Abstract
Real-time modeling and inference are critical for pattern recognition and anomaly detection in dynamic data-driven applications systems (DDDAS), where fast computation and actuation are required for monitoring and active control. An example is mitigation of thermoacoustic instabilities (TAI) in combustion systems; TAI may lead to severe damage in mechanical structures if the frequency of resulting pressure oscillations matches one of the natural frequencies of the combustor structure. The TAI phenomena typically occur on time scales in the order of milliseconds, which must be mitigated by sufficiently fast actuation of control signals. Another example is fatigue damage in mechanical systems, which is a common source of failure in structural materials. Initiation and evolution of fatigue damage are critically dependent on the microstructural initial defects that are usually distributed in a random fashion. Therefore, fatigue damage is a stochastic process, for which early detection is required for condition-based maintenance and service life extension. This chapter addresses the problem of statistical pattern recognition, anomaly detection, and decision-making for control, operation, and maintenance of engineering systems with the aforementioned two applications as typical examples. In particular, a novel framework of symbolic time series analysis (STSA) is developed for pattern recognition and anomaly detection in dynamical systems by using the ergodic theory of measure-preserving transformations (MPTs). Unlike a standard STSA that is built upon the concept of time-homogeneous Markov chains, the MPT-based methodology discussed here generates time-inhomogeneous Markov chains that are capable of significantly enhancing the real-time information of DDDAS by using short-length time series of measurements. The theoretical results have been validated on experimental data collected from two laboratory-scale apparatuses, which consistently show superior performance of the MPT-based STSA with respect to both the standard STSA and a hidden Markov model (HMM).
Original language | English (US) |
---|---|
Title of host publication | Handbook of Dynamic Data Driven Applications Systems |
Subtitle of host publication | Volume 2 |
Publisher | Springer International Publishing |
Pages | 93-120 |
Number of pages | 28 |
Volume | 2 |
ISBN (Electronic) | 9783031279867 |
ISBN (Print) | 9783031279850 |
DOIs | |
State | Published - Jan 1 2023 |
All Science Journal Classification (ASJC) codes
- General Computer Science
- General Mathematics
- General Social Sciences
- General Engineering