Traditional ship design management, centered on drawings and documents, suffers from prominent issues such as discontinuous design processes, scattered data carriers and low collaboration efficiency. To advance the digital transformation of ship design management, this paper constructs a whole-process management system for ship design that takes 3D models as the sole authoritative data source and bill of materials (BOM) collaborative integration as the core. This system covers seven key application scenarios including project management, plan management, review and verification management, submission and return review management, shipyard-institute collaboration management, model data management and archive management. All scenarios feature interlinked data and closely connected processes, forming an integrated whole-process management system. Combined with practical implementation in multiple ship-type projects, the application value of this system has been verified. The new digital design management model proposed in this paper effectively addresses the problems of data silos, version inconsistency and poor collaboration in traditional design management, providing a reference for ship enterprises in the construction of digital design management systems and facilitating their intelligent and digital transformation.

Data preprocessing is a core step in knowledge graph construction, consisting of two main stages: data collection and information extraction. This paper systematically reviews mainstream data preprocessing methods based on rules and lexicons, statistical machine learning, and deep learning, and thoroughly analyzes their technical principles and application limitations in entity recognition and relation extraction. Existing methods rely heavily on manual rules and suffer from weak semantic generalization, making it difficult to achieve cross-domain knowledge transfer. To address these issues, this paper explores a novel paradigm of “semantic-driven and automated extraction” based on large language models. By generating deep semantic embeddings through pre-trained large language models and combining vector similarity computation, it enables unsupervised and context-aware information extraction, driving the intelligent transformation of knowledge graph construction. The current approach is still in the exploratory stage, facing challenges such as high computational cost and low interpretability. Future research should focus on lightweight model design, multimodal semantic alignment, and domain knowledge integration to improve the efficiency of knowledge graph construction and model interpretability.