【作者】 王太勇； 郑明良；

【Author】 ZHENG Mingliang;School of Mechanical Engineering,Tianjin University;

【机构】 天津大学机械工程学院；

【摘要】 在现有数字孪生车间状态数据实时的基础上,针对现有车间调度方法自适应能力差的问题,提出了一种基于实时数据的孪生车间动态调度体系架构。通过对物理车间加工过程资源信息进行调研,搭建高保真的车间数字孪生模型,基于面向加工过程的虚实车间数据交互方法构建数字孪生车间系统。在此基础上,详细阐述了面向加工过程的车间数字孪生模型构建、基于实时数据的孪生车间加工过程异常事件监测和虚实交互、基于深度强化学习的孪生车间加工路线动态预测等关键技术。设计了原型系统,并通过二级减速箱实验车间验证了其可行性,为数字孪生技术在制造车间加工路线规划方面的应用提供了关键支撑。更多还原

【Abstract】 On the basis of the state data perception of the existing digital twin workshop,in order to solve the problem of poor adaptive ability of workshop processing route planning,a dynamic prediction system architecture of processing route based on perception data is proposed.Build a high-fidelity workshop digital twin model by investigating the processing process of the physical workshop,and establish a virtual-real workshop data interaction method oriented to the processing process to construct a digital twin workshop system.On this basis,key technologies such as the construction of a workshop digital twin model for the machining process,the monitoring of abnormal events in the twin workshop based on perception data and the interaction between virtual and reality,and the dynamic prediction of the twin workshop’s processing route based on deep reinforcement learning are elaborated.The prototype system was designed,and its feasibility was verified through the secondary reduction gear box production line,which provided key support for the application of digital twin technology in manufacturing workshop processing route planning.更多还原