Probabilistic Bridge Deterioration Prediction, Applied Analytics, and Predictive Modeling

Designing a new bridge structure or designing the rehabilitation of an existing bridge is a time-consuming process. It is important for bridge owners to proactively assess their structures to make sure that by the time rehabilitation or replacement is required, the design phase elements of such a project, including necessary permits, are ready. Assessing future bridge structure condition also helps the owner to evaluate a variety of design choices or options based on the risk imposed by each alternative. Federal regulations (MAP-21) have required state transportation agencies to implement asset management performance measures. Asset management performance strategies are arguably principal in guiding transportation infrastructure investments. Predicting future asset condition helps businesses and engineers to allocate and utilize the available resources in a direction that keeps the existing bridge inventory in the best possible condition, as well as serves the public. Prediction of bridge future condition also assists engineers and bridge owners in making decisions based upon quality information and well-defined and considered objectives [1]. The core element of any bridge management system is the database containing physical condition–rating data obtained through regular inspection and maintenance activities over a significant amount of time. Consequently, asset management relies on having an accurate inventory and condition assessment of assets in real time. One significant problem lies in the stochastic nature of traffic that transpires over time. This research focused on using the Markov process to predict future bridge condition state based on bridge inspection data inputs. Markov chain modeling was selected for this study because bridge condition ratings have a stochastic nature that requires an appropriate deterioration prediction method. The goal was to create a computational tool to predict bridge and highway degradation and assess such a structure’s long- term performance. As a proposed update to the National Bridge Inspection Standard, this study offers two methods for making changes to bridge inspection intervals. Both ways consider risk a significant contributing factor to the deciding criteria, for which this paper presents methodology to perform such a risk calculation..

Keywords:Bridge condition; Bridge risk; Condition assessment; Bridge asset; Markov chain