The reliability of transport maintenance depot systems is critical for national infrastructure, yet quantitative, predictive assessments of their operational performance are lacking. Current evaluations often rely on static, retrospective analyses, which are inadequate for proactive management and resource allocation. This working paper develops and evaluates a novel time-series forecasting model to assess the reliability of transport maintenance depot systems. The objective is to provide a dynamic, data-driven tool for predicting system failures and scheduling preventive interventions. A seasonal autoregressive integrated moving average (SARIMA) model, specified as (B) (Bˢ) ᵈₛD yₜ = (B) (Bˢ) ₜ, is applied to historical failure data from multiple depots. Model parameters are estimated using maximum likelihood, and forecast uncertainty is quantified with 95% prediction intervals. The model demonstrates strong predictive capability, with a mean absolute percentage error (MAPE) of 12. 7% for a one-month-ahead forecast. A key finding is a pronounced seasonal pattern in system failures, with reliability decreasing by approximately 18% during peak operational periods. The proposed SARIMA model provides a robust methodological framework for forecasting depot system reliability, moving beyond descriptive analysis to enable evidence-based maintenance planning. Implement the forecasting model as a decision-support tool within depot management systems. Future work should integrate exogenous variables, such as parts inventory levels, to improve model specificity. Infrastructure reliability, maintenance management, predictive modelling, SARIMA, systems engineering This paper introduces a novel application of time-series forecasting to quantitatively predict the reliability of transport maintenance depots, providing a new tool for proactive infrastructure management.
Nkosi et al. (2007) studied this question.