Abstract
{ "background": "Maintenance depots are critical nodes for safety and reliability in transport systems, yet systematic evaluations of their operational methodologies are scarce. In the local context, there is a pressing need for evidence-based strategies to mitigate operational risks.", "purpose and objectives": "This study aimed to compare the efficacy of two distinct maintenance depot methodologies—a traditional scheduled maintenance system and a predictive, condition-based system—in reducing operational risk indicators within a major transport network.", "methodology": "A randomised field trial was conducted across multiple depots. Depots were randomly assigned to either the control (traditional) or intervention (predictive) group. Risk reduction was measured using a composite safety-performance index over a defined period. The primary analysis employed a linear mixed-effects model: $Y{ij} = \\beta0 + \\beta1 X{ij} + uj + \\epsilon{ij}$, where $Y{ij}$ is the risk score for depot $j$ at time $i$, $X{ij}$ denotes the intervention, and $u_j$ are depot-level random effects.", "findings": "The predictive maintenance system yielded a statistically significant reduction in the composite risk index compared to the traditional system (mean difference: -12.7 points, 95% CI: -18.3 to -7.1). Specifically, the incidence of critical failure events leading to service delays was reduced by approximately 34% in the intervention group.", "conclusion": "The condition-based predictive methodology demonstrably outperforms the traditional scheduled approach in reducing key operational risks within the studied transport maintenance environment.", "recommendations": "Transport authorities should prioritise investment in condition-monitoring technologies and data analytics capabilities to enable a shift towards predictive maintenance regimes. Further research should investigate long-term cost-benefit analyses and scalability.", "key words": "predictive maintenance, randomised controlled trial, transport infrastructure, operational risk, asset management", "contribution statement": "This paper provides the first empirical evidence from a randomised field