摘要
针对样机设计领域存在的多源异构知识分散、重用效率低以及研发过程智能化水平不足的问题,本研究提出基于知识图谱与智能推理的辅助设计系统。通过构建面向样机设计领域的专业知识图谱,实现了对设计方案、工艺文档等结构化与非结构化数据的语义化整合。本文的创新性在于提出了面向样机设计领域的专用辅助系统,采用知识图谱驱动模式实现设计要素的智能关联与可视化;同时开发了知识更新模块,实现知识的动态更新迭代;提出基于用户交互行为的动态推荐机制,并且融合自然语言理解(Natural Language Understanding, NLU)与大语言模型技术(Large Language Model,LLM),构建了面向样机研制全生命周期的智能问答模块,有效提升了复杂技术问题的解决效率。文章系统阐述了领域知识图谱构建方法、系统架构设计以及推理推荐算法实现路径,并通过典型装备样机的研制过程验证了系统在知识复用率提升和设计周期缩短等方面的应用价值。研究成果为装备制造企业实现知识驱动的智能化设计提供了可复用的技术框架与实践参考。
To address the challenges of dispersed multi-source heterogeneous knowledge,low reuse efficiency,and insufficient intelligence in the prototype design domain,this study proposes a knowledge graph and intelligent reasoning-based design assistance system.By constructing a domain-specific knowledge graph for prototype design,semantic integration of structured and unstructured data such as design schemes and process documentation is achieved.The innovations include:(1)a dedicated prototype design-oriented assistance system that implements intelligent association and visualization of design elements through knowledge graph-driven methods;(2)a knowledge update module enabling dynamic iterative updates of domain knowledge;(3)a dynamic recommendation mechanism based on user interaction behaviors,combined with natural language understanding and large language model technologies to develop an intelligent Q&A module covering the entire prototype development lifecycle,significantly improving the resolution efficiency of complex technical issues.This paper systematically elaborates the domain knowledge graph construction methodology,system architecture design,and implementation approaches for reasoningrecommendation algorithms.The system’s effectiveness in enhancing knowledge reuse rates and shortening design cycles is validated through the development process of a typical equipment prototype.The research provides a reusable technical framework and practical references for knowledge-driven intelligent design transformation in equipment manufacturing enterprises.
作者
朱晓俊
卜凡艳
孙靖
刘洁
梁论飞
ZHU Xiao-jun;BU Fan-yan;SUN Jing;LIU Jie;LIANG Lun-fei(Jianghuai Advance Technology Center,Hefei,Anhui 230088,China;Anhui Provincial Key Laboratory of Humanoid Robots,Hefei,Anhui 230088,China;Anhui Provincial Industry Innovation Center of Humanoid Robot,Hefei,Anhui 230088,China)
出处
《电子技术与软件工程》
2025年第3期26-36,共11页
Electronic Technology & Software Engineering
关键词
知识图谱
知识推理
智能推荐
问答系统
辅助设计
knowledge graph
knowledge reasoning
intelligent recommendation
question-answering system
computer-aided design
