Abstract: To solve problems of lacking design knowledge in describing,decomposing and discovering functions,low availability of function implementation technology and difficulty in obtaining technical features when generating concepts of a new product,a patent data-assisted concept generation method for new product development was proposed.The text related to design tasks in the International patent classification were retrieved using text similarity algorithms to assist in describing main functions of the new product.Based on the C-K theory,the lower level of classification text was considered as K space elements,sub-functions of the new product are considered as C space elements,and main functions decomposition was realized assisted by elements transformation.Then,directions for new product function innovation were determined by combining the patent technology maturity analysis and function innovation strategies,and integration functions of the new product were predicted based on text of co-occurrence classifications,which realized function innovation of the new product.Based on the cross-field patent analysis model and the definition of field convergence,the knowledge breadth and liquidity of patent technology were quantified to prioritize function implementation technologies,and technical features in independent claims of patent were automatically extracted using semantic parsing,which were recombined to generate a conceptual design schema of the new product.Finally,the method was applied to generate a conceptual design schema in the field of heavy truck chassis pre-assembly equipment,which proved the effectiveness of the method.

Key words: concept generation, patent data, function design, patent technology optimum, semantic resolution

摘要： 在生成新产品概念设计方案的过程中,企业常缺乏辅助新产品功能描述、分解及创新的设计知识,且存在功能实现技术检索结果可用性不高、难以有效转化为概念设计方案技术特征的问题。为此,提出一种专利数据辅助的新产品概念设计方案生成方法。首先,使用文本相似度算法,检索国际专利分类表中与设计任务相关的分类号文本,辅助描述新产品的主功能。其次,引入C-K理论将下级分类号释义视为K空间中的元素,新产品子功能视为C空间中的元素,通过元素间转化辅助完成主功能分解。再次,结合专利技术成熟度分析和功能创新策略,确定新产品的功能创新方向,并依据共现分类号释义预测新产品的集成功能,辅助实现新产品的功能创新。接着,应用专利文献跨域特征分析模型,引入技术领域融合度的概念,量化专利技术的知识广度和流动性,辅助优选功能实现技术,并利用语义解析技术自动提取专利独立权利要求中的技术特征,辅助生成新产品概念设计方案。最后,将该方法应用于重型汽车底盘预装设备领域的新产品开发,证明了方法的有效性。

关键词: 概念设计, 专利数据, 功能设计, 专利技术优选, 语义解析