Study of Precision of Transverse Evenness Measurements in FILTER Experiment: Forum of European National Highway Research Laboratories Investigation on Longitudinal and Transverse Evenness of Roads Experiment

Abstract

The results of an analysis of the transverse profile measurement data collected in the FILTER Experiment are presented. The purpose of the analysis was to determine the repeatabilities, reproducibilities, and accuracies of the measurements. Application of the International Organization for Standardization 5725 standard resulted in the detection of a significant proportion of outlying results with most if not all devices. After removal of the outliers, average standard deviations ( SDs) for repeatability were typically 0.1 percent for cross fall, 0.5 to 0.9 mm for rut depth, and 0.25 mm for water depth: cross-fall measurements exhibited an average SD for reproducibility of 0.5 percent; for rut depth measurements the values were 1.7 to 2.7 mm, and for water depth measurements the values were 2.1 to 2.2 mm. The overall SD for the accuracies of the profile measurements was found to be 1.9 mm. The rather wide range of (in)accuracy (0.8 to 3.2 mm) among the different test sections is explained by the influence of longitudinal unevenness (international roughness index). Operating speed in the speed range used in the experiment had no significant influence on most measurements obtained. Moreover, speed had no significant influence on the repeatabilities or the reproducibilities of the indices or on profile measurement accuracy. Averaging distance had a significant influence on the SD for repeatability, which decreased by a factor of 2 to 4 when the averaging distance increased from 50 to 500 m. The reproducibilities of the indices did not significantly depend on the averaging distance in any systematic way.