Use of HMM for evaluation of maintenance activities

Vrignat, Pascal; Avila, Manuel; Duculty, Florent; Kratz, Frédéric

doi:10.1504/ijais.2010.034801

Cited by 8 publications

(22 citation statements)

References 2 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…We chose "λ i " (failure rate) and "µ i " (repair rate) of HMM parameters (Vrignat et al (2010)), to match the maintenance register as closely as possible (Table 1). …”

Section: Simulated Industrial Cmmsmentioning

confidence: 99%

“…HMM topology only depends on matrix elements, where {a ij } = 0 (If all matrix elements are different from zero, we have a "total connectivity matrix"). The transition matrix (A) has been specified in Vrignat et al (2010) …”

Section: Conceptual Synthetic Modelmentioning

confidence: 99%

“…Indeed, it is conceivable for poorly maintained equipment to break down, bringing the entire production line to a halt. In order to help maintenance experts, we built a Synthetic Model based on a reference model, which fits real industrial processes (Vrignat et al (2010)). We showed in Vrignat et al (2010), that this reference model provided good failure prediction.…”

Section: Introductionmentioning

confidence: 99%

“…In order to help maintenance experts, we built a Synthetic Model based on a reference model, which fits real industrial processes (Vrignat et al (2010)). We showed in Vrignat et al (2010), that this reference model provided good failure prediction. This Synthetic Model produced observations or symbols, commonly used in a CMMS 1 .…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Evaluation of Minimal Data Size by Using Entropy, in a HMM Maintenance Manufacturing Use

Robles

Avila

Duculty

et al. 2013

IFAC Proceedings Volumes

Self Cite

View full text Add to dashboard Cite

In this paper, we wish to find a minimal data size in order to better conceptualize industrial maintenance activities. We based our study on data given by a Synthetic Hidden Markov Model. This synthetic model is intended to produce real industrial maintenance observations (or "symbols"), with a corresponding degradation indicator. These time series events are shown as Markov chains, also called "signatures". The production of symbols is generated by using a uniform and a normal distribution. The evaluation is made by applying Shannon entropy on the HMM parameters. The results show a minimal number of data for each distribution studied. After a discussion about the use of a new "Sliding Window" of symbols usable in a Computerized Maintenance Management System, we developed two industrial applications and compare them with the best optimized "signature" previously found.

show abstract

“…We chose "λ i " (failure rate) and "µ i " (repair rate) of HMM parameters (Vrignat et al (2010)), to match the maintenance register as closely as possible (Table 1). …”

Section: Simulated Industrial Cmmsmentioning

confidence: 99%

Section: Conceptual Synthetic Modelmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Evaluation of Minimal Data Size by Using Entropy, in a HMM Maintenance Manufacturing Use

Robles

Avila

Duculty

et al. 2013

IFAC Proceedings Volumes

Self Cite

View full text Add to dashboard Cite

show abstract

“…Un quatrième et dernier niveau (état), indique que le système est en panne, imposant une situation de dépannage. Le modèle de synthèse utilisé est basé sur la topologie proposée par Vrignat et al [16], afin de reproduire les conditions réelles d'un processus industriel soumis à des « perturbations » : pannes, réparations, maintenances préventives, visite, etc. Dans ces conditions, nous produisons ainsi, des séquences d'observations (ou symboles) identiques à celles issues d'une GMAO.…”

Section: Introductionunclassified

Modélisation du niveau de dégradation d’un système industriel à l’aide de modèles de Markov cachés

Robles¹,

Avila²,

Duculty³

et al. 2015

Congrès Lambda Mu 19 De Maîtrise Des Risques Et Sûreté De Fonctionnement

View full text Add to dashboard Cite

RésuméLes modèles de Markov cachés ou HMM 1 sont largement utilisés dans les domaines de la reconnaissance des formes, de la parole ainsi que dans la modélisation de processus complexes. Nous proposons dans cette étude d'évaluer la pertinence des paramètres de modèles de Markov cachés de façon objective sans connaissance a priori. Nous présentons des critères permettant d'évaluer la pertinence d'événements stochastiques issus de modèles de Markov cachés. Nous étayons notre étude en nous appuyant sur l'exemple concret d'un processus industriel. Nous évaluons alors les paramètres de sortie des différents modèles testés sur ce processus, pour finalement s'orienter vers le modèle le plus pertinent. Nous validons nos méthodes de mesures de pertinence en retrouvant la topologie ayant servi à construire le modèle. Nous confrontons nos résultats provenant d'un modèle de synthèse avec des données provenant du processus industriel maintenu. Nous concluons alors sur la topologie qui traduit le mieux la dégradation d'un système réel. SummaryHidden Markov Models are widely used in the fields of pattern recognition, speech and the modeling of complex processes. We propose in this study to assess the relevance of the parameters of hidden Markov models without a priori knowledge. First, we present different criteria usually used in model selection. We study a concrete example of an industrial process. We then evaluate the output parameters of the different models tested in this process, to finally find the most appropriate model. We validate our methods by finding the topology used to build the model. We compare our results with synthetic data. We conclude that the topology 2 best reflects the degradation of a real system.

show abstract