Graduation Semester and Year

Summer 2025

Language

English

Document Type

Dissertation

Degree Name

Doctor of Philosophy in Civil Engineering

Department

Civil Engineering

First Advisor

Nur Yazdani

Second Advisor

Kate Hyun

Third Advisor

Suyun Paul Ham

Fourth Advisor

Mohammad Atiqul Islam

Abstract

The deterioration of bridge elements arises from multiple complex factors such as reinforcement corrosion, concrete degradation, creep, shrinkage, cracking, and fatigue. In the absence of mechanistic models that quantitatively account for these factors, along with environmental influences and maintenance constraints, transportation agencies rely on inspection data to assess rehabilitation, or replacement needs and prioritize maintenance activities. This Ph.D. research addresses these challenges by developing predictive models to forecast bridge deck condition ratings and load ratings using machine learning algorithms. The study leverages historical data from the National Bridge Inventory (NBI) and environmental datasets, incorporating climatic zone classifications from NOAA. Supervised machine learning models, including decision tree, k-nearest neighbors, support vector machines, random forest, Adaboost, CatBoost, and Bagged Tree, were applied to predict future deck condition ratings. Additional data from the Long-Term Bridge Performance (LTBP) program were combined with the bridge features to capture both structural and environmental impacts on deterioration. Since bridge records contain both temporal and spatial data, deep learning approaches such as vanilla and stacked Long Short-Term Memory (LSTM) networks (vLSTM, sLSTM) were employed to capture temporal deterioration trends. Additionally, a Random Forest model was developed to estimate bridge load ratings, integrating diagnostic load test data to improve accuracy beyond inspection-based methods. Results demonstrated high predictive performance across models, highlighting critical features influencing deterioration and enabling data-driven decision-making for bridge safety and maintenance prioritization.

Keywords

Predictive model, Machine learning, Artificial intelligence (AI), Deep learning, Neural networks, Load Rating

Disciplines

Structural Engineering

License

Creative Commons Attribution 4.0 International License
This work is licensed under a Creative Commons Attribution 4.0 International License.

Share

COinS
 
 

To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.