GENERATION OF THE DIGITAL TWIN OF A PROCESS PLANT BASED ON DEEP LEARNING

Abstract

Nowadays, Digital Twin is used both as a subject of scientific research as well as a buzzword to attract existing commercial products and services. It poses one of key innovations in the IT related products and services in the past years. There are numerous types of Digital Twin, in particular in the manufacturing domain. Big challenge still remains how to generate the Digital Twin for existing objects in a plant. In this paper, an approach for the generation of Digital Twin in a built environment based on scan and object recognition is presented which is commercialized as an extension of a well-running portal OpenDESC.com. After the theoretical foundation and the practical requirements are described, the functional approach and the practical implementation are explored. The examples from manufacturing plant and process plant are presented. The method for assessment of results is discussed. Finally, conclusions and outlook summarize the insights and give future directions.