based residual detection with limited drift management, and a governance-optimized two-tier architecture that combines nuisance-constrained thresholds, drift-aware plausibility checks, adaptive confirmation sampling, and staged pressure interventions aligned with evidence strength. Results show that (i) leakage loss and customer impact are dominated by tail behavior in detection and verification latency rather than by mean leak rate, (ii) sensor drift and baseline instability drive false stability that delays detection when fixed thresholds are used, and (iii) reliability improves most when governance reduces long-tail verification times and when pressure management is staged to reduce loss without causing pressure violations. The paper provides copy-ready tables and complete prompts for data-driven figures suitable for Techne submission and for adaptation to utility telemetry datasets.