Quantifying Pavement Preservation Performance Using Probabilistic Deterioration Modeling
Type of DegreePhD Dissertation
MetadataShow full item record
Throughout the last few decades, the pavement industry has slowly transitioned from construction and rehabilitation to pavement preservation. Pavement preservation can be described as a proactive approach for protecting and maintaining existing pavements. Preservation techniques are and will remain indispensable to an industry where sustainability has been a major focus. Today, there are numerous preservation techniques available. These techniques are regularly being refined and expanded, as owners recognize that pavement preservation is a cost-effective approach. The prospects of conducting exclusive research on pavement preservation led the National Center for Asphalt Technology (NCAT) to build innovative preservation test sections to quantify their life-extending benefits. In the summer of 2012, pavement preservation treatments were placed on a low-volume county roadway in Auburn, Alabama that is located close to NCAT’s main facility. The sections are continuously being monitored since then for cracking, rutting, rideability, macrotexture, and skid resistance. Prior to treatment application, the roadway structural condition was assessed to ensure the pavement was in good condition. Moreover, structural integrity was checked through the length of the roadway to ensure that treatments were placed on a pavement with a uniform structural condition, thus, engendering fair performance assessments for different treatments. The half-mile roadway was laid with 23 treatments and 2 sections left untreated as a baseline condition which reflects a ‘do-nothing’ strategy. The pairwise comparisons of treatment performances with the untreated sections help quantify the performance, or more precisely, calculate the benefits that are derived from preservation applications. Among all the performance measures collected, only cracking showed an increasing trend, and hence, it was used as a single performance index to developed models. However, a similar methodology described herein can be applied on other performance indices in the future. Two probabilistic models of Markov chain and Survival analysis were used to model deterioration trends. The Markov chain model is non-parametric. As such, treatments were first categorized based on a hierarchical agglomerative clustering analysis to develop models for each cluster of treatments. Survival analysis was used in the second approach. Non-parametric, semi- and full-parametric survival models were applied on the dataset to quantify treatment performances. Results confirmed that preservation strategies significantly increase the potential for extended service life when compared with a “do-nothing” scenario. It was found that pretreatment condition, treatment family, recycled material usage, and crack sealing application have a significant impact on future deteriorations. In addition, life-extending benefits derived using different methodologies were documented for future reference. It is highly important to note that although results obtained in this study provide a basis for comparisons on various treatments, they should not be interpreted for other climatic regions, traffic levels, pavement types and so forth. The present study is part of a larger study encompassing low- and high-volume roads in hot and cold climates. The present test site was the oldest of these test locations and had enough data for analysis. Data for other test locations are still in their infancy; more years of data collection are needed before comprehensive models can be offered for use in a functioning Pavement Management System (PMS) across the US.