Unsupervised condition change detection in large diesel engines

Author

Pontoppidan, Niels Henrik ; Larsen, Jan

Author_Institution

Informatics & Mathematical Modelling, Tech. Univ. Denmark, Lyngby, Denmark

fYear

2003

fDate

17-19 Sept. 2003

Firstpage

565

Lastpage

574

Abstract

This work presents a new method for unsupervised change detection, which combines independent component modeling, and probabilistic outlier detection. The method further provides a compact data representation, which is amenable to interpretation, i.e., the detected condition changes can be investigated further. The method is successfully applied to unsupervised condition change detection in large diesel engines from acoustical emission sensor signal and compared to more classical techniques based on principal component analysis and Gaussian mixture models.

Keywords

Gaussian processes; acoustic emission; acoustic signal processing; adaptive signal processing; condition monitoring; diesel engines; independent component analysis; principal component analysis; unsupervised learning; Gaussian mixture models; acoustical emission sensor signal; diesel engines; independent component modeling; principal component analysis; probabilistic outlier detection; unsupervised change detection; Acoustic sensors; Acoustic signal detection; Additive noise; Bayesian methods; Diesel engines; Electronic mail; Independent component analysis; Informatics; Principal component analysis; Training data;

fLanguage

English

Publisher

ieee

Conference_Titel

Neural Networks for Signal Processing, 2003. NNSP'03. 2003 IEEE 13th Workshop on

ISSN

1089-3555

Print_ISBN

0-7803-8177-7

Type

conf

DOI

10.1109/NNSP.2003.1318056

Filename

1318056