Evaluation Of Pavement Surface Friction Seasonal Variations

Wet-pavement friction is one of the most important pavement characteristics in relation to highway safety. This property is difficult to measure because it is affected by many vehicle, driver, pavement, and environmental parameters. In particular, it has been observed that both short- and long-term seasonal variations impact wet-pavement friction. Temperature, rainfall, and contaminants accumulated on the pavement surface affect the friction measurements.

The objective of this thesis was to quantify the effect of seasonal variations on pavement surface friction measurements on hot-mix asphalt surfaces. Monthly measurements of friction and texture were collected on nine hot-mix asphalt sections at the Virginia Smart Road for a year and a half. Friction was measured using two locked-wheel trailers and a Dynamic Friction Tester. Measurements with the two types of equipment were conducted in the same day. Macrotexture measurements were taken using a Circular Texture Meter on the same locations used for the DFTester measurements.

In order to compare friction measurements on the different surfaces, the monthly friction values were normalized by dividing the value obtained each month by the August 2007 measurements, which were theoretically the lowest friction numbers. The resulting ratios were considered friction correction factors to bring the friction measurements to the lowest value. After studying the friction variation throughout the year, sinusoidal models were fitted to the data to predict monthly correction factors for measurements at different speeds using both devices.

The main conclusion of this investigation is that seasonal variation has a significant effect on pavement friction measurements. The general trend observed is that the measurements are higher in the winter months than in the summer months. This tendency follows a cyclical sinusoidal pattern throughout the year, similar to the air temperature variations. This suggested that temperature was at least one of the factors that affected the fiction correction factors. Better coefficients of determination were obtained for the DFTester models than for those for the locked-wheel devices. However, the sinusoidal model determined for the locked-wheel device at 64 kph (40 mph), which is the standard test velocity, fit relatively well the measured friction correction factors. Average friction correction factors for the Commonwealth of Virginia were proposed using these models.

The study also showed that the friction correction factors are speed-dependent and are affected by the macrotexture of the pavement surface. The maximum (winter) friction correction factors were found to decrease with increased macrotexture for both devices at all speeds. The effect is more pronounced, however, for the locked-wheel measurements than for the DFTester measurements.