ADVANCED MATERIALS & PROCESSES | JULY/AUGUST 2024 10 MACHINE LEARNING | AI DEEP MACHINE LEARNING DETECTS DEFECTS Researchers at the University of Illinois Urbana-Champaign developed a new method to find defects in additively manufactured (AM) parts. Because AM is often used to print components with complex 3D shapes and internal features, it can be especially challenging to locate these defects. The technology uses deep machine learning to make this task much easier. To build their model, the team used computer simulations to generate tens of thousands of synthetic defects that exist only in the computer. Each of these computer-generated defects has a different size, shape, and location, allowing the deep learning model to train on a wide variety of possible defects and to recognize the difference between good and bad components. The algorithm was then tested on numerous real parts, some defective and some perfect. It was able to correctly identify hundreds of defects in the physical parts that were not previously seen by the deep learning model. The team employed x-ray computed tomography (CT) to inspect the interior of the 3D components with internal features and defects hidden from view. illinois.edu. AI ACCELERATES MATERIALS DISCOVERY Researchers at the DOE’s Oak Ridge National Laboratory (ORNL), Tenn., are developing ways to accelerate discovery by combining automated experiments, artificial intelligence, and high performance computing. A new tool developed at the lab leverages those technologies to demonstrate that AI can influence materials synthesis and conduct associated experiments—without human supervision. This autonomous materials synthesis tool uses pulsed laser deposition (PLD) to deposit a thin layer of On May 15, a bipartisan U.S. Senate working group announced a new legislative plan for artificial intelligence. Several subject matter experts from Carnegie Mellon University, Pittsburgh, contributed knowledge to the group as the plan was developed over many months. The roadmap, “Driving U.S. Innovation in Artificial Intelligence,” directs Congress to infuse billions of dollars into R&D and takes a step forward in regulating AI. cmu.edu. BRIEF Longitudinal (top) and axial (bottom) images of x-ray CT data of parts with six internal defects: spherical clog, stellated shaped clog, cone shaped void, blob shaped void, elliptical warp of inner channel, and a nonconcentric center nozzle. substance onto a base material. It then employs AI to analyze how the quality of the newly created material relates to the synthesis conditions such as temperature, pressure, and energy emitted during the PLD process. The AI then suggests a revised set of conditions that may yield improved quality and then controls the PLD equipment to conduct the next experiment and so on. “We built computer control of all processes into the system and incorporated some hardware innovations to enable AI to drive experimentation,” says researcher Sumner Harris of ORNL’s Center for Nanophase Materials Sciences. “The automation allows us to perform our work 10 times faster, and the AI can understand huge parameter spaces with far fewer samples.” ornl.gov. An automated deposition system places a new material onto a base material (purple beam, right) as the last sample that was made is analyzed and sent to the AI (green beams, brain, left). The AI tells the PLD machine what to do next (data cable, bottom). Courtesy of Chris Rouleau/ORNL.
RkJQdWJsaXNoZXIy MTYyMzk3NQ==