Virginia Bridge Information Systems Laboratory

2014PDFTech ReportCAIT-UTC-005DTRT12-G-UTC16BridgesBridge management systemsData miningData analysisDeteriorationMarkov chainsRegression analysisService lifeResearch HubBridgesBridge management systemsData analysisVirginiaUnited StatesRutgers University. Center for Advanced Infrastructure and TransportationChase, Steven B.University of Virginia. Dept. of Civil and Environmental EngineeringUnited States. Federal Highway AdministrationUnited States. Dept. of Transportation. Research and Innovative Technology AdministrationVirginia. Dept. of TransportationUS Transportation CollectionThis report presents the results of applied data mining of legacy bridge databases, focusing on the Pontis and National Bridge Inventory databases maintained by the Virginia Department of Transportation (VDOT). Data analysis was performed using a variety of information technology tools and statistical methods including Microsoft Access and Excel and the R Statistics System. The resulting information consists of models which were of interest to the Virginia Department of Transportation. Deterioration models were developed for Interstate bridges in Virginia. Two sub-studies were conducted in response to VDOT interests. First, Markov Chain models were developed for condition states for the most common Pontis bridge elements on the Interstate bridges. Second, regression models for condition ratings were developed for these same elements. Two additional special studies were conducted by the Virginia Bridge Information Systems Laboratory. A special study at the National Scale was performed, examining 20 years of NBI data. This study summarized typical changes in bridge performance metrics by identifying a sample of bridges with temporally contiguous data for the period from 1992 to 2012. This study uncovered the signi\ufb01cance of maintenance and repair actions on bridge performance. Another special study examined the characteristics of bridges which were taken out of service in this same 20 year period. Statistical summaries of service life data were developed