Optimized multi-shot imaging inspection design.

Area of Science:

• Industrial Quality Assurance
• Non-Destructive Evaluation (NDE)
• Computational Imaging

Background:

• Inspecting complex-shaped components, common in additive manufacturing, poses significant challenges for industrial quality assurance.
• Traditional methods like X-ray computed tomography have drawbacks, while 2D radiography offers limited capability.
• Designing detailed inspections for complex geometries is labor-intensive and requires substantial expert knowledge.

Purpose of the Study:

• To develop a computational framework for optimizing data acquisition in image-based inspection modalities.
• To advance the capabilities of radiography for inspecting complex components, with potential applicability to other imaging techniques.
• To automate the design of inspection configurations, including optimal component poses relative to the imaging system.

Main Methods:

• Development of an optimization algorithm leveraging prior inspection information and modality simulations.
• Computation of spatial maps to visualize and understand spatially varying limits of detection.
• Innovation in evaluating defect detectability for complex indications and creating application-specific optimization algorithms.

Main Results:

• The developed framework successfully determines optimal inspection configurations, including component poses.
• Spatial maps of inspection performance provide insights into detection limits across different regions.
• Initial trials demonstrated good results, validating the algorithm's effectiveness in enhancing inspection capabilities.

Conclusions:

• The computational framework offers a significant advancement in optimizing image-based inspection for complex components.
• The approach enhances defect detectability and reduces the expert input required for inspection design.
• This methodology holds promise for improving non-destructive evaluation in industrial quality assurance, particularly for additively manufactured parts.