Multi-stage survival modelling for reliability assessment of asphalt pavements under fatigue and rutting

This study proposes a novel piecewise survival model to characterize the phased deterioration of asphalt pavements subjected to fatigue and rutting. Traditional survival models often fall short in representing the distinct failure pattern observed in real-world infrastructure. To address this, a three-phase survival framework is developed, capturing early-life decay, service life and aging deterioration. Each phase was modelled by interpretable parametric forms with smooth transitions. Compared to the classical Weibull model, the proposed structure offers enhanced flexibility and better physical alignment with pavement lifecycle dynamics. Sensitivity analysis highlighted how key parameters influenced different life stages, while reliable life estimates provided actionable thresholds for maintenance planning. The model’s robustness was further validated through Monte Carlo-based uncertainty quantification yielding 95% confidence bands for survival predictions. Results showed that the model maintains stability under moderate parameter variations and effectively supports risk-informed decision-making. This work offers a practical, analytically consistent tool for pavement infrastructure reliability analysis and can be adapted for broader applications in lifecycle modelling and asset management.