A Bibliometric Review and Analysis of Data-Driven Fault Detection and Diagnosis Methods for Process Systems

Accident prevention is one of the most desired and challenging goals in process industries. For accident prevention, fault detection and diagnosis (FDD) is critical. FDD has been an active area of research for decades. The focus of the current review is on the data-driven techniques as we are now in a digital era and data analytics is getting more emphasis in all areas including process industries. The analysis is done to address the following fundamental questions: (i) How are the leading areas evolving? (ii) Who are the contributing authors? (iii) What are the key sources and domains of publications? (iv) Which countries are active in this research area? Furthermore, we briefly described four techniques, principal component analysis, partial least-squares, independent component analysis, and the Gaussian mixture model, to represent the state-of-the-art algorithms from different periods. It was observed that significant work in this field is being carried out throughout the world, including both developed and the developing countries. China is emerging as the leading contributor to the total number of publications while Singapore is the country with the highest per-capita publication. Finally, the link between different types of publications, especially between the engineering journals and the industrial journals, is growing. This indicates that these techniques are gaining industrial importance. It can be concluded that the data-based process monitoring is developing rapidly and being applied in process industries; nevertheless, the pace of application in the process industries is not at par with the pace of theoretical development.