Dynamic knowledge graph modeling method for customized furniture panel production process

Compared with traditional batch production models, customized furniture panel production generates vast heterogeneous data throughout the order life cycle, including order processing, process design, and manufacturing execution. This data exhibits strong dynamics and complex relationships. However, conventional relational databases used to store life-cycle data suffer from high redundancy, weak connectivity, and limited scalability. To address these issues, this paper proposes a dynamic knowledge graph modeling method for customized furniture panel production. First, we outline the production workflow and define node attributes and relationships. Next, we develop an ontology-based knowledge model using enterprise production data. Then, we design a knowledge graph construction algorithm leveraging an ALBERT–BiLSTM–Attention–CRF entity extraction model. Finally, we validate our approach through a case study at a customized furniture factory, demonstrating improved knowledge management and decision-making efficiency. The comprehensive knowledge ontology for customized furniture panel production processes constructed in this study comprises 413,251 nodes and encompasses 10 distinct relationship types. This demonstrates the effectiveness of the proposed approach in enhancing knowledge management and decision-making efficiency. This work holds significant implications for advancing the development and application of knowledge graph technologies within the smart manufacturing domain, particularly by offering new perspectives and solutions for the intelligent optimization of manufacturing workflows.