Session: ASME Undergraduate Student Design Expo

Paper Number: 165072

Integration of Robotic Inspection, Machine Vision, and Artificial Intelligence to Enhance Welding Inspection 

To reduce the time, expenses, risks, and operational interruptions linked to structural inspections, researchers have increasingly explored robotic systems as a means to enhance existing inspection methods. These robotic systems, equipped with advanced sensing technologies such as ultrasonic, infrared, LiDAR, and optical cameras, allow for remote, autonomous, and high-precision inspections in challenging environments. By integrating artificial intelligence (AI) and machine learning (ML) algorithms, these systems can process vast amounts of data in real time, improving defect detection accuracy and reducing human error. Additionally, robotic platforms, such as drones for aerial inspections, crawlers for confined spaces, and robotic arms for complex geometries, enable access to areas that would otherwise be hazardous or impractical for human inspectors. The adoption of robotic inspection not only enhances safety and reliability but also enables continuous structural health monitoring (SHM), allowing for predictive maintenance strategies that mitigate the risk of catastrophic failures. As research advances, the incorporation of sensor fusion techniques, edge computing, and wireless data transmission is further optimizing robotic inspection systems, making them more adaptive, cost-effective, and scalable for widespread industrial applications. With the expanding range of industrial robots, the adoption of these systems has significantly increased. Their applications now cover a wide spectrum of civil infrastructure projects, with new robotic technologies being rapidly implemented. However, due to the highly interdisciplinary nature of this research area, every component of robotic inspection must be meticulously designed and analyzed by experts in the field. One critical aspect of robotic inspection is the integration of advanced nondestructive testing (NDT) methods, which ensure accurate defect detection and structural assessment. NDT techniques, such as advanced phased array ultrasonic testing (PAUT), serve as a cornerstone for quality inspections in infrastructure projects, including the evaluation of weld integrity in structural components. Standards and codes, such as AWS D1.5 for welding inspection in bridges, are also developed based on these NDT testing methods to address the quality inspection requirements of the structures in various conditions which must be met. This study examines the integration of collaborative robots (cobots) into PAUT systems for NDT of infrastructures. The findings indicate that while employing cobots for NDT inspection of welded structural elements often involves balancing the need for advanced sensing capabilities with the challenges of maneuvering in complex field environments, their use offers distinct advantages. These include enhanced repeatability of results and consistent inspection procedure. This research seeks to overcome the traditional inspection limitations by developing an automated inspection framework that combines robotic scanning with AI-driven data analysis and Machine Vision (MV), enabling real-time, high-precision detection and characterization of welding flaws.

Presenting Author: Julia Oubre Georgia Southern University

Presenting Author Biography: Julia Oubre is affiliated with the Mechanical Engineering and the Manufacturing Engineering departments, as well as the Honors College at Georgia Southern University. She is pursuing dual bachelor’s degrees in mechanical engineering and manufacturing engineering, as well as a master’s degree in mechanical engineering. Her research interests include robotic automation, additive manufacturing, nondestructive testing, and mechanism design.

Authors:

Julia Oubre Georgia Southern University
Elsie Lappin Georgia Southern University
Hossein Taheri Georgia Southern University