A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Developing Probabilistic Condition Assessment Models of Concrete Bridges Utilizing Text Mining
https://orcid.org/http://orcid.org/0000-0003-4696-1678
Bridges are one of the most critical systems of transportation infrastructure. A bridge’s structural safety and serviceability can be negatively impacted by serious deterioration caused by harsh environmental conditions, increased traffic volumes, aging, and deferred maintenance. Transportation agencies conduct comprehensive inspections and assessments to provide a complete picture of a bridge’s present condition and determine accordingly required interventions. The collected inspection data include essential information pertinent to the different types of defects in each bridge component. However, such inspection data are stored in unstructured textual reports without effectively being used. In addition, current condition assessment and deterioration models consider only quantitative data and overlook such descriptive text-type data commonly included in inspection reports. This paper introduces a text-mining-based approach to standardize and automate the condition assessment and deterioration modeling of reinforced concrete bridge decks using Bayesian belief networks. Bayesian belief networks are efficient tools for modeling and visualizing complex system interdependencies in a probabilistic manner by means of marginal and conditional probabilities established on a specific model structure. First, text mining techniques are employed to analyze inspection comments related to bridge decks in order to get insights into these comments. Then, text mining results are used to build Bayesian belief networks by calculating conditional and marginal probabilities for different network nodes. Bridge deck conditions and different defect types, including corrosion, cracking, delamination, spalling, and scaling, are modeled. The model relies on prior knowledge and what-if analysis to make probabilistic inferences about the system. The practical contribution of the proposed approach lies in extracting the necessary information buried in inspection reports and utilizing such information in building probabilistic graphical models for bridge conditions and deterioration. These models can be further used in making informed decisions regarding bridge intervention strategies.
Developing Probabilistic Condition Assessment Models of Concrete Bridges Utilizing Text Mining
https://orcid.org/http://orcid.org/0000-0003-4696-1678
Bridges are one of the most critical systems of transportation infrastructure. A bridge’s structural safety and serviceability can be negatively impacted by serious deterioration caused by harsh environmental conditions, increased traffic volumes, aging, and deferred maintenance. Transportation agencies conduct comprehensive inspections and assessments to provide a complete picture of a bridge’s present condition and determine accordingly required interventions. The collected inspection data include essential information pertinent to the different types of defects in each bridge component. However, such inspection data are stored in unstructured textual reports without effectively being used. In addition, current condition assessment and deterioration models consider only quantitative data and overlook such descriptive text-type data commonly included in inspection reports. This paper introduces a text-mining-based approach to standardize and automate the condition assessment and deterioration modeling of reinforced concrete bridge decks using Bayesian belief networks. Bayesian belief networks are efficient tools for modeling and visualizing complex system interdependencies in a probabilistic manner by means of marginal and conditional probabilities established on a specific model structure. First, text mining techniques are employed to analyze inspection comments related to bridge decks in order to get insights into these comments. Then, text mining results are used to build Bayesian belief networks by calculating conditional and marginal probabilities for different network nodes. Bridge deck conditions and different defect types, including corrosion, cracking, delamination, spalling, and scaling, are modeled. The model relies on prior knowledge and what-if analysis to make probabilistic inferences about the system. The practical contribution of the proposed approach lies in extracting the necessary information buried in inspection reports and utilizing such information in building probabilistic graphical models for bridge conditions and deterioration. These models can be further used in making informed decisions regarding bridge intervention strategies.
Developing Probabilistic Condition Assessment Models of Concrete Bridges Utilizing Text Mining
https://orcid.org/http://orcid.org/0000-0003-4696-1678
Bridges are one of the most critical systems of transportation infrastructure. A bridge’s structural safety and serviceability can be negatively impacted by serious deterioration caused by harsh environmental conditions, increased traffic volumes, aging, and deferred maintenance. Transportation agencies conduct comprehensive inspections and assessments to provide a complete picture of a bridge’s present condition and determine accordingly required interventions. The collected inspection data include essential information pertinent to the different types of defects in each bridge component. However, such inspection data are stored in unstructured textual reports without effectively being used. In addition, current condition assessment and deterioration models consider only quantitative data and overlook such descriptive text-type data commonly included in inspection reports. This paper introduces a text-mining-based approach to standardize and automate the condition assessment and deterioration modeling of reinforced concrete bridge decks using Bayesian belief networks. Bayesian belief networks are efficient tools for modeling and visualizing complex system interdependencies in a probabilistic manner by means of marginal and conditional probabilities established on a specific model structure. First, text mining techniques are employed to analyze inspection comments related to bridge decks in order to get insights into these comments. Then, text mining results are used to build Bayesian belief networks by calculating conditional and marginal probabilities for different network nodes. Bridge deck conditions and different defect types, including corrosion, cracking, delamination, spalling, and scaling, are modeled. The model relies on prior knowledge and what-if analysis to make probabilistic inferences about the system. The practical contribution of the proposed approach lies in extracting the necessary information buried in inspection reports and utilizing such information in building probabilistic graphical models for bridge conditions and deterioration. These models can be further used in making informed decisions regarding bridge intervention strategies.
Developing Probabilistic Condition Assessment Models of Concrete Bridges Utilizing Text Mining
Lecture Notes in Civil Engineering
Desjardins, Serge (editor) / Poitras, Gérard J. (editor) / Nik-Bakht, Mazdak (editor) / Omar, Abdelhady (author) / Moselhi, Osama (author)
Canadian Society of Civil Engineering Annual Conference ; 2023 ; Moncton, NB, Canada
Proceedings of the Canadian Society for Civil Engineering Annual Conference 2023, Volume 3 ; Chapter: 21 ; 287-301
2024-10-16
15 pages
Article/Chapter (Book)
Electronic Resource
English
Condition Assessment of Suspension Bridges - A Probabilistic Approach
| British Library Conference Proceedings | 2000
Condition assessment of concrete half joint bridges
| Tema Archive | 2004
Condition Assessment for Existing Reinforced Concrete Bridges
| British Library Conference Proceedings | 2008
Probabilistic models of material properties for design and assessment of concrete bridges
| British Library Conference Proceedings | 2008
CONDITION ASSESSMENT OF CONCRETE BRIDGES DURING DEMOLITION
| British Library Conference Proceedings | 2006