Graduation Semester and Year
Summer 2026
Language
English
Document Type
Dissertation
Degree Name
Doctor of Philosophy in Civil Engineering
Department
Civil Engineering
First Advisor
Mohammad Najafi
Second Advisor
Vinayak Kaushal
Abstract
Municipal sewer infrastructure across the United States faces challenges due to aging systems, while comprehensive condition assessment remains costly and time-intensive for many jurisdictions. Traditional inspection methods, such as closed-circuit television (CCTV), are expensive and slow, limiting the systematic evaluation of large municipal sewer networks. This dissertation develops a machine-learning framework using transfer learning for the condition assessment of sanitary sewer pipes. The proposed framework enables knowledge transfer between municipalities with different data availability, allowing condition estimation in target systems lacking inspection-based PACP ratings. A baseline municipal dataset containing PACP-rated sewer condition data and a target dataset from the City of Los Angeles were used to develop and evaluate the framework. The methodology integrates inspection-priority screening, baseline condition modeling, transfer-learning-based condition-rating generation, and validation using multiple machine learning algorithms, including Random Forest, XGBoost, LightGBM, AdaBoost, Artificial Neural Networks, and stacking hybrid meta-models. Hyperparameter tuning, cross-validation, and comparative performance evaluation were performed. The results demonstrate model performance, with 92.3% accuracy for condition assessment. Feature-importance analysis identified pipe age, length, and material as the most influential factors affecting sewer pipe deterioration. The findings indicate that core relationships between infrastructure attributes and condition patterns can be transferred across jurisdictions to support data-limited sewer systems. The developed framework provides a scalable and cost-effective approach for municipal sewer condition assessment, supporting improved maintenance planning and inspection prioritization through strategic transfer learning.
Keywords
Machine learning, Condition assessment.
Disciplines
Civil and Environmental Engineering | Computer Sciences
License
This work is licensed under a Creative Commons Attribution 4.0 International License.
Recommended Citation
Mohammadagha, Mohsen, "DEVELOPMENT OF A MACHINE LEARNING MODEL USING TRANSFER LEARNING FOR CONDITION ASSESSMENT OF SANITARY SEWER PIPES" (2026). Civil Engineering Dissertations. 7.
https://mavmatrix.uta.edu/civilengineering_dissertations2/7