Performance of Jensen Shannon Divergence in Incipient Fault Detection and Estimation
Xiaoxia Zhang, Claude Delpha, Demba Diallo
Abstract
The diagnosis (Detection, Estimation and Isolation) of incipient faults, i.e. faults with severity variation <;10% of the healthy signal, plays an important role in health monitoring of complex system for earliest maintenance. In this paper, Jensen-Shannon divergence (JSD) is proposed to evaluate detection and estimation of incipient fault severity in a multivariate data driven process. At first, the dimensional space is reduced, thanks to the Principal Component Analysis (PCA). Thus, the fault effect is theoretically modelled using the JSD considering Gaussian distributed signals. Then, the estimation of the fault severity is derived from this model. The fault detection performances are then computed and compared with the Hotelling's T <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> statistics. The fault estimation performances validates the theoretical modeling for incipient faults in noisy environments. An estimation error lower than 3% is obtained even for a Signal to Noise Ratio (SNR) as low as 25dB.
BibTeX
@inproceedings{icassp2019_performanceofjen,
title = {Performance of Jensen Shannon Divergence in Incipient Fault Detection and Estimation},
author = {Xiaoxia Zhang and Claude Delpha and Demba Diallo},
booktitle = {ICASSP 2019},
year = {2019}
}