A digital twin-based big data virtual and real fusion learning reference framework supported by industrial internet towards smart manufacturing

Abstract

Digital twin takes Industrial Internet as a carrier deeply coordinating and integrating virtual spaces with physical spaces, which effectively promotes smart factory development. Digital twin-based big data learning and analysis (BDLA) deepens virtual and real fusion, interaction and closed-loop iterative optimization in smart factories. This paper proposes a digital twin-based big data virtual and real fusion (DT-BDVRL) reference framework supported by Industrial Internet towards smart manufacturing. The reference framework is synthetically designed from three perspectives. The first one is an overall framework of DT-BDVRL supported by Industrial Internet. The second one is the establishment method and flow of BDLA models based on digital twin. The final one is digital thread of DT-BDVRL in virtual and real fusion analysis, iteration and closed-loop feedback in product full life cycle processes. For different virtual scenes, iterative optimization and verification methods and processes of BDLA models in virtual spaces are established. Moreover, the BDLA results can drive digital twin running in virtual spaces. By this, the BDLA results can be validated iteratively multiple times in virtual spaces. At same time, the BDLA results that run in virtual spaces are synchronized and executed in physical spaces through Industrial Internet platforms, effectively improving the physical execution effect of BDLA models. Finally, the above contents were applied and verified in the actual production case study of power switchgear equipment.