Next Generation of Process Monitoring and Diagnostics: Applications of AI and Machine Learning to Enable Early Equipment Fault Prediction and Diagnostics

Kaushik Ghosh; Gokula Krishnan Sivaprakasam; Aparnadevi Minisankar

doi:doi:10.11648/j.ajai.20220601.13

| Peer-Reviewed

Next Generation of Process Monitoring and Diagnostics: Applications of AI and Machine Learning to Enable Early Equipment Fault Prediction and Diagnostics

Kaushik Ghosh, Gokula Krishnan Sivaprakasam, Aparnadevi Minisankar

Published in American Journal of Artificial Intelligence (Volume 6, Issue 1)

Received: 20 January 2022 Accepted: 10 March 2022 Published: 9 April 2022

Views: Downloads:

Download PDF

Share This Article

Twitter
Linked In
Facebook

Abstract

Several rotating equipment such as – centrifugal pumps and positive displacement pumps are extensively used in Water treatment plant for producing potable water from raw water. Centrifugal pumps are required for delivering water from one unit of the plant to the others, while the positive displacement pumps are used for dosing different chemicals at the various stages of water treatment process. Smooth normal operation of these pumps is essential for ensuring both the production quality and quantity. It is extremely important to detect any anomaly or malfunction in this rotating equipment at an early stage. This helps to take the appropriate corrective maintenance actions and prevent any catastrophic failure, equipment down time, quality deviation and/or production loss. However, there are very few methods available in the literature for detecting faults or anomalies in the pumps, particularly for the positive displacement pumps in real industrial application using only routinely available process data -such as: flow, speed, stroke, discharge pressure etc. In this paper, a machine-learning based Early Fault Detection & Diagnostic system is developed to monitor the rotating equipment in operation, detect a fault at initiation, pinpoint the root cause, and to send out alerts for corrective maintenance with suggested remedial actions. The detection works by building a baseline machine learning model of the equipment performance under normal operating conditions which is then used to monitor the health deviation of the equipment in real time and predict a fault at a very early stage, much before it is observed by operations personnel. The proposed fault detection method relies only on routine process data – flow, speed, stroke etc. and does not require any additional measurements like vibration, motor current, acoustic emission data. The diagnostics tool identifies the most probable root causes of the failures and provides the possible failure resolution methods based on the historical maintenance records of similar equipment. The proposed algorithm combines data-driven and knowledge-based approaches. The efficacy of the proposed method was demonstrated to detect and identify incipient faults in positive displacement chemical dosing pumps in a water treatment plant. The detected and identified faults were validated using the maintenance records of the pumps.

Published in	American Journal of Artificial Intelligence (Volume 6, Issue 1)
DOI	10.11648/j.ajai.20220601.13
Page(s)	20-26
Creative Commons	This is an Open Access article, distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution and reproduction in any medium or format, provided the original work is properly cited.
Copyright	Copyright © The Author(s), 2024. Published by Science Publishing Group

Keywords

Anomaly Detection, Machine Learning, Natural Language Processing, Predictive Maintenance, Rotating Equipment, Pumps

References

[1]	R. Rayner, (1995), Pump Users Handbook, 263-281.
[2]	S. Moran. (2015), An Applied Guide to Process and Plant Design, 175-199.
[3]	Isermann, R. (2006), Fault-Diagnosis Systems: An Introduction from Fault Detection to Fault Tolerance; Springer Science & Business Media: Berlin, Germany.
[4]	Tinga, T.; Loendersloot, R. (2019), Physical Model-Based Prognostics and Health Monitoring to Enable Predictive Maintenance. In Predictive Maintenance in Dynamic Systems: Advanced Methods, Decision Support Tools, and Real-World Applications; Lughofer, E., Sayed-Mouchaweh, M., Eds.; Springer International Publishing: Cham, Switzerland, 2019; pp. 313–353.
[5]	Jiang, G.; Xie, P.; He, H.; Yan, J. (2017), Wind turbine fault detection using a denoising autoencoder with temporal information. IEEE ASME Trans. Mechatron. 2017, 23, 89–100.
[6]	Mba, D.; Rao, R. B. Development of acoustic emission technology for condition monitoring and diagnosis of rotating machines; bearings, pumps, gearboxes, engines, and rotating structures. Shock Vibr. Dig. 2006, 38, 3–16.
[7]	Alfayez, L.; Mba, D.; Dyson, G. The application of acoustic emission for detecting incipient cavitation and the best efficiency point of a 60 kW centrifugal pump: Case study. NDT E Int. 2005, 38, 354–358.
[8]	Sakthivel, N. R.; Sugumaran, V.; Babudevasenapati, S. Vibration based fault diagnosis of monoblock centrifugal pump using decision tree. Expert Syst. Appl. 2010, 37, 4040–4049.
[9]	Marton, A. I. Sánchez, S. Carlosa, S. Martorell. (2013), Application of Data Driven Methods for Condition Monitoring Maintenance, Chemical Engineering Transactions, 33: 301-306.
[10]	Zouari, R.; Sieg-Zieba, S.; Sidahmed, M. Fault detection system for centrifugal pumps using neural networks and neuro-fuzzy techniques. Surveillance 2004, 5, 11–13.
[11]	Rajakarunakaran, S.; Venkumar, P.; Devaraj, D.; Rao, K. S. P. Artificial neural network approach for fault detection in rotary system. Appl. Soft Comput. 2008, 8, 740–748.
[12]	Ahmed, M.; Baqqar, M.; Gu, F.; Ball, A. D. Fault detection and diagnosis using Principal Component Analysis of vibration data from a reciprocating compressor. In Proceedings of the 2012 UKACC International Conference on Control, Cardi, UK, 3–5 September 2012; pp. 461–466.
[13]	Xiaoxia Liang; Fang Duan 1; Ian Bennett 2 and David Mba, (2020), A Sparse Autoencoder-Based Unsupervised Scheme for Pump Fault Detection and Isolation. Applied Sciences, Applied Sciences, 2020, 10 (19).
[14]	M. Devaney, A. Ram, H. Qiu, J. Lee. (2005), Preventing failures by mining maintenance logs with case-based reasoning, 59th Meeting of the Society for Machinery Failure Prevention Technology (MFPT-59).

Cite This Article

Plain Text BibTeX RIS

APA Style

Kaushik Ghosh, Gokula Krishnan Sivaprakasam, Aparnadevi Minisankar. (2022). Next Generation of Process Monitoring and Diagnostics: Applications of AI and Machine Learning to Enable Early Equipment Fault Prediction and Diagnostics. American Journal of Artificial Intelligence, 6(1), 20-26. https://doi.org/10.11648/j.ajai.20220601.13

Copy | Download

ACS Style

Kaushik Ghosh; Gokula Krishnan Sivaprakasam; Aparnadevi Minisankar. Next Generation of Process Monitoring and Diagnostics: Applications of AI and Machine Learning to Enable Early Equipment Fault Prediction and Diagnostics. Am. J. Artif. Intell. 2022, 6(1), 20-26. doi: 10.11648/j.ajai.20220601.13

Copy | Download

AMA Style

Kaushik Ghosh, Gokula Krishnan Sivaprakasam, Aparnadevi Minisankar. Next Generation of Process Monitoring and Diagnostics: Applications of AI and Machine Learning to Enable Early Equipment Fault Prediction and Diagnostics. Am J Artif Intell. 2022;6(1):20-26. doi: 10.11648/j.ajai.20220601.13

Copy | Download

@article{10.11648/j.ajai.20220601.13,
  author = {Kaushik Ghosh and Gokula Krishnan Sivaprakasam and Aparnadevi Minisankar},
  title = {Next Generation of Process Monitoring and Diagnostics: Applications of AI and Machine Learning to Enable Early Equipment Fault Prediction and Diagnostics},
  journal = {American Journal of Artificial Intelligence},
  volume = {6},
  number = {1},
  pages = {20-26},
  doi = {10.11648/j.ajai.20220601.13},
  url = {https://doi.org/10.11648/j.ajai.20220601.13},
  eprint = {https://article.sciencepublishinggroup.com/pdf/10.11648.j.ajai.20220601.13},
  abstract = {Several rotating equipment such as – centrifugal pumps and positive displacement pumps are extensively used in Water treatment plant for producing potable water from raw water. Centrifugal pumps are required for delivering water from one unit of the plant to the others, while the positive displacement pumps are used for dosing different chemicals at the various stages of water treatment process. Smooth normal operation of these pumps is essential for ensuring both the production quality and quantity. It is extremely important to detect any anomaly or malfunction in this rotating equipment at an early stage. This helps to take the appropriate corrective maintenance actions and prevent any catastrophic failure, equipment down time, quality deviation and/or production loss. However, there are very few methods available in the literature for detecting faults or anomalies in the pumps, particularly for the positive displacement pumps in real industrial application using only routinely available process data -such as: flow, speed, stroke, discharge pressure etc. In this paper, a machine-learning based Early Fault Detection & Diagnostic system is developed to monitor the rotating equipment in operation, detect a fault at initiation, pinpoint the root cause, and to send out alerts for corrective maintenance with suggested remedial actions. The detection works by building a baseline machine learning model of the equipment performance under normal operating conditions which is then used to monitor the health deviation of the equipment in real time and predict a fault at a very early stage, much before it is observed by operations personnel. The proposed fault detection method relies only on routine process data – flow, speed, stroke etc. and does not require any additional measurements like vibration, motor current, acoustic emission data. The diagnostics tool identifies the most probable root causes of the failures and provides the possible failure resolution methods based on the historical maintenance records of similar equipment. The proposed algorithm combines data-driven and knowledge-based approaches. The efficacy of the proposed method was demonstrated to detect and identify incipient faults in positive displacement chemical dosing pumps in a water treatment plant. The detected and identified faults were validated using the maintenance records of the pumps.},
 year = {2022}
}

Copy | Download

TY  - JOUR
T1  - Next Generation of Process Monitoring and Diagnostics: Applications of AI and Machine Learning to Enable Early Equipment Fault Prediction and Diagnostics
AU  - Kaushik Ghosh
AU  - Gokula Krishnan Sivaprakasam
AU  - Aparnadevi Minisankar
Y1  - 2022/04/09
PY  - 2022
N1  - https://doi.org/10.11648/j.ajai.20220601.13
DO  - 10.11648/j.ajai.20220601.13
T2  - American Journal of Artificial Intelligence
JF  - American Journal of Artificial Intelligence
JO  - American Journal of Artificial Intelligence
SP  - 20
EP  - 26
PB  - Science Publishing Group
SN  - 2639-9733
UR  - https://doi.org/10.11648/j.ajai.20220601.13
AB  - Several rotating equipment such as – centrifugal pumps and positive displacement pumps are extensively used in Water treatment plant for producing potable water from raw water. Centrifugal pumps are required for delivering water from one unit of the plant to the others, while the positive displacement pumps are used for dosing different chemicals at the various stages of water treatment process. Smooth normal operation of these pumps is essential for ensuring both the production quality and quantity. It is extremely important to detect any anomaly or malfunction in this rotating equipment at an early stage. This helps to take the appropriate corrective maintenance actions and prevent any catastrophic failure, equipment down time, quality deviation and/or production loss. However, there are very few methods available in the literature for detecting faults or anomalies in the pumps, particularly for the positive displacement pumps in real industrial application using only routinely available process data -such as: flow, speed, stroke, discharge pressure etc. In this paper, a machine-learning based Early Fault Detection & Diagnostic system is developed to monitor the rotating equipment in operation, detect a fault at initiation, pinpoint the root cause, and to send out alerts for corrective maintenance with suggested remedial actions. The detection works by building a baseline machine learning model of the equipment performance under normal operating conditions which is then used to monitor the health deviation of the equipment in real time and predict a fault at a very early stage, much before it is observed by operations personnel. The proposed fault detection method relies only on routine process data – flow, speed, stroke etc. and does not require any additional measurements like vibration, motor current, acoustic emission data. The diagnostics tool identifies the most probable root causes of the failures and provides the possible failure resolution methods based on the historical maintenance records of similar equipment. The proposed algorithm combines data-driven and knowledge-based approaches. The efficacy of the proposed method was demonstrated to detect and identify incipient faults in positive displacement chemical dosing pumps in a water treatment plant. The detected and identified faults were validated using the maintenance records of the pumps.
VL  - 6
IS  - 1
ER  -

Copy | Download

Author Information

Kaushik Ghosh

Yokogawa Engineering Asia Pte. Ltd., Singapore
Gokula Krishnan Sivaprakasam

Yokogawa Engineering Asia Pte. Ltd., Singapore
Aparnadevi Minisankar

Yokogawa Engineering Asia Pte. Ltd., Singapore

Download PDF

Sections

Plain Text BibTeX RIS

APA Style

Kaushik Ghosh, Gokula Krishnan Sivaprakasam, Aparnadevi Minisankar. (2022). Next Generation of Process Monitoring and Diagnostics: Applications of AI and Machine Learning to Enable Early Equipment Fault Prediction and Diagnostics. American Journal of Artificial Intelligence, 6(1), 20-26. https://doi.org/10.11648/j.ajai.20220601.13

Copy | Download

ACS Style

Kaushik Ghosh; Gokula Krishnan Sivaprakasam; Aparnadevi Minisankar. Next Generation of Process Monitoring and Diagnostics: Applications of AI and Machine Learning to Enable Early Equipment Fault Prediction and Diagnostics. Am. J. Artif. Intell. 2022, 6(1), 20-26. doi: 10.11648/j.ajai.20220601.13

Copy | Download

AMA Style

Kaushik Ghosh, Gokula Krishnan Sivaprakasam, Aparnadevi Minisankar. Next Generation of Process Monitoring and Diagnostics: Applications of AI and Machine Learning to Enable Early Equipment Fault Prediction and Diagnostics. Am J Artif Intell. 2022;6(1):20-26. doi: 10.11648/j.ajai.20220601.13

Copy | Download

@article{10.11648/j.ajai.20220601.13,
  author = {Kaushik Ghosh and Gokula Krishnan Sivaprakasam and Aparnadevi Minisankar},
  title = {Next Generation of Process Monitoring and Diagnostics: Applications of AI and Machine Learning to Enable Early Equipment Fault Prediction and Diagnostics},
  journal = {American Journal of Artificial Intelligence},
  volume = {6},
  number = {1},
  pages = {20-26},
  doi = {10.11648/j.ajai.20220601.13},
  url = {https://doi.org/10.11648/j.ajai.20220601.13},
  eprint = {https://article.sciencepublishinggroup.com/pdf/10.11648.j.ajai.20220601.13},
  abstract = {Several rotating equipment such as – centrifugal pumps and positive displacement pumps are extensively used in Water treatment plant for producing potable water from raw water. Centrifugal pumps are required for delivering water from one unit of the plant to the others, while the positive displacement pumps are used for dosing different chemicals at the various stages of water treatment process. Smooth normal operation of these pumps is essential for ensuring both the production quality and quantity. It is extremely important to detect any anomaly or malfunction in this rotating equipment at an early stage. This helps to take the appropriate corrective maintenance actions and prevent any catastrophic failure, equipment down time, quality deviation and/or production loss. However, there are very few methods available in the literature for detecting faults or anomalies in the pumps, particularly for the positive displacement pumps in real industrial application using only routinely available process data -such as: flow, speed, stroke, discharge pressure etc. In this paper, a machine-learning based Early Fault Detection & Diagnostic system is developed to monitor the rotating equipment in operation, detect a fault at initiation, pinpoint the root cause, and to send out alerts for corrective maintenance with suggested remedial actions. The detection works by building a baseline machine learning model of the equipment performance under normal operating conditions which is then used to monitor the health deviation of the equipment in real time and predict a fault at a very early stage, much before it is observed by operations personnel. The proposed fault detection method relies only on routine process data – flow, speed, stroke etc. and does not require any additional measurements like vibration, motor current, acoustic emission data. The diagnostics tool identifies the most probable root causes of the failures and provides the possible failure resolution methods based on the historical maintenance records of similar equipment. The proposed algorithm combines data-driven and knowledge-based approaches. The efficacy of the proposed method was demonstrated to detect and identify incipient faults in positive displacement chemical dosing pumps in a water treatment plant. The detected and identified faults were validated using the maintenance records of the pumps.},
 year = {2022}
}

Copy | Download

TY  - JOUR
T1  - Next Generation of Process Monitoring and Diagnostics: Applications of AI and Machine Learning to Enable Early Equipment Fault Prediction and Diagnostics
AU  - Kaushik Ghosh
AU  - Gokula Krishnan Sivaprakasam
AU  - Aparnadevi Minisankar
Y1  - 2022/04/09
PY  - 2022
N1  - https://doi.org/10.11648/j.ajai.20220601.13
DO  - 10.11648/j.ajai.20220601.13
T2  - American Journal of Artificial Intelligence
JF  - American Journal of Artificial Intelligence
JO  - American Journal of Artificial Intelligence
SP  - 20
EP  - 26
PB  - Science Publishing Group
SN  - 2639-9733
UR  - https://doi.org/10.11648/j.ajai.20220601.13
AB  - Several rotating equipment such as – centrifugal pumps and positive displacement pumps are extensively used in Water treatment plant for producing potable water from raw water. Centrifugal pumps are required for delivering water from one unit of the plant to the others, while the positive displacement pumps are used for dosing different chemicals at the various stages of water treatment process. Smooth normal operation of these pumps is essential for ensuring both the production quality and quantity. It is extremely important to detect any anomaly or malfunction in this rotating equipment at an early stage. This helps to take the appropriate corrective maintenance actions and prevent any catastrophic failure, equipment down time, quality deviation and/or production loss. However, there are very few methods available in the literature for detecting faults or anomalies in the pumps, particularly for the positive displacement pumps in real industrial application using only routinely available process data -such as: flow, speed, stroke, discharge pressure etc. In this paper, a machine-learning based Early Fault Detection & Diagnostic system is developed to monitor the rotating equipment in operation, detect a fault at initiation, pinpoint the root cause, and to send out alerts for corrective maintenance with suggested remedial actions. The detection works by building a baseline machine learning model of the equipment performance under normal operating conditions which is then used to monitor the health deviation of the equipment in real time and predict a fault at a very early stage, much before it is observed by operations personnel. The proposed fault detection method relies only on routine process data – flow, speed, stroke etc. and does not require any additional measurements like vibration, motor current, acoustic emission data. The diagnostics tool identifies the most probable root causes of the failures and provides the possible failure resolution methods based on the historical maintenance records of similar equipment. The proposed algorithm combines data-driven and knowledge-based approaches. The efficacy of the proposed method was demonstrated to detect and identify incipient faults in positive displacement chemical dosing pumps in a water treatment plant. The detected and identified faults were validated using the maintenance records of the pumps.
VL  - 6
IS  - 1
ER  -

Copy | Download