Fault Detection in Mechanical Systems With Friction Phenomena: An Online Neural Approximation Approach
Authors:
- George A. RovithakisGRGeorge A. Rovithakis
- Thomas ParisiniTPThomas Parisini
- 1 July 2007
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Neural Networks
- Vol. 18 (4) , 1067-1082
Abstract
In this paper, the problem of fault detection in mechanical systems performing linear motion, under the action of friction phenomena is addressed. The friction effects are modeled through the dynamic LuGre model. The proposed architecture is built upon an online neural network (NN) approximator, which requires only system's position and velocity. The friction internal state is not assumed to be available for measurement. The neural fault detection methodology is analyzed with respect to its robustness and sensitivity properties. Rigorous fault detectability conditions and upper bounds for the detection time are also derived. Extensive simulation results showing the effectiveness of the proposed methodology are provided, including a real case study on an industrial actuator.Keywords
This publication has 32 references indexed in Scilit:
- A robust detection and isolation scheme for abrupt and incipient faults in nonlinear systemsIEEE Transactions on Automatic Control, 2002
- Reinforcement adaptive learning neural-net-based friction compensation control for high speed and precisionIEEE Transactions on Control Systems Technology, 2000
- Incipient fault diagnosis of dynamical systems using online approximatorsIEEE Transactions on Automatic Control, 1998
- Deterministic nonlinear observer-based approaches to fault diagnosis: A surveyControl Engineering Practice, 1997
- State observation, failure detection and isolation (FDI) in bilinear systemsInternational Journal of Control, 1997
- A bilinear fault detection filterInternational Journal of Control, 1997
- Automated fault detection and accommodation: a learning systems approachIEEE Transactions on Systems, Man, and Cybernetics, 1995
- Universal approximation bounds for superpositions of a sigmoidal functionIEEE Transactions on Information Theory, 1993
- Regularization Algorithms for Learning That Are Equivalent to Multilayer NetworksScience, 1990
- The Stone-Weierstrass theorem and its application to neural networksIEEE Transactions on Neural Networks, 1990