Peak Friction Prediction Model Based on Surface Texture Characteristics

This paper proposes a new model for predicting the speed gradient of peak friction values on asphalt pavements on the basis of surface characteristics. The innovative feature of the proposed model is the reliable estimation of peak friction values experienced by vehicles equipped with an antilock brake system at a certain vehicle speed. To define the experimental model, several types of dense asphalt concrete surface layers with various surface characteristics were analyzed by in situ tests. Friction was measured with the Skiddometer BV11 and the British pendulum tester, and texture properties were measured with a laser profilometer. The Rado model was used to predict peak friction values at three vehicle speeds, and these data were used to determine the gradient of peak friction values for each pavement section. The spectral analysis of pavement profile data was used to define a texture parameter negatively correlated with peak friction values; this parameter was introduced in a new formulation of the speed number S_p* that was a measure of the influence of pavement macrotexture on peak friction values. The speed number S_p* was used in the new exponential model proposed for defining the gradient of peak friction values. The results show that the model is highly reliable; because the model allows identification of texture characteristics to be modified to optimize peak friction values, it is particularly useful for optimization of the mix design and maintenance of pavement surfaces.