ADVANCED MATERIALS & PROCESSES | JANUARY/FEBRUARY 2025 10 MACHINE LEARNING | AI The new “Chatbots Decoded: Exploring AI” exhibit at the Computer History Museum, Mountain View, Calif., explores why chatbots matter. Ameca, a robot from Engineered Arts that uses hardware, sensors, and large language models like Open AI’s GPT-4, will interact with visitors. computerhistory.org. BRIEF MACHINE LEARNING ENHANCES DEFECT DETECTION Researchers from Northwestern University, University of Virginia, Car- negie Mellon University, and Argonne National Laboratory made a signi- ficant advancement in defect detec- tion for laser powder bed fusion (LPBF) additive manufacturing. By using sensors such as microphones and photodiodes along with machine learning, they achieved over 90% accuracy with a temporal resolution of 0.1 ms in detecting keyhole pore formation. This breakthrough could lead to a faster certification process for metal AM parts. Detecting keyhole pores in real- time during the printing process has been challenging due to the speed and complexity of LPBF. To address this, the team developed a machine learning-based approach that uses light and sound sensors to monitor the printing process and accurately detect when and where keyhole pores form. The key to this method lies in measuring the oscillations of the keyhole, a vapor depression formed in the melt pool during printing. High-speed synchrotron x-ray imaging was used to help train the machine learning model to recognize conditions that lead to pore formation. With this approach, manufacturers could detect defects during printing, allowing for adjustments to prevent production of flawed parts. northwestern.edu. AI LISTENS FOR BATTERY ABOUT TO CATCH FIRE Researchers at the National Insti- tute of Standards and Technology (NIST) are using sound to detect when lithium- ion batteries are about to catch fire— especially dangerous because a battery can emit a flame up to 2012°F in about a second. “While watching videos of exploding batteries, I noticed something interesting,” says researcher Andy Tam. “Right before the fire started, the safety valve in the battery broke and it made this little noise. I thought we might be able to use that.” Tam was A new method detects keyhole pore formation in laser powder bed fusion. Courtesy of Northwestern University. not the first to make this observation, but he wanted to see what he could do with it. Many of the hard casings used on Li-ion batteries contain a safety valve designed to break and release pressure when needed. This breaking valve is the sound Tam heard in the videos, a distinctive click-hiss. The researchers needed software that could recognize this sound and not detect other noises, so they trained a machine learning algorithm. Through a collaboration with Xi’an University of Science and Technology, they recorded audio from 38 exploding batteries. Then they tweaked those recordings into more than 1000 unique audio samples to teach the software what a breaking safety valve sounds like. Using a microphone mounted on a camera, the team detected the sound of an overheating battery 94% of the time. nist.gov. An experiment records the sounds a Li-ion battery makes before and during thermal runaway. Courtesy of Xi’an University.
RkJQdWJsaXNoZXIy MTYyMzk3NQ==