AI Tools for Rapid Post-Earthquake Damage Assessment of Bridges with Standard and Substandard Columns

Sponsors:

Alaska Department of Transportation and Public Facilities (DOT&PF)

Project Funds: $300,000

Year: 2024-2026

Personnel:

PI: Mostafa Tazarv, PhD, PE

Co-PI: Kwanghee Won, PhD, Electrical Eng. & Computer Science

Graduate Research Assistants: Pawan Acharya

Project Technical Panel: Nicholas Murray, John Malaby, Ben Fetterhoff

Project Summary:

The Alaska Department of Transportation and Public Facilities (DOT&PF) is responsible for condition assessments of more than 1,000 bridges in the state. Visual inspection is currently the most common method of bridge evaluation after an earthquake, which requires sending trained engineers or inspectors to each affected bridge. Since this practice is time consuming and logistically challenging, the functionality of the affected bridges will be unknown for several hours. An alternative method is needed. For multi-span bridges, their seismic performance usually depends on the bent detailing and behavior. Bridge bents with standard columns, those designed with modern codes, may exhibit moderate-to-high levels of damage under strong earthquakes but collapse is prevented. Bents with substandard columns suffer more and different damages than those seen in standard columns and bridges with substandard columns may collapse. Recently, a pilot study at SDSU has developed a framework that utilizes computer vision, a type of image processing that incorporates artificial intelligence (AI) for analyzing the surroundings, for quick and safe assessment of modern RC bridge columns after earthquakes. To demonstrate the feasibility of the proposed framework, cloud-based software was developed to perform “preliminary damage assessment, PDA” and “detailed damage assessment, DDA” for bridges. AI tools estimate the column damage states based on the extent of the damage. At DDA, a pushover analysis is also performed to estimate displacement demand-to-capacity ratio. Nevertheless, the deployment of this tool at professional level is limited since many bridges include substandard columns.

The main goal of the present project is to develop AI and analytical tools that perform post-earthquake PDA and DDA on both standard and substandard columns that are in-service in Alaska. To achieve this goal, Alaka BrM database and other resources such as drawings and inspection reports will be reviewed to extract the column type and detailing most used in Alaska, a comprehensive literature review on the performance of such columns will be performed, a comprehensive experimental database for substandard columns will be generated, large-scale testing will be performed to establish damage pattern for substandard columns common in Alaska, and AI and other analytical tools that can perform PDA and DDA on the Alaska’s standard and substandard columns will be developed. An expert panel will be formed by DOT&PF and will guide the research team for practical outcomes.