Incorporating the Effectiveness of Preservation and Rehabilitation Techniques on Flexible Pavement Service Life Predictions Using Machine Learning Approach

A flexible pavement network may be exposed to traffic changes over time because of detours that may cause unusually high traffic loading on some parts of the pavement sections. Such sections, therefore, need to be preserved earlier than their expected service life. This study focuses on incorporating the effectiveness of preservation and rehabilitation techniques, e.g., functional thin overlay and structural asphalt overlay on flexible pavement service life predictions using a machine learning approach. The developed models are capable of estimating the current and future international roughness index (IRI) and of detecting performance recovery after application of preservation treatment, thereby forecasting the remaining service life. The traffic, pavement structural, and performance data were obtained from the pavement management information system (PMIS) of the Iowa Department of Transportation (DOT). This study describes realistic what-if scenarios considering various treatment techniques to provide a better understanding of the effectiveness of treatments on extending flexible pavement service life. Such outcomes will help facilitate better pavement preservation management strategies.