ADVANCED MATERIALS & PROCESSES | JANUARY/FEBRUARY 2025 29 Manufacturing a completely defect- free part is nearly impossible. Results can get close, but with great expense—as the aerospace industry has learned when pursuing flight certification. In most major industries, a certain amount of imperfection in a finished part is acceptable providing it meets dimensional and strength standards and performs as required in the real world. Yet, when making significant quantities of a part, waiting until the end of the production line to test for quality is hardly cost-efficient. Testing and discarding products that do not meet the standards of a particular industry wastes human resources, time, and materials. Good questions to ask as early as possible in the production cycle include: Is the process as efficient and cost-effective as possible? What enhancements could be made? How can incremental changes and their impact on finished-part metrics be determined? How can design changes improve manufacturing outcomes? Which minor defects—pores, cracks, inclusions—are nothing to worry about and which ones might cause problems? INDUSTRIAL COMPUTED TOMOGRAPHY One method that provides valuable data to help answer such essential questions is industrial computed tomography (CT) scanning. From a TECHNICAL SPOTLIGHT HOW AI CAN MAKE A DIFFERENCE IN THE REAL WORLD OF MANUFACTURING Industrial computed tomography data analysis is harnessing deep learning to both accelerate in-line inspection and build better products. huge volumes of 3D data and readies it for interpretation. The software’s analysis tools can then be used to automatically call out and capture the geometries of a wide variety of physical defects. This enables engineers to compare such flaws against standards, pinpoint design or manufacturing errors that may have contributed, and make changes in designs and/or processes to correct those that are unacceptable. Such capabilities have been evolving rapidly over the past decade, proving their worth in aerospace, automotive, and other industries. Statistical process-control software traditional manufacturing process like casting to the most advanced metal additive manufacturing, this type of nondestructive testing (NDT) at critical stages of design and production is a highly effective way to investigate part quality. The x-ray strength of industrial CT can penetrate almost any material, from composites to metals to advanced alloys. In fact, this technology is the only NDT method that can enable materials analysis of the entire volume of a part (Fig. 1). Advanced visualization and analysis software applied to such scans extracts Fig. 1 — The location and size of foreign particles are displayed in a software analysis of a CT scan of an automotive battery.
RkJQdWJsaXNoZXIy MTYyMzk3NQ==