Missouri has the seventh largest number of bridges nationwide, yet must maintain its inventory with funding from just the fourth lowest gasoline tax in the country. Estimation and prediction of the condition of bridges is necessary to create and optimize future maintenance, repair, and rehabilitation plans as well as to assign the necessary associated budgets. Previous studies have used statistical analysis, fuzzy logic, and Markovian models to develop algorithms for predicting future bridge conditions. Due to the non-linear nature of the relationship between the characteristics of bridges and their deterioration behavior, Artificial Neural Networks (ANN) have shown to be more suitable for discovering and modeling such relationship. As such, there is a gap in the literature when it comes to the ability of bride condition estimating. The goal of this research is to develop an ANN deterioration assessment model in Missouri. To this end, data on long span bridges was used where 80% of the data points were used for training and 20% were used for testing. In addition, a linear regression model was developed to act as a benchmark to assess the performed of the proposed ANN. The developed framework was successfully able to predict future condition of bridges. By using the developed model, the Missouri Department of Transportation will have a better ability to optimize their funding allocation and timing of bridge maintenance, repair, and rehabilitation. While this model was applied to bridges in Missouri, it can be tailored for other bridge assessment operations nationwide.

Meeting Name

2019 Canadian Society for Civil Engineering Annual Conference, CSCE 2019 (2019: Jun. 12-15, Montreal, Canada)


Civil, Architectural and Environmental Engineering

Keywords and Phrases

Budget Control; Deterioration; Forecasting; Fuzzy Logic; Markov Processes; Neural Networks; Regression Analysis, Artificial Neural Network Modeling; Assessment Models; Bridge Assessment; Bridge Deterioration; Bridge Maintenance; Estimation and Predictions; Linear Regression Models; Missouri Department of Transportations, Repair

Document Type

Article - Conference proceedings

Document Version

Final Version

File Type





© 2019 Canadian Society for Civil Engineering (CSCE), All rights reserved.

Publication Date

15 Jun 2019