Profiling of Priority Sections for FWD Surveys Using Prediction Reduals From Deterioration Heterogeneity Modeling

Key Points

• The aim is to predict early deterioration areas of pavement using advanced statistical modeling techniques.
• Utilized a mixed Markov deterioration hazard model to predict pavement deterioration.
• Applied random forest analysis to evaluate the impact of various deterioration factors not included in the main model.
• Extracted high-priority locations for FWD surveys using prediction residuals from the model.
• Identified key areas where FWD surveys provide the most benefit based on deterioration predictions.
• Demonstrated the effectiveness of mixed Markov models in forecasting pavement conditions.
• Enhanced understanding of the influence of various deterioration factors through random forest analysis.

Abstract