A performance evaluation method based on combination of knowledge graph and surrogate model

To satisfy the requirements of individual design and rapid performance evaluation of complex products, this paper proposes a hybrid approach to build a performance evaluation model and perform the rapid evaluation of design schemes. This approach consists of a surrogate model and knowledge graph (KG). Firstly, the KG of complex electromechanical products is established by Web Ontology Language to provide information about parts and evaluation indexes for the sampling process. It includes building ontology and writing inference and query rules at the framework level. Secondly, based on the sample points, a dynamics model is built and used for simulation. Using the Design of Experiments, the variables that have the greatest impact are found. The relevant variables will be input into the model to obtain the data set. According to the data set, a surrogate model based on the radial basis function is built as a performance evaluation model, which can improve computing efficiency to achieve evaluation results rapidly. In this study, the bogie design is used as a test case to evaluate the proposed method. And the results show that it can improve design efficiency for design issues such as part selection.

Automated summary

This paper proposes a hybrid approach to build a performance evaluation model and perform rapid evaluation of design schemes. The approach uses a surrogate model and knowledge graph to establish the model and simulate the design process, improving computational and design efficiency.