Files
Abstract
This thesis investigates element-based bridge inspection data and proposes a method for comparing element–based bridge inspection (EBI) and National Bridge Inventory (NBI) databases. Deteriorating bridge performance is a societal problem in the U.S., and state department of transportation (DOT) agencies face limited resources to address this issue. The Federal Highway Administration (FHWA) requires that states have less than 10 percent of the total deck area that is structurally deficient (SD) when using the NBI data. A bridge is classified as SD when the condition rating in one of the 3 major components, deck, superstructure, and substructure, is 4.0 or below. However, such requirements do not exist for recently collected EBI data. This study aims to establish equivalent requirements for EBI data and understand significant predictors that influence bridge health indices (BHI). Significant predictor variables, in this case, specific elements, are identified through a regression analysis to better understand which elements affect overall BHI. These analysis parameters include element and bridge health indices, age of the structure, and the improvement cost associated with each bridge. Results confirm that deck remains as a significant predictor for determining overall BHI in Georgia, similar to how the overall score is assessed when using the NBI data. Even so, its overall contribution is reduced because there are more than 3 elements in the element data. Next, a meta-analysis is performed to test the hypothesis that a deviation in deck conditions from the two datasets exists nationwide. The results determined that there is no significant relationship between the NBI deck condition ratings and deck health indices nationwide. In a subsequent study, a meta-analysis of Georgia’s concrete bridge decks is performed to test a hypothesis that inspector’s bias affects the deck NBI and EBI correlation within the state which determined there is a linear relationship present. A previous study confirms that such a relationship exists and that there is a 22% discrepancy between NBI and BHI in Georgia. This thesis concludes with recommended threshold requirements consistent with current NBI requirements when using element-data.