A statistical approach for predicting skid resistance of asphalt pavements

There are various models for predicting the friction or skid number of asphalt pavements based on inputs such as aggregate characteristics, aggregate gradation, and traffic level. However, existing models are deterministic as they incorporate mean values of the input parameters and yield an average skid number value for a pavement section. Consequently, these models are not capable of accounting for variability in material properties and associated variation in the predicted skid number values. This study aims to develop a probabilistic approach that determines the mean as well as the variation in skid number in a pavement section. The efficacy of this approach is demonstrated by selecting one of the validated models that incorporate a wide range of aggregate characteristics. The results showed that while the average skid number predicted by the deterministic model may satisfy a required threshold/specification value, a fair percentage of the distribution could fall below the required value. The probabilistic approach is useful for engineers to determine the effect of aggregate characteristics on skid resistance distribution and to identify risks associated with the probability of skid resistance values in a pavement section fall below the required limit.