Incipient Fault Severity Estimation Using Local Mahalanobis Distance

Junjie Yang; Claude Delpha

ICASSP 2022accepted0 citations

Incipient Fault Severity Estimation Using Local Mahalanobis Distance

Abstract

Recently, the Local Mahalanobis Distance (LMD) technique was proposed for incipient fault detection, which was shown to be sensitive, robust and distribution assumption-free. Further, this paper explicitly establishes the relation between fault severity and LMD index to take advantage of those excellent characteristics for fault severity estimation. The performance of the estimation model is evaluated based on a benchmark case of Continuous-flow Stirred Tank Reactor (CSTR) process, which shows a high accuracy even for tiny deviation of signal and large signal-to-noise ratio.

BibTeX

@inproceedings{icassp2022_incipientfaultse,
  title = {Incipient Fault Severity Estimation Using Local Mahalanobis Distance},
  author = {Junjie Yang and Claude Delpha},
  booktitle = {ICASSP 2022},
  year = {2022}
}