8th International Conference on Managing Pavement Assets LONG-TERM PAVEMENT PERFORMANCE EFFECTIVENESS OF FLEXIBLE PAVEMENT REHABILITATION TREATMENTS USING MARKOV CHAIN ALGORITHM

In the long-term pavement performance (LTPP) study, the SPS-5 experiment was designed to assess the performance of different flexible pavement rehabilitation treatments, relative to the performance of untreated control sections. The experiment consists of a control section and two-inch and five-inch overlays, with virgin or recycled asphalt concrete mixes, and with or without pre-overlay milling. There is a need to re-evaluate the results of the field experiment as more performance data become available. This paper treats the SPS-5 sections as a network and uses Markov chain algorithm (MCA) to evaluate the effectiveness of rehabilitation treatments at the network level. For each treatment, the transition matrices were determined from the observed time series performance data for ride quality, fatigue cracking, and rutting. The advantages of using MCA includes the ability to dynamically model pavement deterioration and improvement at the same time, evaluate the impact of initial pavement conditions on the short- and long-term performance, and compare the relative pavement performance of different maintenance or rehabilitation treatments. The results show that different rehabilitation treatments have varying degrees of effectiveness depending on the distress type. For example, thick overlays are more effective in the long-term for improving IRI and cracking while thin overlays seem to be a better choice in case of rutting. Generally, pavement sections with recycled asphalt concrete showed higher potential for cracking when compared with pavements with virgin asphalt concrete mixes. Pre-overlay repairs were found to have noticeable impact on the effectiveness of the thin overlays, especially for cracking. KEY WORDS: long-term pavement performance, rehabilitation treatments, marcov chai