Kullback-Leibler distance-based enhanced detection of incipient anomalies

Fouzi Harrou, Ying Sun, Muddu Madakyaru

Research output: Contribution to journalArticle

16 Citations (Scopus)

Abstract

Accurate and effective anomaly detection and diagnosis of modern engineering systems by monitoring processes ensure reliability and safety of a product while maintaining desired quality. In this paper, an innovative method based on Kullback-Leibler divergence for detecting incipient anomalies in highly correlated multivariate data is presented. We use a partial least square (PLS) method as a modeling framework and a symmetrized Kullback-Leibler distance (KLD) as an anomaly indicator, where it is used to quantify the dissimilarity between current PLS-based residual and reference probability distributions obtained using fault-free data. Furthermore, this paper reports the development of two monitoring charts based on the KLD. The first approach is a KLD-Shewhart chart, where the Shewhart monitoring chart with a three sigma rule is used to monitor the KLD of the response variables residuals from the PLS model. The second approach integrates the KLD statistic into the exponentially weighted moving average monitoring chart. The performance of the PLS-based KLD anomaly-detection methods is illustrated and compared to that of conventional PLS-based anomaly detection methods. Using synthetic data and simulated distillation column data, we demonstrate the greater sensitivity and effectiveness of the developed method over the conventional PLS-based methods, especially when data are highly correlated and small anomalies are of interest. Results indicate that the proposed chart is a very promising KLD-based method because KLD-based charts are, in practice, designed to detect small shifts in process parameters.

Original languageEnglish
Pages (from-to)73-87
Number of pages15
JournalJournal of Loss Prevention in the Process Industries
Volume44
DOIs
Publication statusPublished - 01-11-2016

Fingerprint

least squares
Least-Squares Analysis
Monitoring
monitoring
Process monitoring
Distillation columns
Systems engineering
Probability distributions
methodology
Statistics
process monitoring
product safety
probability distribution
Distillation
distillation
Anomaly
statistics
Charts
Partial least squares
Safety

All Science Journal Classification (ASJC) codes

  • Control and Systems Engineering
  • Food Science
  • Chemical Engineering(all)
  • Safety, Risk, Reliability and Quality
  • Energy Engineering and Power Technology
  • Management Science and Operations Research
  • Industrial and Manufacturing Engineering

Cite this

@article{2893d2f784fd4f2a94f428bce8d5b6ce,
title = "Kullback-Leibler distance-based enhanced detection of incipient anomalies",
abstract = "Accurate and effective anomaly detection and diagnosis of modern engineering systems by monitoring processes ensure reliability and safety of a product while maintaining desired quality. In this paper, an innovative method based on Kullback-Leibler divergence for detecting incipient anomalies in highly correlated multivariate data is presented. We use a partial least square (PLS) method as a modeling framework and a symmetrized Kullback-Leibler distance (KLD) as an anomaly indicator, where it is used to quantify the dissimilarity between current PLS-based residual and reference probability distributions obtained using fault-free data. Furthermore, this paper reports the development of two monitoring charts based on the KLD. The first approach is a KLD-Shewhart chart, where the Shewhart monitoring chart with a three sigma rule is used to monitor the KLD of the response variables residuals from the PLS model. The second approach integrates the KLD statistic into the exponentially weighted moving average monitoring chart. The performance of the PLS-based KLD anomaly-detection methods is illustrated and compared to that of conventional PLS-based anomaly detection methods. Using synthetic data and simulated distillation column data, we demonstrate the greater sensitivity and effectiveness of the developed method over the conventional PLS-based methods, especially when data are highly correlated and small anomalies are of interest. Results indicate that the proposed chart is a very promising KLD-based method because KLD-based charts are, in practice, designed to detect small shifts in process parameters.",
author = "Fouzi Harrou and Ying Sun and Muddu Madakyaru",
year = "2016",
month = "11",
day = "1",
doi = "10.1016/j.jlp.2016.08.020",
language = "English",
volume = "44",
pages = "73--87",
journal = "Journal of Loss Prevention in the Process Industries",
issn = "0950-4230",
publisher = "Elsevier BV",

}

Kullback-Leibler distance-based enhanced detection of incipient anomalies. / Harrou, Fouzi; Sun, Ying; Madakyaru, Muddu.

In: Journal of Loss Prevention in the Process Industries, Vol. 44, 01.11.2016, p. 73-87.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Kullback-Leibler distance-based enhanced detection of incipient anomalies

AU - Harrou, Fouzi

AU - Sun, Ying

AU - Madakyaru, Muddu

PY - 2016/11/1

Y1 - 2016/11/1

N2 - Accurate and effective anomaly detection and diagnosis of modern engineering systems by monitoring processes ensure reliability and safety of a product while maintaining desired quality. In this paper, an innovative method based on Kullback-Leibler divergence for detecting incipient anomalies in highly correlated multivariate data is presented. We use a partial least square (PLS) method as a modeling framework and a symmetrized Kullback-Leibler distance (KLD) as an anomaly indicator, where it is used to quantify the dissimilarity between current PLS-based residual and reference probability distributions obtained using fault-free data. Furthermore, this paper reports the development of two monitoring charts based on the KLD. The first approach is a KLD-Shewhart chart, where the Shewhart monitoring chart with a three sigma rule is used to monitor the KLD of the response variables residuals from the PLS model. The second approach integrates the KLD statistic into the exponentially weighted moving average monitoring chart. The performance of the PLS-based KLD anomaly-detection methods is illustrated and compared to that of conventional PLS-based anomaly detection methods. Using synthetic data and simulated distillation column data, we demonstrate the greater sensitivity and effectiveness of the developed method over the conventional PLS-based methods, especially when data are highly correlated and small anomalies are of interest. Results indicate that the proposed chart is a very promising KLD-based method because KLD-based charts are, in practice, designed to detect small shifts in process parameters.

AB - Accurate and effective anomaly detection and diagnosis of modern engineering systems by monitoring processes ensure reliability and safety of a product while maintaining desired quality. In this paper, an innovative method based on Kullback-Leibler divergence for detecting incipient anomalies in highly correlated multivariate data is presented. We use a partial least square (PLS) method as a modeling framework and a symmetrized Kullback-Leibler distance (KLD) as an anomaly indicator, where it is used to quantify the dissimilarity between current PLS-based residual and reference probability distributions obtained using fault-free data. Furthermore, this paper reports the development of two monitoring charts based on the KLD. The first approach is a KLD-Shewhart chart, where the Shewhart monitoring chart with a three sigma rule is used to monitor the KLD of the response variables residuals from the PLS model. The second approach integrates the KLD statistic into the exponentially weighted moving average monitoring chart. The performance of the PLS-based KLD anomaly-detection methods is illustrated and compared to that of conventional PLS-based anomaly detection methods. Using synthetic data and simulated distillation column data, we demonstrate the greater sensitivity and effectiveness of the developed method over the conventional PLS-based methods, especially when data are highly correlated and small anomalies are of interest. Results indicate that the proposed chart is a very promising KLD-based method because KLD-based charts are, in practice, designed to detect small shifts in process parameters.

UR - http://www.scopus.com/inward/record.url?scp=84985906099&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84985906099&partnerID=8YFLogxK

U2 - 10.1016/j.jlp.2016.08.020

DO - 10.1016/j.jlp.2016.08.020

M3 - Article

AN - SCOPUS:84985906099

VL - 44

SP - 73

EP - 87

JO - Journal of Loss Prevention in the Process Industries

JF - Journal of Loss Prevention in the Process Industries

SN - 0950-4230

ER -