基于数字孪生的智能产品模块管控交互仿真

Interactive Simulation of Intelligent Product Module Management and Control Based on Digital Twin

在信息技术赋予智能产品智能特性的前提下,为了通过智能产品的管控提升工作过程的整体性能,提出了基于数字孪生的智能产品模块管控交互方法。针对智能产品模块具有的物理域和信息域两种特性,将智能产品分为感知模块、交互模块、推理模块、存储模块和执行模块。采用公理化设计理论对智能产品管控指标体系进行构建,整个智能产品实时数据模型主要分为孪生体模型数据和单元工作实时数据两部分,通过自动识别技术对智能模块资源进行标识和采集。采用数字孪生模型及其同步的方法,完成对智能产品的虚拟映射,通过离散化处理将智能产品动力学模型转换为自回归平滑模型,实现对智能产品模块的数学孪生模型描述,由于数字孪生模型参数极易退化,通过递推增广最小二乘算法对数字孪生模型进行优化,提高数字孪生模型的管控精度。实验结果表明,数字孪生模型数据同步方法具有良好的收敛性,系统调整时间小于0.1s,且阶跃响应曲线与设计的阶跃响应曲线拟合程度较好,充分说明数字孪生模型控制策略的有效性。

Under the premise of intelligent characteristics of intelligent products given by information technology,in order to improve the overall performance of the work process through the control of intelligent products, a control and interaction method based on digital twins is proposed. Firstly, according to the physical domain and information domain characteristics of the intelligent product module, the intelligent product was divided into perception module,interaction module, reasoning module, storage module and execution module. Then the axiomatic design theory was used to build the intelligent product management and control index system. The whole intelligent product real-time data model was mainly divided into two parts: twin model data and unit work real-time data. The intelligent module resources were identified and collected by automatic identification technology. Finally, the digital twin model and its synchronization method were used to complete the virtual mapping of the intelligent product, and the dynamic model of the intelligent product was transformed into the autoregressive smoothing model through discretization processing to realize the mathematical twin model description of the intelligent product module. Because the parameters of the digital twin model are easy to degenerate, the recursive augmented least square algorithm was used to optimize the digital twin model to improve the control accuracy of the digital twin model. The experimental results show that the digital twin model data synchronization method has good convergence, the system adjustment time is less than 0. 1s,and the step response curve fits well with the designed step response curve, which fully shows the effectiveness of the digital twin model control strategy.