Bibliometric analysis of digital twin literature: a review of influencing factors and conceptual structure

Abstract

ABSTRACT This article systematically summarises the digital twin technology in the existing scientific field using the bibliometric method, puts forward the critical areas and research directions at different stages of the life cycle of the digital twin research topic. Using R-Studio ‘Biblioshiny’ to measure scientifically, we visually analyze 514 articles related to digital twin in the Web of Science's core set from 2014 to 2021 (including SCI-Expanded and SSCI database). Firstly, we analyzed the core journals, articles, authors, institutions, relevant countries to determine the vital influence of digital twin-related literature. Secondly, we use theme maps to draw the research topics of digital twin-related literature as icons on the conceptual structure and offer four cluster parts (Motor themes, highly developed and isolated themes, Emerging themes, and Basic themes). Finally, we concluded by summarising the limitations of our study and subsequently highlighted future research directions. The study shows that monitoring, maintenance, and computer vision are research topics with good development momentum. Disturbance observer, reinforcement learning, and friction are relatively small research topics with a good development foundation, and scheduling, process modelling, and deep reinforcement learning are emerging research directions.