The Data Mining Approach for Analyzing Infrastructure Operating Conditions

AbstractBridges serve as critical structures to the functionality of road networks within an infrastructure system and require regular inspections to ensure they are in safe operating condition for the public. This research developed a data mining approach to analyze and produce predictions of the future operating condition of bridges. 5-year bridge inspection data provided by the Connecticut Department of Transportation (ConnDOT) are applied as case studies. The approach combines the artificial neural networks with non-linear statistical data modelling tools to study the complex relationship between various bridge parameters and bridge conditions. Some variables considered include information on geometry, construction and service, while the neural net output variables were the condition ratings for the deck, superstructure, and substructure. The data mining model demonstrated the potential to develop a practical computer tool to accurately estimate the future condition ratings of the bridges. The results were discussed with the bridge experts in ConnDOT and suggestions on the model's practical applications were also made