A Sparse Autoencoder-Based Unsupervised Scheme for Pump Fault Detection and Isolation
Pumps are one of the most critical machines in the petrochemical process. Condition monitoring of such parts and detecting faults at an early stage are crucial for reducing downtime in the production line and improving plant safety, efficiency and reliability. This paper develops a fault detection a...
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2020-09-01
|
Series: | Applied Sciences |
Subjects: | |
Online Access: | https://www.mdpi.com/2076-3417/10/19/6789 |
id |
doaj-d4fdd0dbfbb84c74a30a297266fa61ce |
---|---|
record_format |
Article |
spelling |
doaj-d4fdd0dbfbb84c74a30a297266fa61ce2020-11-25T03:27:42ZengMDPI AGApplied Sciences2076-34172020-09-01106789678910.3390/app10196789A Sparse Autoencoder-Based Unsupervised Scheme for Pump Fault Detection and IsolationXiaoxia Liang0Fang Duan1Ian Bennett2David Mba3School of Engineering, London South Bank University, London SE1 0AA, UKSchool of Engineering, London South Bank University, London SE1 0AA, UKTechnology Manager Services, Shell Research Ltd., Floor 21, London Shell Centre, London SE1 7NA, UKFaculty of Computing, Engineering and Media, De Montfort University, Leicester LE1 9BH, UKPumps are one of the most critical machines in the petrochemical process. Condition monitoring of such parts and detecting faults at an early stage are crucial for reducing downtime in the production line and improving plant safety, efficiency and reliability. This paper develops a fault detection and isolation scheme based on an unsupervised machine learning method, sparse autoencoder (SAE), and evaluates the model on industrial multivariate data. The Mahalanobis distance (MD) is employed to calculate the statistical difference of the residual outputs between monitoring and normal states and is used as a system-wide health indicator. Furthermore, fault isolation is achieved by a reconstruction-based two-dimensional contribution map, in which the variables with larger contributions are responsible for the detected fault. To demonstrate the effectiveness of the proposed scheme, two case studies are carried out based on a multivariate data set from a pump system in an oil and petrochemical factory. The classical principal component analysis (PCA) method is compared with the proposed method and results show that SAE performs better in terms of fault detection than PCA, and can effectively isolate the abnormal variables, which can hence help effectively trace the root cause of the detected fault.https://www.mdpi.com/2076-3417/10/19/6789sparse autoencodersunsupervised learningmultivariate datafault detectionpump |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Xiaoxia Liang Fang Duan Ian Bennett David Mba |
spellingShingle |
Xiaoxia Liang Fang Duan Ian Bennett David Mba A Sparse Autoencoder-Based Unsupervised Scheme for Pump Fault Detection and Isolation Applied Sciences sparse autoencoders unsupervised learning multivariate data fault detection pump |
author_facet |
Xiaoxia Liang Fang Duan Ian Bennett David Mba |
author_sort |
Xiaoxia Liang |
title |
A Sparse Autoencoder-Based Unsupervised Scheme for Pump Fault Detection and Isolation |
title_short |
A Sparse Autoencoder-Based Unsupervised Scheme for Pump Fault Detection and Isolation |
title_full |
A Sparse Autoencoder-Based Unsupervised Scheme for Pump Fault Detection and Isolation |
title_fullStr |
A Sparse Autoencoder-Based Unsupervised Scheme for Pump Fault Detection and Isolation |
title_full_unstemmed |
A Sparse Autoencoder-Based Unsupervised Scheme for Pump Fault Detection and Isolation |
title_sort |
sparse autoencoder-based unsupervised scheme for pump fault detection and isolation |
publisher |
MDPI AG |
series |
Applied Sciences |
issn |
2076-3417 |
publishDate |
2020-09-01 |
description |
Pumps are one of the most critical machines in the petrochemical process. Condition monitoring of such parts and detecting faults at an early stage are crucial for reducing downtime in the production line and improving plant safety, efficiency and reliability. This paper develops a fault detection and isolation scheme based on an unsupervised machine learning method, sparse autoencoder (SAE), and evaluates the model on industrial multivariate data. The Mahalanobis distance (MD) is employed to calculate the statistical difference of the residual outputs between monitoring and normal states and is used as a system-wide health indicator. Furthermore, fault isolation is achieved by a reconstruction-based two-dimensional contribution map, in which the variables with larger contributions are responsible for the detected fault. To demonstrate the effectiveness of the proposed scheme, two case studies are carried out based on a multivariate data set from a pump system in an oil and petrochemical factory. The classical principal component analysis (PCA) method is compared with the proposed method and results show that SAE performs better in terms of fault detection than PCA, and can effectively isolate the abnormal variables, which can hence help effectively trace the root cause of the detected fault. |
topic |
sparse autoencoders unsupervised learning multivariate data fault detection pump |
url |
https://www.mdpi.com/2076-3417/10/19/6789 |
work_keys_str_mv |
AT xiaoxialiang asparseautoencoderbasedunsupervisedschemeforpumpfaultdetectionandisolation AT fangduan asparseautoencoderbasedunsupervisedschemeforpumpfaultdetectionandisolation AT ianbennett asparseautoencoderbasedunsupervisedschemeforpumpfaultdetectionandisolation AT davidmba asparseautoencoderbasedunsupervisedschemeforpumpfaultdetectionandisolation AT xiaoxialiang sparseautoencoderbasedunsupervisedschemeforpumpfaultdetectionandisolation AT fangduan sparseautoencoderbasedunsupervisedschemeforpumpfaultdetectionandisolation AT ianbennett sparseautoencoderbasedunsupervisedschemeforpumpfaultdetectionandisolation AT davidmba sparseautoencoderbasedunsupervisedschemeforpumpfaultdetectionandisolation |
_version_ |
1724587596330303488 |