A Framework of Cognitive Intelligence-Enabled Welding Cyber Physical System

Building a welding cyber physical system (WCPS) is an important way to realize intelligent welding. However, the interaction between the current welding cyber world and the physical world is insufficient. On the one hand, it is reflected in the lack of cognition of the welding process that contains quality information. On the other hand, the interaction efficiency based on manual experience is low. A cognitive intelligence-enabled WCPS is proposed. Firstly, the welding knowledge graph is constructed to model and manage the multi-source data, so as to improve the interaction efficiency between welding cyber space and physical space. Secondly, the cognitive capability of the WCPS is improved through perceptual and cognitive computing. Finally, the welding action formed by each knowledge node is optimized based on reinforcement learning to improve the interactive capability of the WCPS.