National Center for Transportation Infrastructure Durability and Life-Extension (TriDurLE) – University Transportation Center (UTC)
Project Funds: $114,104 ($54,499 from TriDurLE and $59,605 match from SDSU)
Year: 2020-2021
PI: Mostafa Tazarv, PhD, PE
Co-PI: Kwanghee Won, PhD, Electrical Eng. & Computer Science
Graduate Research Assistants: Youjeong Jang (EECS), Kallan Hart (CEE), Evan Greeneway (CEE)
Project Technical Panel: Ahmad Abu-hawash (Iowa DOT); Bijan Khaleghi (Washington State DOT); Ebrahim Amiri Hormozaki (WSP); Elmer Marx (Alaska DOT); Harvey Coffman (Coffman Engineers, Inc.); Nickolas Murray (Alaska DOT)
Modern seismic design codes ensure a large displacement capacity and prevent total collapse for bridges. However, this performance objective is usually attained at the cost of damage to target ductile members. For reinforced concrete bridges, the columns are usually the main source of ductility during an earthquake in which concrete cover, core, and reinforcement may damage, and the column may experience a large permanent lateral deformation. A significant number of the US bridges will experience large earthquakes in the next 50 years that may result in the bridge closure due to excessive damage. A quick assessment of bridges immediately after severe events is needed to maximize serviceability and access to the affected sites, and to minimize casualties and costs.
The main goal of this proposal, which is the first phase of a multi-phase project, is to accelerate post-earthquake bridge inspection using “computer vision”. Instead of sending trained personnel to the affect bridges, a drone can be used as a fast inspection device. If the drone is equipped with an image processor, which can relate bridge apparent damages to seismic demands, it will be feasible to quickly assess the post-event serviceability of the bridge and to tag the structure (to be opened, closed, or have limited access). Such assessment will save lives and costs since the bridge serviceability will be known to the public and emergency responders. The main product in this phase of the project will be an open-source computer program that can assess bridge bent damage, determine the bridge demand using post-earthquake conditions, and tag a bridge. In the second phase of the project, the software will be implemented in drones and/or mobile applications. Subsequently, other bridge types or elements may be considered for damage assessment in the following phases.
Task 1: Perform a literature review on visual methods of bridge column assessment, performance, and current methods of computer vision,
Task 2: Develop a comprehensive database of the RC bridge column performance including test data and damage photographs,
Task 3: Relate apparent damage to displacement demands through damage index,
Task 4: Develop software that can assess bridge performance based on the visual damage after earthquakes,
Task 5: Validate the software,
Task 6: Prepare a final report documenting all aspects of the project.
Tazarv, M., Won, K., Jang, Y., Hart, K., and Greeneway, E. (2022). “Post-Earthquake Serviceability Assessment of Standard RC Bridge Columns Using Computer Vision and Seismic Analyses,” Engineering Structures, Vol. 272, 115002, 17 pp. (Link).
Tazarv, M., Won, K., Jang, Y., Hart, K., Greeneway, E., and Harshvardhan, A. (2021). “Post-Earthquake Serviceability Assessment of RC Bridge Columns Using Computer Vision,” National Center for Transportation Infrastructure Durability and Life Extension (TriDurLE) Report No: 2020-SDSU-01, Washington State University, Pullman, WA, 338 pp. (Link)
Hart, K., Greeneway, E., and Tazarv, M. (2021). “Modern RC Bridge Column Experimental Database,” in the Post-Earthquake Serviceability of RC Bridge Columns Using Visual Inspection project. DesignSafe-CI PRJ 3294 (Link).
Tazarv, M., Won, K., Jang, Y., Hart, K., and Greeneway, E. (2021). “AI-Based Damage Assessment of RC Bridge Columns after Earthquakes,” The 2022 TRB Conference, Washington, DC, Jan. 11.
Tazarv, M., Kwangee Won, Youjeong Jang, Kallan Hart, Evan Greeneway (2021) “Post-Earthquake Serviceability of RC Bridge Columns Using Computer Vision,” The 2021 TriDurLE Virtual Symposium, Washington State University, Dec. 6-7.
Link to Cloud-Based Post-Earthquake Damage Assessment Tools:
Bridge Damage Assessment Tools (BrDats): Website
Crack Angle Detection
Sample Results of Computer-Vision Software
Samples Results for Two Columns at Damage State 1
Samples Results for Two Columns at Damage State 2
Samples Results for Two Columns at Damage State 3
Samples Results for Two Columns at Damage State 4
Samples Results for Two Columns at Damage State 5
Samples Assessment of Cloud-Based Software