Date of Award

Summer 2011

Document Type


Degree Name

Doctor of Philosophy (PhD)


Materials and Infrastructure Systems

First Advisor

Erez N. Allouche


A culvert is typically a hydraulic passage, normally placed perpendicular to the road alignment, which connects the upstream and downstream sections underneath an embankment, while also providing structural support for earth and traffic loads. The structural condition of culverts continues to deteriorate due to aging, limited maintenance budgets, and increased traffic loads. Maintaining the performance of culverts at acceptable levels is a priority for the U.S. Department of Transportation (DOT), and an effective maintenance of culvert structures can be greatly improved by introducing asset management practices. A priority list generated by traditional condition assessment might not provide optimum solutions, and benefits of culvert asset management practices can be maximized by incorporating prediction of deterioration trends. This dissertation includes the development of a decision making chart for culvert inspection, the development of a culvert rating methodology using the Analytic Hierarchy Process (AFIP) based on an expert opinion survey and the development of a Markovian model to predict the deterioration rate of culvert structures at the network level.

The literature review is presented in three parts: culvert asset management systems in the U.S.; Non-destructive Technologies (NDT) for culvert inspection (concrete, metal, and thermoplastic culvert structures); and statistical approaches for estimating the deterioration rate for infrastructure. A review of available NDT methods was performed to identify methods applicable for culvert inspection.

To identify practices currently used for culvert asset management, culvert inventory data requests were sent to 34 DOTs. The responses revealed that a relatively small number of DOTs manage their culvert assets using formal asset management systems and, while a number of DOTs have inventory databases, many do not have a methodology in place to convert them to priority lists. In addition, when making decisions, DOTs do not incorporate future deterioration rate information into the decision making process. The objective of this work was to narrow the gap between research and application.

The culvert inventory database provides basic information support for culvert asset management. Preliminary data analysis of datasets provided by selected DOTs was performed to demonstrate the differences among them. An expert opinion survey using AHP was performed to confirm the weight of 23 factors, which was believed to contribute to the hydraulic & structural performance of culvert structures, so as to establish the culvert rating methodology.

A homogenous Markov model, which was calibrated using the Metropolis-Hastings Algorithm, was utilized in the computation of the deterioration rate of culverts at the network level. A real world case study consisting of datasets of three highways inspected regularly by Oregon DOT is also presented. The performance of the model was validated using Pearson's chi-square test.