A scoping review on multi-fault diagnosis of industrial rotating machines using multi-sensor data fusion

• Published in: The Artificial intelligence review Vol. 56; no. 5; pp. 4711 - 4764
• Main Authors: Gawde, Shreyas, Patil, Shruti, Kumar, Satish, Kotecha, Ketan
• Format: Journal Article
• Language: English
• Published: Dordrecht Springer Netherlands 01.05.2023; Springer; Springer Nature B.V
• Subjects: Algorithms; Artificial Intelligence; Bibliometrics; Citation analysis; Computer Science; Data integration; Deep learning; Fault diagnosis; Forecasts and trends; Industry 4.0; Literature reviews; Machine learning; Machinery; Magneto-electric machines; Maintenance and repair; Manufacturing; Multisensor fusion; Network analysis; Predictive maintenance; Qualitative analysis; Rotating machinery; Rotating machines; Sensors; Fault diagnosis; Industrial rotating machines; ProKnow-C; Industry 4.0; Bibliometric analysis; Artificial intelligence; Predictive maintenance
• ISSN: 0269-2821; 1573-7462
• DOI: 10.1007/s10462-022-10243-z

Rotating machines is an essential part of any manufacturing industry. The sudden breakdown of such machines due to improper maintenance can also lead to the industries' shutdown. The era of the 4th industrial revolution is taking its major shape concerning maintenance strategies, notable being in predictive maintenance. Fault prediction and diagnosis is the major concern in predictive maintenance as this is the major issue faced by all the maintenance engineers. Most of the bibliometric literature review studies that are accessible focus on fault diagnosis in rotating machines, mainly focusing on a single type of fault. However, there isn't a thorough analysis of the literature that focuses on the "multi-fault diagnosis using multi-sensor data" aspect of rotating machines. In this regard, this paper reviews the literature on the “multi-Fault diagnosis using multi-sensor data fusion” of Industrial Rotating Machines employing Machine learning/Deep learning techniques. A hybrid bibliometric approach was used to analyze articles from the “Web of Science” and “Scopus” Database for the last 10 years. The method for literature analysis used, is quantitative as well as qualitative, as not only the traditional approach (bibliometric and network analysis) but also a novel method named ProKnow-C is used, and it entails a number of phases, that includes intelligent and extensive filtering from the large set of results and finally selecting the articles that are more pertinent to the research theme. Based on available publications, an analysis is performed on year-by-year publication data, article types, linguistic distribution of articles, funding sponsors, affiliations, citation analysis and the relationship between keywords, authors, etc. to provide an in-depth vision of research trends in the related area. The paper also focuses on the maintenance strategies, predictive maintenance approaches, AI algorithms, Multi sensor data fusion, challenges, and future directions in “multi-fault diagnosis using multi-sensor data fusion” in rotating machines. The foundational work done in the field, the most prolific papers and the key research themes within the research area are all identified in this bibliometric survey.