IDENTIFICATION USING THE ALGORITHM OF SINGULAR VALUE DECOMPOSITION—AN APPLICATION TO THE REALISATION OF DYNAMIC SYSTEMS AND TO FAULT DETECTION AND LOCALISATION
✍ Scribed by A. Lenzen; H. Waller
- Publisher
- Elsevier Science
- Year
- 1997
- Tongue
- English
- Weight
- 238 KB
- Volume
- 11
- Category
- Article
- ISSN
- 0888-3270
No coin nor oath required. For personal study only.
✦ Synopsis
A new algorithm-the generalised singular value decomposition (GSVD)-is used in the field of identification and fault detection and localisation. By using the GSVD a simultaneous factorisation of two matrices A and B is possible. Therefore, a state space realisation of a structure can be established from measured data of a real system. The measured data are arranged in Hankel matrices and the parameters of the state space equation can be found with a special procedure using the generalised singular value decomposition. Changes in the dynamic system behaviour that can be detected by new measurements and realisations can be used for fault detection and localisation. A sensitive method for this task is established, with the singular value decomposition in combination with dynamic influence coefficients, that result from a spectral decomposition of the weighting operator.