Turbopump condition monitoring using incremental clustering and one-class support vector machine

Lei Hu, Niaoqing Hu, Guojun Qin, Fengshou Gu

Research output: Contribution to journalArticle

5 Citations (Scopus)

Abstract

Turbopump condition monitoring is a significant approach to ensure the safety of liquid rocket engine (LRE). Because of lack of fault samples, a monitoring system cannot be trained on all possible condition patterns. Thus it is important to differentiate abnormal or unknown patterns from normal pattern with novelty detection methods. One-class support vector machine (OCSVM) that has been commonly used for novelty detection cannot deal well with large scale samples. In order to model the normal pattern of the turbopump with OCSVM and so as to monitor the condition of the turbopump, a monitoring method that integrates OCSVM with incremental clustering is presented. In this method, the incremental clustering is used for sample reduction by extracting representative vectors from a large training set. The representative vectors are supposed to distribute uniformly in the object region and fulfill the region. And training OCSVM on these representative vectors yields a novelty detector. By applying this method to the analysis of the turbopump's historical test data, it shows that the incremental clustering algorithm can extract 91 representative points from more than 36 000 training vectors, and the OCSVM detector trained on these 91 representative points can recognize spikes in vibration signals caused by different abnormal events such as vane shedding, rub-impact and sensor faults. This monitoring method does not need fault samples during training as classical recognition methods. The method resolves the learning problem of large samples and is an alternative method for condition monitoring of the LRE turbopump.

Original languageEnglish
Pages (from-to)474-479
Number of pages6
JournalChinese Journal of Mechanical Engineering (English Edition)
Volume24
Issue number3
DOIs
Publication statusPublished - 1 May 2011

Fingerprint

Condition monitoring
Support vector machines
Rocket engines
Monitoring
Detectors
Liquids
Clustering algorithms
Sensors

Cite this

@article{a09a2e497bd048a4b33c702629fdd0e7,
title = "Turbopump condition monitoring using incremental clustering and one-class support vector machine",
abstract = "Turbopump condition monitoring is a significant approach to ensure the safety of liquid rocket engine (LRE). Because of lack of fault samples, a monitoring system cannot be trained on all possible condition patterns. Thus it is important to differentiate abnormal or unknown patterns from normal pattern with novelty detection methods. One-class support vector machine (OCSVM) that has been commonly used for novelty detection cannot deal well with large scale samples. In order to model the normal pattern of the turbopump with OCSVM and so as to monitor the condition of the turbopump, a monitoring method that integrates OCSVM with incremental clustering is presented. In this method, the incremental clustering is used for sample reduction by extracting representative vectors from a large training set. The representative vectors are supposed to distribute uniformly in the object region and fulfill the region. And training OCSVM on these representative vectors yields a novelty detector. By applying this method to the analysis of the turbopump's historical test data, it shows that the incremental clustering algorithm can extract 91 representative points from more than 36 000 training vectors, and the OCSVM detector trained on these 91 representative points can recognize spikes in vibration signals caused by different abnormal events such as vane shedding, rub-impact and sensor faults. This monitoring method does not need fault samples during training as classical recognition methods. The method resolves the learning problem of large samples and is an alternative method for condition monitoring of the LRE turbopump.",
keywords = "Condition monitoring, Incremental clustering, Novelty detection, One-class support vector machine, Turbopump",
author = "Lei Hu and Niaoqing Hu and Guojun Qin and Fengshou Gu",
year = "2011",
month = "5",
day = "1",
doi = "10.3901/CJME.2011.03.474",
language = "English",
volume = "24",
pages = "474--479",
journal = "Chinese Journal of Mechanical Engineering (English Edition)",
issn = "1000-9345",
publisher = "Springer Verlag",
number = "3",

}

Turbopump condition monitoring using incremental clustering and one-class support vector machine. / Hu, Lei; Hu, Niaoqing; Qin, Guojun; Gu, Fengshou.

In: Chinese Journal of Mechanical Engineering (English Edition), Vol. 24, No. 3, 01.05.2011, p. 474-479.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Turbopump condition monitoring using incremental clustering and one-class support vector machine

AU - Hu, Lei

AU - Hu, Niaoqing

AU - Qin, Guojun

AU - Gu, Fengshou

PY - 2011/5/1

Y1 - 2011/5/1

N2 - Turbopump condition monitoring is a significant approach to ensure the safety of liquid rocket engine (LRE). Because of lack of fault samples, a monitoring system cannot be trained on all possible condition patterns. Thus it is important to differentiate abnormal or unknown patterns from normal pattern with novelty detection methods. One-class support vector machine (OCSVM) that has been commonly used for novelty detection cannot deal well with large scale samples. In order to model the normal pattern of the turbopump with OCSVM and so as to monitor the condition of the turbopump, a monitoring method that integrates OCSVM with incremental clustering is presented. In this method, the incremental clustering is used for sample reduction by extracting representative vectors from a large training set. The representative vectors are supposed to distribute uniformly in the object region and fulfill the region. And training OCSVM on these representative vectors yields a novelty detector. By applying this method to the analysis of the turbopump's historical test data, it shows that the incremental clustering algorithm can extract 91 representative points from more than 36 000 training vectors, and the OCSVM detector trained on these 91 representative points can recognize spikes in vibration signals caused by different abnormal events such as vane shedding, rub-impact and sensor faults. This monitoring method does not need fault samples during training as classical recognition methods. The method resolves the learning problem of large samples and is an alternative method for condition monitoring of the LRE turbopump.

AB - Turbopump condition monitoring is a significant approach to ensure the safety of liquid rocket engine (LRE). Because of lack of fault samples, a monitoring system cannot be trained on all possible condition patterns. Thus it is important to differentiate abnormal or unknown patterns from normal pattern with novelty detection methods. One-class support vector machine (OCSVM) that has been commonly used for novelty detection cannot deal well with large scale samples. In order to model the normal pattern of the turbopump with OCSVM and so as to monitor the condition of the turbopump, a monitoring method that integrates OCSVM with incremental clustering is presented. In this method, the incremental clustering is used for sample reduction by extracting representative vectors from a large training set. The representative vectors are supposed to distribute uniformly in the object region and fulfill the region. And training OCSVM on these representative vectors yields a novelty detector. By applying this method to the analysis of the turbopump's historical test data, it shows that the incremental clustering algorithm can extract 91 representative points from more than 36 000 training vectors, and the OCSVM detector trained on these 91 representative points can recognize spikes in vibration signals caused by different abnormal events such as vane shedding, rub-impact and sensor faults. This monitoring method does not need fault samples during training as classical recognition methods. The method resolves the learning problem of large samples and is an alternative method for condition monitoring of the LRE turbopump.

KW - Condition monitoring

KW - Incremental clustering

KW - Novelty detection

KW - One-class support vector machine

KW - Turbopump

UR - http://www.scopus.com/inward/record.url?scp=79958002377&partnerID=8YFLogxK

U2 - 10.3901/CJME.2011.03.474

DO - 10.3901/CJME.2011.03.474

M3 - Article

VL - 24

SP - 474

EP - 479

JO - Chinese Journal of Mechanical Engineering (English Edition)

JF - Chinese Journal of Mechanical Engineering (English Edition)

SN - 1000-9345

IS - 3

ER -