Header menu link for other important links
X
Application of machine learning algorithm for anomaly detection for industrial pumps
N. Dutta, , U. Subramaniam
Published in Springer
2021
Volume: 907
   
Pages: 237 - 263
Abstract
Automation technology has brought a pragmatic change in the field of industrial sector, commerce and agricultural sector etc. where machine learning algorithm is one of the pioneers of this. Machine learning has a broad scale application among that anomaly detection is one of the applications. Industrial pumps are essential parts of any kind of industry which requires proper maintenance which is recognized as condition monitoring. The application of machine learning in industrial sectors is going to 65% up to 2018. More will increase in later future. Condition monitoring is the process which cannot prevent failure but can predict the possibility of failure, fault condition by measuring certain machine parameters. If machine learning algorithm can be implemented then the system will be more efficient and it is possible to detect the problems in the ground level which can help to increase the lifespan of the pumping system. Various machine learning algorithms are available among which most of the cases classification and regression analysis are used to detect the anomalies. When there is discrete system generally classification is used and when continuous function is there regression is used. There are more other machine learning algorithms which can analyze the anomalies in the system with the help of predictive control model. The predictive control hybrid model is the new socket of study where the researchers can forecast to shrink the vigor loss of resource and time and can make the system flawless. Thus, there is a big challenge before the researchers regarding the application of machine learning for detecting the anomalies in the pumping system. © 2021, The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG.
About the journal
JournalData powered by TypesetStudies in Computational Intelligence
PublisherData powered by TypesetSpringer
ISSN1860949X