Estimating the Load Rating of Reinforced Concrete Slab Bridges Without Plans

Abstract

This article documents the development of a multiple regression model for the estimation of the inventory rating of reinforced concrete slab bridges. Given the lack of information on the superstructure elements in the National Bridge Inventory, the authors developed a data set by using the information found in as-built drawings provided by multiple state departments of transportation. The data consisted of 50 reinforced concrete slabs for which their inventory rating was computed. Various model combinations were considered, and their estimated errors were evaluated by performing a repeated tenfold cross validation. A five-variable model that included the year built, the span length, the slab thickness, the support type, and the skew was selected based on the lowest mean error and variance. The model had an adjusted coefficient of determination (adjusted R²) of 0.817, which was considerably higher than a previously developed model. The difference in adjusted R² is attributed to the inclusion of slab thickness as an independent variable as well as to the quality of the data used.