Knowledge driven multiview bill of material reconfiguration for complex products in the digital twin workshop

For the problem of long modeling time and large workload of multiview bill of material (XBOM) reconstruction process in digital twin shop, this paper proposes a knowledge-driven XBOM reconstruction method for complex products. Through the research of knowledge base construction, the modeling and simulation analysis of XBOM reconstruction process in digital twin workshop are supported, so as to shorten the cycle and improve efficiency and quality. Taking the maintenance history data of the typical representative electric multiple unit (EMU) bogies in complex products as the research object, the bidirectional long short-term memory neural network with conditional random field (BiLSTM-CRF) algorithm is used to complete the entity recognition of maintenance BOM (WBOM) reconstructed parts. Finally, taking the XBOM reconstruction process of the bogie of an enterprise as an example, the XBOM reconstruction knowledge base interaction system of the EMU bogie is built. It verifies the feasibility of the method proposed in this paper, and provides knowledge support for the XBOM reconstruction process of complex products in the digital twin workshop.