基于知识的复杂产品装配工艺快速编制方法

Aiming at the problems of low preparation efficiency, high quality fluctuation and low automation and intelligence in the process of complex product assembly process design, a fast preparation method for assembly process with semantic-oriented knowledge instance retrieval and reuse is proposed. A two-layer assembly process knowledge model based on the knowledge retrieval layer and the knowledge instance layer is established by analyzing the assembly process of complex product, and the knowledge is classified in terms of content, type, function and application direction. On this basis, a process document semantic analysis method based on the Sentence-BERT sentence embeddings model is proposed for the existing knowledge instances. The semantic retrieval method of process knowledge is given combined with the cosine similarity algorithm, and the rapid reuse of knowledge in the design process of assembly process is realized. Taking aerospace products as an example, a rapid design management system of aerospace product assembly process is developed, and a knowledge base for the assembly process is constructed and put into operation in an aerospace enterprise to verify the feasibility of the proposed method, resulting in a significant increase in process preparation efficiency.