ADVANCED MATERIALS & PROCESSES | MAY/JUNE 2024 10 MACHINE LEARNING | AI VISION TRANSFORMERS OPTIMIZE 3D PRINTING Researchers at Carnegie Mellon University, Pittsburgh, developed a new system that employs ultra-high-speed in situ imaging and vision transformers to optimize process parameters for 3D printing a variety of metal alloys. These parameters include critical process details such as printing speed, laser power, and layer thickness of the deposited material. Vision transformers are a type of machine learning that applies neural network architectures originally developed for natural language processing tasks to computer vision tasks such as image classification. Video vision transformers go even further by using video sequences instead of still images to capture both spatial and temporal relationships that enable the model to learn complex patterns from video data. “We are excited to have developed an AI method that leverages temporal features in AM imaging data to detect different types of defects. Demonstrating the generalizability of the AI method using different AM metals is groundbreaking and reveals that the same trained AI model can be employed without costly retraining using additional data,” explains Conrad Tucker, mechanical engineering professor. The team used funding from the Army Research Laboratory to develop a high-speed imaging setup to capture clear features of molten metal and a machine learning model that could see the patterns associated with the defects they were trying to detect and prevent. By using vision transformers to classify different kinds of defects that can occur during 3D printing, the algorithmic accuracy surpassed 90% depending on the material. The scientists developed an off- axial imaging setup using a high-speed video camera and magnification lens to capture high-frequency oscillation in the melt pool shape. Video was recorded with a temporal resolution of over 50,000 frames per second. The videos were classified into four categories: one desirable regime and three undesirable regimes that produced defects including keyholing, balling, and lack of fusion. These four printing regimes were then tested on 316L stainless steel, Ti-6AL-4V, and Inconel 718. The team found that video vision transformers with temporal embedding can enable in situ detection of melt pool defects with a simple off-axial imaging setup. Using this video data, researchers are able to generate process maps that could accelerate qualification of printability for newly developed metal alloys specifically designed for 3D printing. engineering. cmu.edu. AI TRAINING BOOSTS BRITISH METALS INDUSTRY A new training center at the University of Leicester in the U.K. plans to boost the metals industry by developing postgraduates with excellent skills in data and artificial intelligence. The new $22 million Molten pools of Ti-6Al-4V captured by videos of different printing regimes. Centre for Doctoral Training (CDT) in Digital Transformation of Metals Industry (DigitalMetal) was funded by the Engineering and Physical Sciences Research Council (EPSRC), five U.K. partner universities (Birmingham, Leicester, Loughborough, Nottingham, and Warwick), and 35 industrial members. The effort is part of the U.K.’s largest investment ever in engineering and physical sciences doctoral skills, totaling more than $1 billion. Sixty-five CDTs will support leading research in AI, quantum technologies, semiconductors, telecommunications, and engineering biology. The DigitalMetal CDT is designed to meet a national strategic need for training a new generation of technical leaders able to guide the digital transformation of the metals industry and its supply chain with the objective of increasing agility, productivity, and international competitiveness. The pro- gram will provide postgraduate training that combines metals and alloys engineering with digital technology and AI skills, to help the U.K.’s domestic metals and manufacturing industries reap the benefits of big data. The goal is to develop future industry leaders who can rapidly take advantage of the latest discoveries in manufacturing processes through digital twinning to enable defect-free manufacturing at reduced costs. le.ac.uk. Enhanced AI training aims to boost the U.K. metals industry. Courtesy of www.vpnsrus.com.
RkJQdWJsaXNoZXIy MTYyMzk3NQ==