LTPP PROFILE VARIABILITY

A study has been conducted to evaluate the quality and variability of the International Roughness Index (IRI) data in the Long Term Pavement Performance (LTPP) database. All LTPP profiles collected between June 1989 and October 1997 were visually reviewed for saturation spikes, lost lock, shifted starts, wrong location, out of study, and other equipment- and operator-related problems. Noted problems are identified in the report. Repair, replacement, or deletion of these files was completed in conjunction with this study. Remaining or replaced good IRI data were used for analysis of variability. Analysis of variance focused on the run-to-run and visit-to-visit IRI variability in the LTPP sections studied. Mean IRI values for these asphalt concrete (AC), portland cement concrete (PC), and AC/PC overlaid sections ranged from 0.5 to 4.3 m/km. Based on the upgraded data set, confidence limits were developed for expected variability between repeated profile testing runs and for the expected change in IRI between subsequent visits. Additional analysis is also presented regarding the effects of saturation spikes, shifted profiles, and wheelpath testing location on IRI. Study was also completed on the effect of the number of runs collected at a site, the effect of moist pavement surfaces, and the results of small changes in the coefficients used to compute IRI. Recommendations of the study address the methods used for field profile collection, office review methods, Information Management System (IMS) database design, and profile calibration.