Abstract
{ "background": "Transport maintenance depots are critical infrastructure for road freight and public transport networks. Current reliability assessments often rely on deterministic models or aggregated failure data, which fail to account for operational heterogeneity and uncertainty inherent in complex, multi-component systems.", "purpose and objectives": "This study aims to develop and validate a novel probabilistic framework for quantifying the reliability of maintenance depot systems, explicitly modelling variability between individual depots and integrating multiple sources of uncertainty.", "methodology": "A Bayesian hierarchical model was formulated, treating depot-specific reliability parameters as drawn from a common population distribution. The core reliability model for a depot $d$ is $Rd(t) = \\exp(-(\\lambdad t)^{\\kappad})$, where $\\lambdad$ and $\\kappad$ are modelled hierarchically. Operational data, including time-failure for key subsystems, were collected from a sample of depots. Posterior distributions were estimated using Markov chain Monte Carlo sampling.", "findings": "The model identified significant heterogeneity in failure rates ($\\lambdad$) across depots, with a 95% credible interval for the population mean spanning [0.021, 0.034] failures per day. Depots in coastal regions showed, on average, a 15% higher reliability over a 30-day period compared to inland counterparts, after controlling for asset age and workload.", "conclusion": "The hierarchical Bayesian approach provides a robust and nuanced tool for system reliability analysis, effectively separating common trends from depot-specific idiosyncrasies and formally incorporating parameter uncertainty.", "recommendations": "Infrastructure managers should adopt probabilistic reliability models to inform targeted maintenance investments. Future data collection should standardise the recording of contextual factors, such as environmental conditions and maintenance histories, to enhance model granularity.", "key words": "Bayesian inference, infrastructure management, maintenance engineering, probabilistic modelling, system reliability, Weibull distribution", "contribution statement": "This paper presents the first application of a Bayesian hierarchical model to the reliability assessment of