Static and dynamic novelty detection methods for jet engine health monitoring
- 12 December 2006
- journal article
- review article
- Published by The Royal Society in Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences
- Vol. 365 (1851) , 493-514
Abstract
Novelty detection requires models of normality to be learnt from training data known to be normal. The first model considered in this paper is a static model trained to detect novel events associated with changes in the vibration spectra recorded from a jet engine. We describe how the distribution of energy across the harmonics of a rotating shaft can be learnt by a support vector machine model of normality. The second model is a dynamic model partially learnt from data using an expectation–maximization-based method. This model uses a Kalman filter to fuse performance data in order to characterize normal engine behaviour. Deviations from normal operation are detected using the normalized innovations squared from the Kalman filter.Keywords
This publication has 12 references indexed in Scilit:
- Novelty detection: a review—part 1: statistical approachesSignal Processing, 2003
- Constructing boosting algorithms from SVMs: an application to one-class classificationIEEE Transactions on Pattern Analysis and Machine Intelligence, 2002
- Particle filtering-based fault detection in non-linear stochastic systemsInternational Journal of Systems Science, 2002
- A Unifying Review of Linear Gaussian ModelsNeural Computation, 1999
- Robust Model-Based Fault Diagnosis for Dynamic Systems1999
- A System for the Analysis of Jet Engine Vibration DataIntegrated Computer-Aided Engineering, 1999
- Shadow targetsNeurocomputing, 1998
- The Nature of Statistical Learning Theory1995
- A training algorithm for optimal margin classifiers1992
- A Nonlinear Mapping for Data Structure AnalysisIEEE Transactions on Computers, 1969