A knowledge-based system for quality analysis in model-based design

Abstract

The quality of model-based definition (MBD) models is one of the significant challenges to achieving model-based enterprise (MBE). Different stakeholders in MBE put forward different quality requirements for MBD models, while any quality defect may hinder the production. Therefore, it is essential to analyze the quality of MBD models adequately before the usage. However, existing quality tools are difficult to apply in MBE. A knowledge-based quality analysis system for MBD part models is presented in this paper. The system utilizes model quality knowledge to analyze model quality from the perspective of various stakeholders in MBE. And it can be extended by adding new MBD part model quality knowledge. In order to judge model quality, a concept of the model definition unit is introduced for converting model information into the description towards model semantics quality, and a quality analysis method based on this description is proposed. For the ease of model quality knowledge representation and maintenance, a knowledge representation framework based on the object chain and parameter table is developed. Finally, a prototype system is implemented to verify the effectiveness and practicality. With 46 knowledge items derived from eight model-used stages, in total 24 potential quality defects are found in two case models. And additional two defects are discovered by extending the system with new model quality knowledge. In this way, the model quality is improved from the perspective of corresponding users.