Flexible Pavement Distresses Prediction Models using AASHTOWare

Pavement performance prediction is widely considered as a significant element of road infrastructure asset-management systems or Pavement Management Systems (PMS) by pavement researchers and practitioners. Predicting pavement performance significantly reduces the huge costs of constructing roads, especially in the case of countries that made incredible investments in road construction. This study mainly focuses on the implementation of the mechanistic-empirical (M-E) analysis method using the AASHTOWare Pavement ME Design (AASHTOWare PMED) software for flexible pavement distress prediction-models generation. To achieve that four steps were followed. First, the most accurate assessment that shows the combined impact of the most important parameters that affect flexible pavement performance was used to perform the AASHTOWare runs. In which, 378 design combinations of (3 traffic speed levels × 3 traffic load levels ×3 climatic zones ×7 Surface HMA mixes widely used in Egypt) at two input levels of the AASHTOWare PMED hierarchy (levels 1 &2) that typically are required for binders and hot-mix-asphalt (HMA) were used. Second, a sensitivity analysis to study the combined effect and impact of the investigated parameters on AASHTOWare PMED-predicted performance (cracking, rutting, and roughness) was conducted at the two input levels. Third, a Multiple Linear Regression (MLR) was implemented as a modeling approach to develop five performance prediction models for flexible pavements based on the AASHTOWare PMED software results. The proposed MLR models predicted each distress as a function of climatic factors, the surface HMA properties, different regions' speed levels, and traffic volume levels. Finally, a validation process of the proposed MLR prediction models was conducted. Results indicated that the proposed models yield an overall good prediction, asserting the robustness of the proposed process. Proposed MLR prediction models can be perceived as a function of Average Annual Daily Truck Traffic, Traffic speed, mean annual air temperature, and the percentage of air voids. This study provides a procedure to develop the performance prediction models of flexible pavements based on the AASHTOWare PMED approach and in accordance with different regions’ input levels.