| The objective is to advance the design of pavement surfaces. The goal is to design a pavement with low wear and high friction for tires. The objective is accomplished by relating the pavement texture to its performance and finishing technique. Pavement texture is represented by characterization parameters calculated from topographic data sets. Data sets are acquired using a Scanning Laser Position Sensor (SLPS, Johnsen 1995) capable of measuring over a 150 x 150 x 50 mm (x, y, z) volume, at a resolution of 25 {dollar}mu{dollar}m on each axis. Standard and scale-sensitive fractal analysis methods are used and a computer model is developed to calculate characterization parameters from the topographic data sets. The fractal parameters complexity (ASFC) and smooth-rough crossover (SRC) are able to differentiate certain pavement textures. The computer model, when constrained to various settings, correlates with the ASTM E965 sand-patch volumetric technique with a slope of 1.156 {dollar}pm{dollar} 0.247 and a correlation coefficient (r{dollar}sp2{dollar}) of 0.961 {dollar}pm{dollar} 0.016. Neural networks are used to relate pavement texture to in-situ performance. Nearly 75% of the neural network input arrays can predict the friction coefficient with a correlation coefficient of greater than 0.90. The parameters ASFC and SRC increase, and mean texture depth measurements made by a van or the beam method (MTDVB) decrease, this statistical significance. A new device is developed to measure 2500 coefficients of friction per second to investigate friction and wear of tires over a 60 millisecond impact on varying pavement specimens. The mean and standard deviation of the measured coefficient of friction curves differentiate the direction of the pavement brush finish. It is shown that low wear rates are possible with high coefficients of friction. |
