Network-aware deterioration risk and renewal prioritization for aging urban sewers: a gis-based simulation framework applied to Nam-gu, Ulsan, South Korea

Municipalities must decide which of their ageing sewers to renew first, yet many hold no inspection record at all. This paper asks how far renewal prioritization can be taken using only the attributes a utility already holds in its sewer geographic information system (GIS), and proposes a network-aware simulation framework to answer it: a material-specific Weibull model converts pipe age and its environmental stressors into a failure likelihood; a Manning-based hydraulic model on a geometrically rebuilt network graph converts pipe position into a failure consequence; their product defines pipe risk; an interpretable gradient-boosted surrogate emulates the deterioration engine; and a multi-objective genetic algorithm traces the trade-off between renewal spending and residual network risk. The framework is demonstrated on the complete public sewer network of Nam-gu, Ulsan, South Korea — 19,849 gravity pipes over 886.5 km, 68% of them older than thirty years, and without any inspection record. Its principal methodological contribution is a decision-relevant accuracy benchmark: although the surrogate reproduces only 57% of the variance of the simulated condition score, prioritizing on predicted rather than engine-computed risk retains 98.6–99.7% of the achievable risk reduction, because renewal decisions consume a rank order rather than a point estimate. Deterioration likelihood and failure consequence were approximately uncorrelated in rank across this network ( ρ = −0.03), so network position, and not age alone, sets the priority. All results are internal to the simulation: without inspection labels the surrogate emulates a mathematical process rather than measured defects, and its accuracy is not evidence of field-diagnostic skill. The framework is offered as a reproducible screening procedure for utilities holding a sewer GIS and no inspection data, designed to be recalibrated as inspection records arrive.