Pavement Condition Rating Based on Factor Analysis of Pavement Condition Measurements (05-2307) - MP-06
Antti Ruotoistenmaki, Helsinki School of Economics, Finland

The vast amount of information conveyed by a large number of measured pavement condition variables has to be summarized for use at the strategic-level decision-making of pavement maintenance. For this purpose, several condition indices exist, the most famous of them being the Pavement Condition Index (PCI) and the Present Serviceability Index (PSI). In this paper, a pavement condition rating is derived using factor analysis of measured pavement condition variables. This condition rating can be calibrated to derive any existing or new condition index values from pavement condition measurements. Factor analysis summarizes the information contained in the original variables into few distinct dimensions - factors - that describe the more general, usually not directly measurable properties of the pavement. In this paper, first a preliminary analysis is done to find three factors to be extracted from the data set, one by one, in the second stage of the analysis: the structural factor, the roughness factor and the transversal unevenness factor. The data set consists of 14,416 observations of 100-meter (328 ft.) pavement sections with both pavement surface profile and deflections measured in 2003. Logarithmic transformation is applied to the original variables to facilitate the calculation of normally distributed factor scores as the mean of the standardized values of variables belonging to each factor. The condition rating is calculated as the weighted sum of the factor scores. Condition rating values from different pavement sections with varying condition variables are directly comparable.