A knowledge graph – based requirement identification model for products remanufacturing design

Abstract

The effective identification of remanufacturing design requirements is crucial to ensure that remanufactured products meet the required standards and demands. However, in addition to considering the performance demands of customers for products, remanufactured products should also take into account the various types and degrees of failure across diverse parts. To this end, under the dual demands, a knowledge graph – based requirement identification model is proposed to intelligently transform complex remanufacturing demands into standardised design requirements. Firstly, excavation of design information from representative cases is conducted, including failure characteristics, customer demands and design requirements (FCDR). A FCDR ontology model is proposed by establishing the mapping relationship between demands and design requirements. Secondly, the triplet of design information is integrated into the BERT-BiLSTM-CRF model to obtain the accurate entity, and the ALBERT-BiLSTM-Attention model is used to extract the FCDR relationship, so that a knowledge graph can be constructed. Thirdly, by using coupling weighting technique to consider the relevance of the dual demands, normalised failure characteristics and customer demands are accurately extracted and mapped to similar design requirement nodes in knowledge graph. Finally, a machine tool remanufacturing design is used as an example to verify the effectiveness of the proposed model.