A High-Dimensional and Small-Sample Submersible Fault Detection Method Based on Feature Selection and Data Augmentation
The fault detection of manned submersibles plays a very important role in protecting the safety of submersible equipment and personnel. However, the diving sensor data is scarce and high-dimensional, so this paper proposes a submersible fault detection method, which is made up of feature selection m...
Main Authors: | , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2021-12-01
|
Series: | Sensors |
Subjects: | |
Online Access: | https://www.mdpi.com/1424-8220/22/1/204 |
_version_ | 1797497599027052544 |
---|---|
author | Penghui Zhao Qinghe Zheng Zhongjun Ding Yi Zhang Hongjun Wang Yang Yang |
author_facet | Penghui Zhao Qinghe Zheng Zhongjun Ding Yi Zhang Hongjun Wang Yang Yang |
author_sort | Penghui Zhao |
collection | DOAJ |
description | The fault detection of manned submersibles plays a very important role in protecting the safety of submersible equipment and personnel. However, the diving sensor data is scarce and high-dimensional, so this paper proposes a submersible fault detection method, which is made up of feature selection module based on hierarchical clustering and Autoencoder (AE), the improved Deep Convolutional Generative Adversarial Networks (DCGAN)-based data augmentation module and fault detection module using Convolutional Neural Network (CNN) with LeNet-5 structure. First, feature selection is developed to select the features that have a strong correlation with failure event. Second, data augmentation model is conducted to generate sufficient data for training the CNN model, including rough data generation and data refiners. Finally, a fault detection framework with LeNet-5 is trained and fine-tuned by synthetic data, and tested using real data. Experiment results based on sensor data from submersible hydraulic system demonstrate that our proposed method can successfully detect the fault samples. The detection accuracy of proposed method can reach 97% and our method significantly outperforms other classic detection algorithms. |
first_indexed | 2024-03-10T03:21:32Z |
format | Article |
id | doaj.art-d43dd09dadc243fab7a60924e8094c84 |
institution | Directory Open Access Journal |
issn | 1424-8220 |
language | English |
last_indexed | 2024-03-10T03:21:32Z |
publishDate | 2021-12-01 |
publisher | MDPI AG |
record_format | Article |
series | Sensors |
spelling | doaj.art-d43dd09dadc243fab7a60924e8094c842023-11-23T12:18:36ZengMDPI AGSensors1424-82202021-12-0122120410.3390/s22010204A High-Dimensional and Small-Sample Submersible Fault Detection Method Based on Feature Selection and Data AugmentationPenghui Zhao0Qinghe Zheng1Zhongjun Ding2Yi Zhang3Hongjun Wang4Yang Yang5School of Information Science and Engineering, Shandong University, Qingdao 266237, ChinaSchool of Information Science and Engineering, Shandong University, Qingdao 266237, ChinaChina National Deep Sea Center, Qingdao 266237, ChinaChina National Deep Sea Center, Qingdao 266237, ChinaSchool of Information Science and Engineering, Shandong University, Qingdao 266237, ChinaSchool of Information Science and Engineering, Shandong University, Qingdao 266237, ChinaThe fault detection of manned submersibles plays a very important role in protecting the safety of submersible equipment and personnel. However, the diving sensor data is scarce and high-dimensional, so this paper proposes a submersible fault detection method, which is made up of feature selection module based on hierarchical clustering and Autoencoder (AE), the improved Deep Convolutional Generative Adversarial Networks (DCGAN)-based data augmentation module and fault detection module using Convolutional Neural Network (CNN) with LeNet-5 structure. First, feature selection is developed to select the features that have a strong correlation with failure event. Second, data augmentation model is conducted to generate sufficient data for training the CNN model, including rough data generation and data refiners. Finally, a fault detection framework with LeNet-5 is trained and fine-tuned by synthetic data, and tested using real data. Experiment results based on sensor data from submersible hydraulic system demonstrate that our proposed method can successfully detect the fault samples. The detection accuracy of proposed method can reach 97% and our method significantly outperforms other classic detection algorithms.https://www.mdpi.com/1424-8220/22/1/204fault detectionfeature selectiondata augmentationhigh-dimensional sensor datalimited fault eventmanned submersible |
spellingShingle | Penghui Zhao Qinghe Zheng Zhongjun Ding Yi Zhang Hongjun Wang Yang Yang A High-Dimensional and Small-Sample Submersible Fault Detection Method Based on Feature Selection and Data Augmentation Sensors fault detection feature selection data augmentation high-dimensional sensor data limited fault event manned submersible |
title | A High-Dimensional and Small-Sample Submersible Fault Detection Method Based on Feature Selection and Data Augmentation |
title_full | A High-Dimensional and Small-Sample Submersible Fault Detection Method Based on Feature Selection and Data Augmentation |
title_fullStr | A High-Dimensional and Small-Sample Submersible Fault Detection Method Based on Feature Selection and Data Augmentation |
title_full_unstemmed | A High-Dimensional and Small-Sample Submersible Fault Detection Method Based on Feature Selection and Data Augmentation |
title_short | A High-Dimensional and Small-Sample Submersible Fault Detection Method Based on Feature Selection and Data Augmentation |
title_sort | high dimensional and small sample submersible fault detection method based on feature selection and data augmentation |
topic | fault detection feature selection data augmentation high-dimensional sensor data limited fault event manned submersible |
url | https://www.mdpi.com/1424-8220/22/1/204 |
work_keys_str_mv | AT penghuizhao ahighdimensionalandsmallsamplesubmersiblefaultdetectionmethodbasedonfeatureselectionanddataaugmentation AT qinghezheng ahighdimensionalandsmallsamplesubmersiblefaultdetectionmethodbasedonfeatureselectionanddataaugmentation AT zhongjunding ahighdimensionalandsmallsamplesubmersiblefaultdetectionmethodbasedonfeatureselectionanddataaugmentation AT yizhang ahighdimensionalandsmallsamplesubmersiblefaultdetectionmethodbasedonfeatureselectionanddataaugmentation AT hongjunwang ahighdimensionalandsmallsamplesubmersiblefaultdetectionmethodbasedonfeatureselectionanddataaugmentation AT yangyang ahighdimensionalandsmallsamplesubmersiblefaultdetectionmethodbasedonfeatureselectionanddataaugmentation AT penghuizhao highdimensionalandsmallsamplesubmersiblefaultdetectionmethodbasedonfeatureselectionanddataaugmentation AT qinghezheng highdimensionalandsmallsamplesubmersiblefaultdetectionmethodbasedonfeatureselectionanddataaugmentation AT zhongjunding highdimensionalandsmallsamplesubmersiblefaultdetectionmethodbasedonfeatureselectionanddataaugmentation AT yizhang highdimensionalandsmallsamplesubmersiblefaultdetectionmethodbasedonfeatureselectionanddataaugmentation AT hongjunwang highdimensionalandsmallsamplesubmersiblefaultdetectionmethodbasedonfeatureselectionanddataaugmentation AT yangyang highdimensionalandsmallsamplesubmersiblefaultdetectionmethodbasedonfeatureselectionanddataaugmentation |