Health Index Estimation Through Integration of General Knowledge with Unsupervised Learning
Authors:
Kristupas Bajarunas,
Marcia L. Baptista,
Kai Goebel,
Manuel A. Chao
Abstract:
Accurately estimating a Health Index (HI) from condition monitoring data (CM) is essential for reliable and interpretable prognostics and health management (PHM) in complex systems. In most scenarios, complex systems operate under varying operating conditions and can exhibit different fault modes, making unsupervised inference of an HI from CM data a significant challenge. Hybrid models combining…
▽ More
Accurately estimating a Health Index (HI) from condition monitoring data (CM) is essential for reliable and interpretable prognostics and health management (PHM) in complex systems. In most scenarios, complex systems operate under varying operating conditions and can exhibit different fault modes, making unsupervised inference of an HI from CM data a significant challenge. Hybrid models combining prior knowledge about degradation with deep learning models have been proposed to overcome this challenge. However, previously suggested hybrid models for HI estimation usually rely heavily on system-specific information, limiting their transferability to other systems. In this work, we propose an unsupervised hybrid method for HI estimation that integrates general knowledge about degradation into the convolutional autoencoder's model architecture and learning algorithm, enhancing its applicability across various systems. The effectiveness of the proposed method is demonstrated in two case studies from different domains: turbofan engines and lithium batteries. The results show that the proposed method outperforms other competitive alternatives, including residual-based methods, in terms of HI quality and their utility for Remaining Useful Life (RUL) predictions. The case studies also highlight the comparable performance of our proposed method with a supervised model trained with HI labels.
△ Less
Submitted 8 May, 2024;
originally announced May 2024.
Room-Temperature Plasmon-Assisted Resonant THz Detection in Single-layer Graphene Transistors
Authors:
José M. Caridad,
Óscar Castelló,
Sofía M. López Baptista,
Takashi Taniguchi,
Kenji Watanabe,
Hartmut G. Roskos,
Juan A. Delgado-Notario
Abstract:
Frequency-selective or even frequency-tunable Terahertz (THz) photodevices are critical components for many technological applications that require nanoscale manipulation, control and confinement of light. Within this context, gate-tunable phototransistors based on plasmonic resonances are often regarded as the most promising devices for frequency-selective detection of THz fields. The exploitatio…
▽ More
Frequency-selective or even frequency-tunable Terahertz (THz) photodevices are critical components for many technological applications that require nanoscale manipulation, control and confinement of light. Within this context, gate-tunable phototransistors based on plasmonic resonances are often regarded as the most promising devices for frequency-selective detection of THz fields. The exploitation of constructive interference of plasma waves in such detectors not only promises frequency selectivity, but also a pronounced sensitivity enhancement at the target frequencies. However, clear signatures of plasmon-assisted resonances in THz detectors have been only revealed at cryogenic temperatures so far, and remain unobserved at application-relevant room-temperature conditions. In this work, we demonstrate the sought-after room-temperature resonant detection of THz radiation in short-channel gated photodetectors made from high-quality single-layer graphene. The survival of this intriguing resonant regime at room-temperature ultimately relies on the weak intrinsic electron-phonon scattering in graphene, which avoids the dam** of the plasma oscillations.
△ Less
Submitted 8 January, 2024;
originally announced January 2024.
