Supervised Health Stage Prediction Using Convolutional Neural Networks for Bearing Wear
Early detection of faults in rotating machinery systems is crucial in preventing system failure, increasing safety, and reducing maintenance costs. Current methods of fault detection suffer from the lack of efficient feature extraction method, the need for designating a threshold producing minimal f...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2020-10-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/20/20/5846 |
| id |
doaj-df9038760d474379bce6d425e32232af |
|---|---|
| record_format |
Article |
| spelling |
doaj-df9038760d474379bce6d425e32232af2020-11-25T02:25:26ZengMDPI AGSensors1424-82202020-10-01205846584610.3390/s20205846Supervised Health Stage Prediction Using Convolutional Neural Networks for Bearing WearSungho Suh0Joel Jang1Seungjae Won2Mayank Shekhar Jha3Yong Oh Lee4Smart Convergence Group, Korea Institute of Science and Technology Europe Forschungsgesellschaft mbH, 66123 Saarbruecken, GermanySmart Convergence Group, Korea Institute of Science and Technology Europe Forschungsgesellschaft mbH, 66123 Saarbruecken, GermanySmart Convergence Group, Korea Institute of Science and Technology Europe Forschungsgesellschaft mbH, 66123 Saarbruecken, GermanyCentre de Recherche en Automatique de Nancy (CRAN), UMR 7039, CNRS, University of Lorraine, 54506 Vandoeuvre CEDEX, FranceSmart Convergence Group, Korea Institute of Science and Technology Europe Forschungsgesellschaft mbH, 66123 Saarbruecken, GermanyEarly detection of faults in rotating machinery systems is crucial in preventing system failure, increasing safety, and reducing maintenance costs. Current methods of fault detection suffer from the lack of efficient feature extraction method, the need for designating a threshold producing minimal false alarm rates, and the need for expert domain knowledge, which is costly. In this paper, we propose a novel data-driven health division method based on convolutional neural networks using a graphical representation of time series data, called a nested scatter plot. The proposed method trains the model with a small amount of labeled data and does not require a threshold value to predict the health state of rotary machines. Notwithstanding the lack of datasets that show the ground truth of health stages, our experiments with two open datasets of run-to-failure bearing demonstrated that our method is able to detect the early symptoms of bearing wear earlier and more efficiently than other threshold-based health indicator methods.https://www.mdpi.com/1424-8220/20/20/5846fault detectionconvolutional neural networksfeature extractionmachinery prognosticshealth stage division |
| collection |
DOAJ |
| language |
English |
| format |
Article |
| sources |
DOAJ |
| author |
Sungho Suh Joel Jang Seungjae Won Mayank Shekhar Jha Yong Oh Lee |
| spellingShingle |
Sungho Suh Joel Jang Seungjae Won Mayank Shekhar Jha Yong Oh Lee Supervised Health Stage Prediction Using Convolutional Neural Networks for Bearing Wear Sensors fault detection convolutional neural networks feature extraction machinery prognostics health stage division |
| author_facet |
Sungho Suh Joel Jang Seungjae Won Mayank Shekhar Jha Yong Oh Lee |
| author_sort |
Sungho Suh |
| title |
Supervised Health Stage Prediction Using Convolutional Neural Networks for Bearing Wear |
| title_short |
Supervised Health Stage Prediction Using Convolutional Neural Networks for Bearing Wear |
| title_full |
Supervised Health Stage Prediction Using Convolutional Neural Networks for Bearing Wear |
| title_fullStr |
Supervised Health Stage Prediction Using Convolutional Neural Networks for Bearing Wear |
| title_full_unstemmed |
Supervised Health Stage Prediction Using Convolutional Neural Networks for Bearing Wear |
| title_sort |
supervised health stage prediction using convolutional neural networks for bearing wear |
| publisher |
MDPI AG |
| series |
Sensors |
| issn |
1424-8220 |
| publishDate |
2020-10-01 |
| description |
Early detection of faults in rotating machinery systems is crucial in preventing system failure, increasing safety, and reducing maintenance costs. Current methods of fault detection suffer from the lack of efficient feature extraction method, the need for designating a threshold producing minimal false alarm rates, and the need for expert domain knowledge, which is costly. In this paper, we propose a novel data-driven health division method based on convolutional neural networks using a graphical representation of time series data, called a nested scatter plot. The proposed method trains the model with a small amount of labeled data and does not require a threshold value to predict the health state of rotary machines. Notwithstanding the lack of datasets that show the ground truth of health stages, our experiments with two open datasets of run-to-failure bearing demonstrated that our method is able to detect the early symptoms of bearing wear earlier and more efficiently than other threshold-based health indicator methods. |
| topic |
fault detection convolutional neural networks feature extraction machinery prognostics health stage division |
| url |
https://www.mdpi.com/1424-8220/20/20/5846 |
| work_keys_str_mv |
AT sunghosuh supervisedhealthstagepredictionusingconvolutionalneuralnetworksforbearingwear AT joeljang supervisedhealthstagepredictionusingconvolutionalneuralnetworksforbearingwear AT seungjaewon supervisedhealthstagepredictionusingconvolutionalneuralnetworksforbearingwear AT mayankshekharjha supervisedhealthstagepredictionusingconvolutionalneuralnetworksforbearingwear AT yongohlee supervisedhealthstagepredictionusingconvolutionalneuralnetworksforbearingwear |
| _version_ |
1724851360779730944 |