Use of PMS Data for the PMED Model Calibration in Michigan

The PMED performance models have been nationally calibrated and may not represent local conditions in a region. Therefore, these models should be calibrated considering the local construction practices, performance, and climatic conditions. An essential step in the calibration process is to select pavement sections. This paper presents a systematic approach to obtaining observed performance and selecting sections for local calibration of flexible and rigid pavement models. The authors analyzed the distress data retrieved from MDOT's PMS and sensor database and successfully converted it into units compatible with the PMED. The converted data was checked to filter for anomalies and cut-off limits, ensuring the practicality of measured performance. The paper also defines specific criteria for project selection that include the need for a minimum number of data points, positive distress trends, and maintenance history. After selection, each section was categorized into good, fair, and poor-performing pavements and compared against all MDOT sections. This was to ensure that the selected sections represent overall data. The paper documents a stepwise approach to consider observed performance and select representative pavement section samples for optimal local calibration.