Unsupervised Methods for Anomalies Detection through Intelligent Monitoring Systems

Carrascal, Alberto; Dı́ez, A.; Azpeitia, Ander

doi:10.1007/978-3-642-02319-4_17

Cited by 6 publications

(6 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Operational parameters values are normalized between 0 and 1. It can be seen as a simplified Kohonen map or network with a small number of nodes, from lower to higher engine load, and no neighborhood function when updating the BMU (best matching unit) at each iteration [35] .…”

Section: Resultsmentioning

confidence: 99%

Data-driven prognostics using a combination of constrained K-means clustering, fuzzy modeling and LOF-based score

et al. 2017

View full text Add to dashboard Cite

Today, failure modes characterization and early detection is a key issue in complex assets. This is due to the negative impact of corrective operations and the conservative strategies usually put in practice, focused on preventive maintenance. In this paper anomaly detection issue is addressed in new monitoring sensor data by characterizing and modeling operational behaviors. The learning framework is performed on the basis of a machine learning approach that combines constrained K-means clustering for outlier detection and fuzzy modeling of distances to normality. A final score is also calculated over time, considering the membership degree to resulting fuzzy sets and a local outlier factor. Proposed solution is deployed in a CBM+ platform for online monitoring of the assets. In order to show the validity of the approach, experiments have been conducted on real operational faults in an auxiliary marine diesel engine. Experimental results show a fully comprehensive yet accurate prognostics approach, improving detection capabilities and knowledge management. The performance achieved is quite high (precision, sensitivity and specificity above 93% and κ = 0. 93), even more so given that a very small percentage of real faults are present in data.

show abstract

Section: Resultsmentioning

confidence: 99%

Data-driven prognostics using a combination of constrained K-means clustering, fuzzy modeling and LOF-based score

et al. 2017

View full text Add to dashboard Cite

show abstract

“…There exists other kind of classifiers, such as non-supervised or semi-supervised, usually applied to detect anomalies when no labels are provided or the knowledge on data under study is limited [19].…”

Section: Evaluation Of Assessment Values Sensitivity To Blind Fastenimentioning

confidence: 99%

A non-destructive quality assessment for blind-fastener installations based on the combination of ultrasound techniques and real-time monitoring of the fastening process

Camacho

Rivero

Veiga

et al. 2015

Int J Adv Manuf Technol

View full text Add to dashboard Cite

This paper proposes and evaluates a non-destructive method in assessing the quality of blind-fastener installations. The method is based on combining the time of flight of an ultrasonic pulse along the pin of the fastener and the fastener installation time. The former has to be measured after the fastening process; the latter has to be acquired at the machine during the fastening. The tests performed on carbon fibre reinforced plastic laminates joined by means of ABS0257 blind fasteners have proved that the method can distinguish incorrectly installed fasteners due to improper grip selection. The hit rate of the developed non-destructive method is above 95 % for the tested range of fastener sizes. The implementation of this method might prevent the use of extra fasteners, currently required, in order to compensate for the uncertainty of blind-fastener installations.

show abstract

“…This is due to the nature of the data set used, each experiment or sample is classified according to resulting burr and s values. There exist other classification methods, such as non-supervised or supervised classification, whose applicability is more oriented to detect anomalies when knowledge of data behaviour is limited or even inexistent [7 supervised classification paradigm [8] consists of a set of N facts, each of t 1 variables; first n variables, X 1 , X 2 ,…, X n , would be predictor variables and the variable with index n+1, identified as C, would be the These data can be represented in table format using the following notation:…”

Section: Supervised Classificationmentioning

confidence: 99%

A Multiclassifier Approach for Drill Wear Prediction

Dı́ez

Carrascal

2012

Machine Learning and Data Mining in Pattern Recognition

Self Cite

View full text Add to dashboard Cite

Abstract. Classification methods have been widely used during last years in order to predict patterns and trends of interest in data. In present paper, a multiclassifier approach that combines the output of some of the most popular data mining algorithms is shown. The approach is based on voting criteria, by estimating the confidence distributions of each algorithm individually and combining them according to three different methods: confidence voting, weighted voting and majority voting. To illustrate its applicability in a real problem, the drill wear detection in machine-tool sector is addressed. In this study, the accuracy obtained by each isolated classifier is compared with the performance of the multiclassifier when characterizing the patterns of interest involved in the drilling process and predicting the drill wear. Experimental results show that, in general, false positives obtained by the classifiers can be slightly reduced by using the multiclassifier approach.

show abstract

Unsupervised Methods for Anomalies Detection through Intelligent Monitoring Systems

Cited by 6 publications

References 6 publications

Data-driven prognostics using a combination of constrained K-means clustering, fuzzy modeling and LOF-based score

Data-driven prognostics using a combination of constrained K-means clustering, fuzzy modeling and LOF-based score

A non-destructive quality assessment for blind-fastener installations based on the combination of ultrasound techniques and real-time monitoring of the fastening process

A Multiclassifier Approach for Drill Wear Prediction

Contact Info

Product

Resources

About