An NLP-Based Extraction and Display Method for Augmented Maintenance Information

With the advancement of industrial intelligence, stakeholders are increasingly adopting Augmented Reality for industrial operation and maintenance. However, the extraction and display of augmented information still rely heavily on subjective expertise, leading to significant annotation discrepancies among personnel, which is labor-intensive and time-consuming. To address this, this paper proposes an NLP (Natural Language Processing)-based method for extracting and displaying augmented maintenance information. Firstly, an NLP-based maintenance information extraction method is proposed. This approach achieves automated text annotation and classification through keyword property matching and weighted-similarity algorithms. Subsequently, leveraging the extracted information, an augmented maintenance information display method is developed to plan AR device layouts. Finally, feasibility is validated via a maintenance case study of oil seepage in hollow bolts on distribution oil seats. This method successfully overcomes the subjective limitations and cost inefficiencies of manual annotation, advancing the efficient adoption of Augmented Reality assisted maintenance.