Link to Latest Report: Final Report
Background:
Departments of Transportation (DOTs) and other government agencies that oversee the maintenance of bridges will benefit from an objective approach to the allocation of funds for maintenance projects. The outcomes of this research project will provide a framework for an automated criticality analysis to support funding decisions, assist in forecasting maintenance expenditures, and strengthen the case for budget allocations with data-driven justification. For example, the Federal Highway Administration has funds set aside through the Bridge Investment Program for rehabilitating bridges in poor condition or fair condition at risk of falling into poor condition. With an objective analysis, these “fair condition at risk of falling into poor condition” bridges can be clearly identified. More importantly, the automated criticality analysis is not limited to bridges in fair condition only but can be applied to bridges in all conditions and ages, and for managing deferred maintenance backlogs.
Objective :
The objective of this project is to use data from Oklahoma, Texas, Florida, and Georgia bridges sourced from the LTBP InfoBridge platform to develop a framework for automated criticality analysis. The framework that results from this project will be used as the basis for developing a proof-of-concept for automated criticality analysis using real-world cases. The proof of concept will be used to further develop a more robust framework, augmented with more data, that will inform the architecture for licensable software.
Scope :
Task 1 – Data collection
The sample for data collection for this project will include a diverse set of bridges from Oklahoma, Texas, Florida, and Georgia, representing a range of conditions, ages, traffic volumes, and structural types. The project will rely primarily on the LTBP InfoBridge Platform’s NBI, NBE, and MERRA-2 Data, and support these with state and local DOT datasets to conduct regional scale investigations in Oklahoma, Texas, Florida, and Georgia that will result in the synthesis of bridge element condition states, codification of bridge maintenance experts’ knowledge, historical bridge maintenance records, and fiscal resources for maintenance into a high-fidelity universal criticality index.
Data collection will involve collaboration with the Oklahoma DOT and other agencies responsible for infrastructure maintenance. The data gathered will include element-level bridge inspection records, maintenance histories, and codified workflows used by DOT personnel for bridge deferred maintenance decision-making. A Memorandum of Understanding (MOU) will be established with participating DOTs to facilitate access to their data repositories, including any available automated monitoring data, maintenance records, and recent inspection records. This collaboration will ensure that the data gathered aligns with real-world needs and priorities. DOT staff will be invited to provide input on maintenance priorities, which will help tailor the criticality analysis to their specific operational and budgetary needs.
Task 2 – Data Analysis/Organization
The goal of the framework is to compute indices that represent the overall condition of individual bridges, through an automated evaluation of the element-level condition, historical maintenance records, and potential impacts on the transportation system. For example, to prioritize one bridge over another for maintenance action, the element-level conditions will be ranked based on the codified expert knowledge, and the system will automatically calculate a composite index of the criticality of its undesirable condition. The bridge with the higher index will be deemed to have a higher priority for maintenance action over that with a lower index. This automatically computed index will help bridge owners, DOTs, and broadly, the Federal Highway Administration (FHWA) rapidly respond to maintenance needs from a resource planning and maintenance scheduling perspective in a data-driven fashion and promptly.
Collected data will be organized into categories based on bridge age, structural condition (current condition based on NBI element data, LTBP data, or DOT data), traffic volume (number and type of vehicles using the bridge daily), environmental exposure (exposure to saltwater, changes in water levels, extreme temperatures, flooding history), and maintenance history. A criticality index will be developed to rank bridges based on urgency and the potential impact of maintenance deferral on the bridge and transportation needs. Data analysis will employ statistical and artificial intelligence/machine learning (AI) approaches like random forest and deep neural networks.
First, we will clean the data to remove inconsistencies, identify, and apply appropriate methods to handle missing data. Data cleaning protocols will be determined together with experts from the DOTs that we will engage with. To ensure uniformity, we will normalize the varying datasets involved in the analysis.
Task 3 – Framework Integration
The automated criticality analysis framework will integrate data from various sources described in earlier sections to produce an objective, evidence-based tool for prioritizing bridge maintenance. The framework will be designed to be adaptable to incorporate data from other types of infrastructure assets in future applications.
Task 4 – Evaluation of Framework
We will perform a cost-benefit analysis to evaluate the usefulness of the model from a time- and cost-saving perspective. Specifically, we will conduct a limited deployment of the framework on selected bridges, chosen in collaboration with our DOT partners. To validate the outcomes of the framework, we will have the same bridges considered for maintenance decision-making by well-qualified DOT personnel using their conventional approach. We will compare the time taken to arrive at prioritizing maintenance needs using the framework with the time taken by using a conventional approach. This will help provide evidence of the utility and benefit of the framework. We can also assess the validity of the framework’s output by comparing the rankings from both tests.
Task 5 – Reporting and Dissemination
A final report will be produced to document the research findings, development of the framework, and evaluation outcomes, including the cost-benefit analysis. We will produce a roadmap for scaling up the proposed solution that specifies the required technical and strategic steps for implementation in other states. Dissemination activities will involve submitting research papers to peer-reviewed journals, presenting at academic and industry conferences, and sharing results with DOTs to encourage the adoption of the automated criticality framework.
Research Team:
Principal Investigator: Kofi Asare, Ph.D.
Co-Principal Investigator: Matthew Reyes and Somik Ghosh
Previous Progress Reports: