Knowledge Acquisition and Reasoning Model for Welding Information Integration Based on CNN and Knowledge Graph

Knowledge acquisition and reasoning are essential in intelligent welding decisions. However, the challenges of unstructured knowledge weak linkage across phases limit development intelligence, especially integration domain information engineering. This paper proposes a cognitive model combining image recognition graph. A CNN is used as perception layer to obtain direct information. Automated logic rules based on graph described enable domain. In addition, bogie frame was constructed entity relationship recognition. models with different network structures were compared trained under supervised conditions. results, InceptionV1 obtained high score (0.758 for thickness relation, 0.642 groove form, 0.704 joint type, 0.835 base material form). The proposed showed positive performance terms accuracy, interpretation, coverage, scalability, portability several other methods. can effectively address abovementioned limitations important manufacturing engineering integration.