A virtual model knowledge updating method driven by data

In the initial construction of unit-level virtual models in discrete production system, understanding physical entities and actual operating environment is lacking, need to be iteratively updated resemble more closely. To solve above problems, an updating method for knowledge model driven by data studied. Firstly, behavior can divided into basic functional such as forward kinematics, trajectory planning, material generation, components support computation are constructed. Secondly, evaluation indexes quality, maintainability, reliability established, appropriate component combinations selected from base. Then, based on measurements, it used triggering judgment, a combination with smallest difference between simulation output real multiple realize iterative update model. Finally, feasibility effectiveness verified examples.