Real-Time Machine Learning for High-Rate Structural Health Monitoring
Date
2021-10-24
Authors
Major Professor
Advisor
Committee Member
Journal Title
Journal ISSN
Volume Title
Publisher
Springer, Cham
Abstract
Advances in science and engineering are empowering high-rate dynamic systems, such as hypersonic vehicles, advanced weaponries, and active shock and blast mitigation strategies. The real-time estimation of the structural health of high-rate systems, termed high-rate structural health monitoring (HRSHM), is critical in designing decision mechanisms that can ensure structural integrity and performance. However, this is a difficult task, because three aspects uniquely characterize these systems: (1) large uncertainties in the external loads; (2) high levels of non-stationarities and heavy disturbances; and (3) unmodeled dynamics generated from changes in system configurations. In addition, because these systems are experiencing events of high amplitudes (often beyond 100 g) over short durations (under 100 ms), a successful feedback mechanism is one that can operate under 1 ms. A solution to the unique system characteristics and temporal constraint is the design and application of real-time learning algorithms. Here, we review and discuss a real-time learning algorithm for HRSHM applications. In particular, after introducing the HRSHM challenge, we explore fast real-time learning for time series prediction using conventional and deep neural networks and discuss a path to rapid real-time state estimation.
Series Number
Journal Issue
Is Version Of
Versions
Series
Type
Book chapter
Comments
This chapter is published as Laflamme S., Hu C., Dodson J. (2022) Real-Time Machine Learning for High-Rate Structural Health Monitoring. In: Cury A., Ribeiro D., Ubertini F., Todd M.D. (eds) Structural Health Monitoring Based on Data Science Techniques. Structural Integrity, vol 21. Springer, Cham. DOI: 10.1007/978-3-030-81716-9_4. Works produced by employees of the U.S. Government as part of their official duties are not copyrighted within the U.S. The content of this document is not copyrighted.