Construction of Knowledge Graph of Glass Defects

The glass industry constitutes a crucial component of China's industrial framework. However, challenges persist in systematizing professional knowledge due to the extensive variety of glass products, intricate manufacturing processes, and fragmented expertise. Knowledge graph construction technology offers a solution by structurally integrating critical information—including glass classifications, production workflows, defect characteristics, and their underlying causes—thereby establishing a comprehensive glass knowledge network. This study employs a hybrid approach combining bottom-up and top-down methodologies to construct the knowledge graph across two dimensions: Schema Layer Defines 8 core entity types and 16 relational frameworks based on industry standards. Data Layer Implements a purpose-built relation extraction model to dynamically mine entity and relationship data.