 Event Information |
|
CALCE Webinar - Digital Twin-Driven Prognostics and Health Management for Digital Circuits
This seminar presents an overview of CALCE's research into a digital twin-based prognostics and health management (PHM) approach for digital circuits. Building upon CALCE's previous PHM work for analog circuits that provided component-level fault diagnosis and prognosis using only the output waveform from the circuit, this new approach focuses on analyzing the transition behavior of digital circuit binary output voltage signals. It integrates transfer learning and digital twin technologies to enhance model robustness and accuracy. Temporal, spectral, and hybrid features are extracted from these transitions. For fault type classification, transfer learning methods are employed for feature transformation and selection, which are fed into a classification model that localizes the degraded circuit components. A regression model is then used to estimate the severity of the degradation. This fault value can be used in conjunction with component degradation models to complete the prognosis task, which provides the remaining useful life for the circuit. The digital twin operates through iterative updating of the models to improve their fidelity to the real world, ensuring that the predictions can be used for effective decision-making for product sustainment and failure prevention. About the Presenters: Huwei Dong received his B.S. degree in Mechanical Engineering from the University of Cincinnati, Cincinnati, OH, USA, in 2022, and his B.Eng. degree in Mechanical Engineering from Chongqing University, Chongqing, China, in 2022. He is currently pursuing a Ph.D. in Mechanical Engineering at the University of Maryland, College Park, MD, USA. His research interests include prognostics and health management, digital twins, and machine learning. Michael Azarian is a Research Scientist at the Center for Advanced Life Cycle Engineering (CALCE) at the University of Maryland. He holds a Masters and Ph.D. in Materials Science and Engineering from Carnegie Mellon University and a Bachelor’s degree in Chemical Engineering from Princeton University. His research is focused on the analysis, detection, prediction, and prevention of failures in electronic and electromechanical products. He has over 150 publications on electronic packaging, component reliability, prognostics and health management, and tribology. He holds 6 U.S. patents. Prior to joining CALCE in 2004, he spent over 13 years in the data storage, advanced materials, and fiber optics industries. --> This Event is For: Public
|
Digital circuits are the brains and the communication centers for modern industrial systems, so their failures can cause significant system downtime in industrial sectors such as manufacturing, telecommunications, transportation, and healthcare. Traditional testing methods can help identify the causes of failures after they occur, but they do not provide early warning prior to a malfunction, thus risking failures and unplanned maintenance.
