A modeling method of complex assembly based on digital twin

Assembly quality as a key indicator of the performance of complex products requires the development of a predictive product model that integrates part and process information. Furthermore, deriving a high-quality assembly model is important for accurate assembly modeling and simulation. At present, assembly modeling mainly considers the geometric errors of the parts induced by manufacturing processes, and ignores a number of parameters in the real environment, which leads to inconsistencies when assembling the parts, and difficulties in meeting the performance requirements of high-precision products. Therefore, to further improve the accuracy of assembly simulation, digital technologies are extensively used to reproduce the assembly process of complex products digitally and guide the actual production and assembly of parts. In this paper, by collecting and processing real data of complex products, a hierarchical model of a complex assembly including geometric shape, position constraint and physical state is constructed, and the corresponding reverse engineering model is built. Digital twins are thereafter constructed and mapped to the real parts. Based on this, real-time data of the resource layer, the process layer and the environment layer in the assembly process are integrated, and a digital representation of complex assembly is proposed. A case study of a bearing seat highlights the applicability of the proposed approach. The data collected from the scanned bearing seat integrates the constraint information of the assembly features and the bearing seat's physical information to construct a digital twin of the bearing seat.