Implication of Aggregates in the Construction and Performance of Seal Coat Pavement Overlays

The ultimate aim of this research was to formulate statistical models that can be used for predicting the frictional performance of seal coat pavement overlays. The methodology involved establishing fifty-nine seal coat test sections in many parts of the State of Texas, including all four environmental regions, and monitoring their performances over time. Many factors, believed to have an influence on performance level and identified in the literature and Texas districts surveys, were considered in this study. These included aggregate properties, construction variables, traffic, and environment and weather variables. Frictional performance was measured by a skid trailer and expressed as a friction number (FN). Eight sets of FN measurements, spanned over about five years, have been obtained which were used in the analysis. In addition, three rounds of British pendulum and sand patch testing were performed on most of the test sections. The data were used for correlation purposes with the FN and the interpretation of the trends in frictional performance. Weather data relevant to the period prior to field testing were also collected. Construction data were gathered which basically dealt with construction application rates and types of aggregates and asphalt. Aggregate samples obtained from construction sites were tested in the laboratory for their basic properties, polish susceptibility, resistance to weathering action, and resistance to abrasion and impact actions. Twenty of the samples were also examined for their mineralogical and petrographical properties. In this examination, the mineralogical constituents were estimated, and the textural characteristics were evaluated. Correlations among the laboratory tests and among the field tests were studied. The performance data was graphed to detect the sources of variations and were grouped according to the different considered variables. The grouping gave insights into which variables controlled the observed differences in frictional performance. The grouping was followed by extensive statistical modeling which pinpointed the significant variables.