Abstract: The diverse range of airborne products and the extensive industrial chain involve a complex testing system, necessitating comprehensive metrological work. However, airborne product data resources primarily exist in unstructured, fragmented, and multimodal forms, making it difficult to conduct holistic analysis of various testing elements or trace the standardization of testing and product quality under a unified framework, thereby posing challenges to metrological tasks.To address this issue, the construction of knowledge graph for Metrological Traceability of Airborne Products (MT-AP) was explored by leveraging generative Large Language Models (LLMs). First, the resource types and metrological traceability process were systematically categorized, and an ontological model for the Knowledge Graph (KG) was constructed. Second, LLMs-based workflow modules were designed and integrated into workflow chains, a method for constructing the MT-AP knowledge graph based on these chains and prompt templates was proposed. Experiments were conducted using airborne product case data and workflow chains. The knowledge comprehension and naming capability scored above 0.91, while the text segmentation and knowledge decoupling capability scored above 0.83. The complex parameter extraction and structuring capability maintained scores above 0.85. The results demonstrate that the proposed method exhibits satisfactory performance in key tasks of MT-AP knowledge graph construction, providing technical support for metrological engineering of airborne products.

Key words: Large Language Model (LLM), multi-source heterogeneous data, airborne product, metrological traceability, knowledge graph

摘要： 机载产品类型多样、产业链广泛，存在复杂的检测体系，需要开展完备的计量检定工作。然而，机载产品数据资源主要以非结构化、碎片化和多模态的形式存在，不利于对各类测试要素开展统筹分析，难以在统一框架下溯源测试规范性与产品质量，给计量工作带来了障碍。面对这一问题，结合生成式大语言模型（LLMs）开展机载产品计量溯源（MT-AP）知识图谱构建研究。首先，梳理资源类型和计量溯源环节，构建知识图谱（KG）本体模型；其次，设计LLMs工作模块并组合成工作链，提出基于工作链与提示模板的MT-AP知识图谱的构建方法。结合机载产品实例数据和多个工作链进行验证实验，其中知识理解与命名能力评分基本在0.91以上，文本拆分与知识解耦能力评分基本在0.83以上，复杂参数的知识提取与结构化能力评分基本在0.85以上。实验结果表明，所提方法在MT-AP知识图谱构建关键任务上体现了较好的工作性能，为机载产品计量工程提供技术支撑。

关键词: 大语言模型, 多源异构数据, 机载产品, 计量溯源, 知识图谱