Abstract Fault isolation is essential to fault monitoring, which can be used to detect the cause of the fault. Commonly used methods include contribution plots, LASSO, Nonnegative garrote, construction-based methods, branch and bound algorithm (B & B), etc. However, these existing methods have shortcomings limiting their implementation when there exist vertical outliers and leverage points, Therefore, to further improve the fault prediction accuracy, this paper present a strategy based on robust nonnegative garrote (R-NNG) variable selection algorithm, which is proved to be robust to outliers in the TE process.
Multivariate Fault Isolation in Presence of Outliers Based on Robust Nonnegative Garrote
2017-01-01
10 pages
Article/Chapter (Book)
Electronic Resource
English
Robust Self-tuning Controller under Outliers
British Library Online Contents | 2014
|Robust State Estimation with Sparse Outliers
AIAA | 2015
|Robust State Estimation with Sparse Outliers
Online Contents | 2015
|Robust image matching with cascaded outliers removal
British Library Online Contents | 2017
|