Data-driven fault diagnosis of oil rig motor pumps applying automatic definition and selection of features

Wandekokem, Estefhan Dazzi; Franzosi, Frederico Thomaz de Aquino; Rauber, Thomas W.; Varejão, Flávio Miguel; Batista, Rodrigo J.

doi:10.1109/demped.2009.5292765

Cited by 4 publications

(10 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this work we use automatic feature selection methods, that have shown better results so far [3], contrary to the traditional manual way to select features.…”

Section: Introductionmentioning

confidence: 93%

“…The criterion for the selection of features in each step of BSFS was the area under the curve (AUC) [13] resulting from the 10-fold cross-validation done by the SVM on the training data [3]. Having created all these diverse groups of features, we build the classifiers, now diversifying also the parameters C and γ mentioned in Section III.…”

Section: B Experiments Specificationsmentioning

confidence: 99%

“…The second method used in this paper is Best Selected Feature Subsets (BSFS) [3], and uses basically two different greedy search algorithms to perform its selection, SFS and SBS [6].…”

Section: Best Selected Feature Subsetsmentioning

confidence: 99%

“…The other method consists of using Sequential Forward Selection (SFS) and Sequential Backward Selection (SBS) [6], to create a large number of classifiers, and then select some of them to form the final ensemble. This method is called Best Selected Feature Subsets (BSFS) [3]. The main motivation for comparing these two methods is to analyze their strong and weak points and determine which one works better for this particular problem.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

A Comparison of Two Feature-Based Ensemble Methods for Constructing Motor Pump Fault Diagnosis Classifiers

Oliveira

Wandekokem

Mendel

et al. 2010

2010 22nd IEEE International Conference on Tools With Artificial Intelligence

View full text Add to dashboard Cite

This paper presents the results achieved by fault classifier ensembles based on a model-free supervised learning approach for diagnosing faults on oil rigs motor pumps. The main goal is to compare two feature-based ensemble construction methods, and present a third variation from one of them. The use of ensembles instead of single classifier systems has been widely applied in classification problems lately. The diversification of classifiers performed by the methods presented in this work is obtained by varying the feature set each classifier uses, and also at one point, alternating the intrinsic parameters for the training algorithm. We show results obtained with the established genetic algorithm GEFS and our recently developed approach called BSFS, which has a lower computational cost. We rely on a database of real data, with 2000 acquisitions of vibration signals extracted from operational motor pumps. Our results compare the outcomes from the two methods mentioned, and present a modification in one of them that improved the accuracy, reinforcing the motivation for the usage of that method.

show abstract

“…In this work we use automatic feature selection methods, that have shown better results so far [3], contrary to the traditional manual way to select features.…”

Section: Introductionmentioning

confidence: 93%

Section: B Experiments Specificationsmentioning

confidence: 99%

“…The second method used in this paper is Best Selected Feature Subsets (BSFS) [3], and uses basically two different greedy search algorithms to perform its selection, SFS and SBS [6].…”

Section: Best Selected Feature Subsetsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

A Comparison of Two Feature-Based Ensemble Methods for Constructing Motor Pump Fault Diagnosis Classifiers

Oliveira

Wandekokem

Mendel

et al. 2010

2010 22nd IEEE International Conference on Tools With Artificial Intelligence

View full text Add to dashboard Cite

show abstract

“…Though sets with a greater number of features completely contain the sets with fewer features in a sequential search algorithm like SFFS, the classifiers still present different results, and this difference is emphasized by the usage of a cross-validation method in order to automatically tune the numerical parameters for each of the final selected feature sets. This approach was described in (Wandekokem et al, 2009) and the final classifier results will be shown here. To perform the experiments, we divided the complete database into a pair of training data and test data, each one with data obtained from oil rigs that are not used in the complementary base, and keeping the approximated proportion of 2/3 of the examples in the training database and the remaining 1/3 in the test database.…”

Section: Misalignment Fault Diagnosismentioning

confidence: 99%

Pattern Recognition based Fault Diagnosis in Industrial Processes: Review and Application

Rauber¹,

Nascimento²,

Wandekokem³

et al. 2010

Pattern Recognition Recent Advances

View full text Add to dashboard Cite

Constructing feature-based ensemble classifiers for real-world machines fault diagnosis

Marcelo

Estefhan

Mendel

et al. 2010

IECON 2010 - 36th Annual Conference on IEEE Industrial Electronics Society

View full text Add to dashboard Cite

This paper presents the results achieved by fault classifier ensembles based on supervised learning for diagnosing faults on oil rigs motor pumps. The main goal is to apply two feature-based ensemble construction methods to a real-world problem. Recent studies have shown that the use of ensembles of classifiers that are accurate and at the same time have diversifying results can improve the final classification accuracy, compared to a single accurate classifier. The diversification performed by the methods presented in this work is obtained by varying the feature set each classifier uses. We show results obtained with the established genetic algorithm GEFS and a recently developed approach called BSFS, which has lower computational cost. We rely on a database of real data, with 2000 acquisitions of vibration signals extracted from operational motor pumps. Our results show that the ensemble methods had a higher classification accuracy solving a real-world fault diagnosis problem than single classifiers. Also, we present promising results in our experiments with both algorithms, that successfully solves the problem.

show abstract

Data-driven fault diagnosis of oil rig motor pumps applying automatic definition and selection of features

Cited by 4 publications

References 8 publications

A Comparison of Two Feature-Based Ensemble Methods for Constructing Motor Pump Fault Diagnosis Classifiers

A Comparison of Two Feature-Based Ensemble Methods for Constructing Motor Pump Fault Diagnosis Classifiers

Pattern Recognition based Fault Diagnosis in Industrial Processes: Review and Application

Constructing feature-based ensemble classifiers for real-world machines fault diagnosis

Contact Info

Product

Resources

About