Transforming NDT Inspection Through AI and Workflow Automation: A Digital Leap Toward Smarter Inspections
Tracks
NDT UNLOCKED
Knowledge Level - NDT Level I/NDT Level II
Knowledge Level - NDT Level III
Knowledge Level - Student
Presentation Topic Level - Advanced
Presentation Topic Level - Intermediate
Presentation Topic Level - Novice
Target Audience - General Interest
Target Audience - Level III Managers
Target Audience - Research/Academics
Target Audience - Small Business Managers
Target Audience- NDT Engineers
Target Audience- Technicians/Inspectors
Wednesday, October 8, 2025 |
1:30 PM - 2:00 PM |
Monterey 2-3 |
Speaker
Kathiravan Selvam
Data Science Director
Bakerhughes
Transforming NDT Inspection Through AI and Workflow Automation: A Digital Leap Toward Smarter Inspections
Presentation Description
The Non-Destructive Testing (NDT) industry is undergoing a significant transformation driven by the convergence of Artificial Intelligence (AI) and workflow automation. This presentation explores how these technologies are reshaping inspection processes by increasing speed, accuracy, consistency —while minimizing manual effort and fatigue.
We showcase how AI not only accelerates automatic or assisted defect detection but also enriches the inspection workflow by automatically identifying components and critical defect locations, quantifying their dimensions and severity, and aligning findings with maintenance protocols for faster, more informed decisions. Seamless workflow automation ensures end-to-end efficiency—from data acquisition to reporting—within integrated digital ecosystems.
Latest techniques such as AI-powered denoising significantly improve imaging quality and reduce inspection time in CT scans, while a Synthetic Defect Simulator supports effective AI model training in data-scarce real-world scenarios. The presentation also covers ethical data practices, showcasing how image anonymization enables secure, scalable use of inspection data while maintaining privacy and model quality. We further illustrate how AI itself can be leveraged to enhance data privacy and anonymization processes.
We present real-world case studies, including AI-powered borescope inspections of aircraft engines, enhanced CT scan analysis using AI denoising, defect detection in ultrasonic inspections of rail wheels and thickness evaluation, and synthetic defect generation for radiographic weld inspections.
Crucially, the success of these innovations lies in strong collaboration between NDT subject matter experts, data scientists, and automation technologies. This presentation highlights the multidisciplinary effort required to bring AI from prototype to production, ensuring both technical and operational alignment.
This session offer a forward-looking view of how AI and digital workflow automation are becoming integral to the digital future of NDT, empowering smarter, faster, and more reliable inspections across the industry.
We showcase how AI not only accelerates automatic or assisted defect detection but also enriches the inspection workflow by automatically identifying components and critical defect locations, quantifying their dimensions and severity, and aligning findings with maintenance protocols for faster, more informed decisions. Seamless workflow automation ensures end-to-end efficiency—from data acquisition to reporting—within integrated digital ecosystems.
Latest techniques such as AI-powered denoising significantly improve imaging quality and reduce inspection time in CT scans, while a Synthetic Defect Simulator supports effective AI model training in data-scarce real-world scenarios. The presentation also covers ethical data practices, showcasing how image anonymization enables secure, scalable use of inspection data while maintaining privacy and model quality. We further illustrate how AI itself can be leveraged to enhance data privacy and anonymization processes.
We present real-world case studies, including AI-powered borescope inspections of aircraft engines, enhanced CT scan analysis using AI denoising, defect detection in ultrasonic inspections of rail wheels and thickness evaluation, and synthetic defect generation for radiographic weld inspections.
Crucially, the success of these innovations lies in strong collaboration between NDT subject matter experts, data scientists, and automation technologies. This presentation highlights the multidisciplinary effort required to bring AI from prototype to production, ensuring both technical and operational alignment.
This session offer a forward-looking view of how AI and digital workflow automation are becoming integral to the digital future of NDT, empowering smarter, faster, and more reliable inspections across the industry.
Short Course Description
Biography
Kathiravan Selvam is the Director of Data Science at Waygate Technologies, a Baker Hughes business. He leads the development of advanced AI and cloud-based solutions to digitally transform Non-Destructive Testing (NDT) inspections.
