A comprehensive framework for evaluating the impact of rehabilitation strategies on pavement sustainability

Effective pavement maintenance is essential for ensuring road safety, minimizing lifecycle costs, and improving ride quality. One of the key challenges in pavement management is accurately evaluating the performance of rehabilitation treatments under varying environmental and structural conditions. This study proposes a comprehensive framework for evaluating the effectiveness of asphalt overlay treatments on flexible pavements, using the International Roughness Index (IRI) as the primary performance metric. Drawing on data from the Long-Term Pavement Performance (LTPP) program’s Specific Pavement Study 5 (SPS-5) sites, this study evaluates both the short-term and long-term effectiveness of pavement rehabilitation treatments. A Multiple Linear Regression (MLR) model was developed to predict immediate improvements in IRI, achieving an adjusted R² of 0.8237 based on five key explanatory variables. For long-term performance, an Artificial Neural Network (ANN) model captured complex deterioration patterns with a high prediction accuracy (R = 0.987). Two performance estimators were introduced: Rehabilitation Service Life (RSL), representing the duration until pavement reaches a critical IRI threshold, and the bounded area (AB), quantifying overall performance gain over time. The proposed evaluation indicators also serve as valuable training data for artificial intelligence models, enabling them to learn and predict pavement behavior under diverse conditions more accurately. Results highlight that the pre-overlay pavement condition, overlay thickness, and environmental factors significantly influence treatment effectiveness. Notably, thicker overlays and recycled asphalt mixtures extended service life more than thinner or virgin overlay. While the framework is conceptually applicable to flexible pavements worldwide, its current validation is limited to the SPS-5 dataset, which reflects North American climatic and traffic conditions. Broader validation with independent datasets from other regions is recommended to confirm global applicability. The proposed framework supports Pavement Management Systems (PMS) by enabling data-driven comparisons of rehabilitation strategies, aiding in more informed infrastructure planning. • MLR model predicts short-term IRI improvements with R² = 0.82 using key variables. • ANN simulates long-term pavement deterioration with high accuracy (R = 0.98). • RSL and AB indicators quantify service life and total performance gain over time. • Thicker overlays and recycled asphalt mixtures significantly extend lifespan. • Proposed framework helps PMS compare rehab strategies using LTPP SPS-5 real data.