Title: Fault detection in sensors by using multiresolution signal decomposition

Authors: Subrat Kumar Mohanty, N.P. Mahalik

Addresses: Department of Mechatronics, Gwangju Institute of Science and Technology, 1 Oryong-dong, Buk-gu, Gwangju 500–712, Korea. ' Department of Mechatronics, Gwangju Institute of Science and Technology, 1 Oryong-dong, Buk-gu, Gwangju 500–712, Korea

Abstract: The detection of faults in sensors is very important as these faults may pose a threat to the safety of the system. In particular, the detection of abrupt faults is very essential in order to improve the safety barrier. In this paper, we have proposed a new and powerful method for the detection of faults. The approach is based on wavelet transform analysis. The main concept in this approach is to decompose the sensor output signal response into other signals, which represent the smoothed and detailed version of the original signal. If there is a change in the original signal because of the fault, then it will be reflected in the decomposed signals and by comparing the decomposed signals with the decomposed versions of a healthy signal, we can the detect and localise the fault.

Keywords: sensors; fault detection; wavelet transform analysis; time-frequency representation; multiresolution signal decomposition; MSD; validation; self-diagnostics; system safety.

DOI: 10.1504/IJISTA.2007.014034

International Journal of Intelligent Systems Technologies and Applications, 2007 Vol.2 No.4, pp.416 - 426

Published online: 12 Jun 2007 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article