面向数字孪生的动态数据驱动建模与仿真方法

信息物理系统(cyber-physical system, CPS)是支撑信息化和工业化深度融合的综合技术体系,传统的建模与仿真方法难以满足其应用需求。因此,提出面向数字孪生的动态数据驱动建模与仿真方法,通过随机有限集对CPS中的物理实体和传感器进行数据驱动建模,并使用基于贝叶斯推理的预测与校正过程支持数据驱动的仿真模型运行。该方法能够很好地实现数字孪生机制下虚实结合的仿真运行,并实现数据驱动的CPS仿真模型解算,通过以实利虚的方式有效提高了仿真结果的准确性和可信性。最后,通过水面舰艇作战应用案例,验证了该方法的可行性和有效性。

Dynamic data driven modeling and simulation method for digital twin

Cyber-physical system (CPS) is a comprehensive technical system that supports the deep integration of information and industrialization. The classical modeling and simulation method fail to satisfy the application requirements of the CPS. Therefore, a digital twin oriented dynamic data driven modeling and simulation method is proposed, and a random finite set (RFS) is used to build the data driven simulation model of physical entities and sensors of CPS. The Bayesian inference based prediction and update method is proposed to run the CPS model. The proposed method can implement the virtual and real combined simulation running for digital twin very well. It can also enable the data driven model running, and improve the accuracy and credibility of simulation results by using the data from the real space. Finally, the feasibility and validity of the proposed method is certified by using the operational application cases of the surface ships.