A virtual model knowledge updating method driven by data

Abstract

The initial construction of the unit-level virtual model in the discrete production line lacks the understanding of physical entities and the actual operating environment, which makes it difficult for the virtual model knowledge to correctly drive the behavior of physical entities. In order to solve the above problems, a virtual model knowledge updating method driven by data is studied. Firstly, the knowledge in the virtual model is divided into basic functional units according to functions, and the knowledge of each functional unit is encapsulated into knowledge components such as forward kinematics, motion control, material generation, etc., and the knowledge component base is constructed. Secondly, this method establishes the evaluation index of effectiveness, usability, and reliability of knowledge components, and can screen out appropriate groups of knowledge component combinations from the knowledge component base. Then, based on the measurement value, it is used to update the trigger judgment and select a group of knowledge components with the smallest difference between the simulation output of virtual model and the real data of physical entities from multiple groups of knowledge combinations, so as to realize the iterative update of the model. Finally, the feasibility and effectiveness of the method are verified by examples.