Computer Integrated Manufacturing System ›› 2023, Vol. 29 ›› Issue (6): 1811-1823.DOI: 10.13196/j.cims.2023.06.003

Previous Articles     Next Articles

Distributed rendering method of digital twin model for complex production scene

YANG Yanfang1,2,GAO Jujian1,WANG Qi1,SHU Liang2+   

  1. 1.School of Transportation and Logistics Engineering,Wuhan University of Technology
    2.Low Voltage Apparatus Technology Research Center of Zhejiang Province,Wenzhou University
    3.State Key Laboratory of Maritime Technology and Safety,Wuhan University of Technology
  • Online:2023-06-30 Published:2023-07-10
  • Supported by:
    Project supported by the Key Research and Development Program of Zhejiang Province,China(No.2021C01046),and the Key Industrial Project of Wenzhou Science and Technology Major Project,China(No.ZG2020049).

面向复杂生产场景的数字孪生模型分布式渲染方法

杨艳芳1,2,高居建1,王奇1,舒亮2+   

  1. 1.武汉理工大学交通与物流工程学院
    2.温州大学浙江省低压电器工程技术研究中心
    3.武汉理工大学水路交通控制全国重点实验室
  • 基金资助:
    浙江省重点研发计划资助项目(2021C01046);温州市重大科技专项重点工业资助项目(ZG2020049)。

Abstract: In view of the problems that the large number of digital twin scene models,complex logical relationship,many motion model states,large dynamic information in production line digital twin scene,which led to the single-machine rendering scheme was difficult to meet the real-time requirements of logical calculation and rendering output,a distributed rendering method with parallel processing of logic computing and real-time rendering was proposed.The scene data was divided into logical data and real-time rendering data,the logical data was distributed in the control port and the rendering data was distributed in the rendering port,which realized multi-machine parallel rendering.Mealwhile,synchronizing the behavior state of the dynamic model at each rendering port was taken as the entry point to design a multi-threading mechanism of Graphics Processing Unit (GPU),and the calculation of logical data and the synchronous determination of moving entities was realized.Then the synchronized data package including workshop state data,observation and projection matrix was constructed,and the fast and accurate synchronization of motion models was realized by combining the master-slave synchronized data transfer and parsing mechanism.The validation results showed that the distributed rendering method could increase the running frame rate from 22.1 to 66.6 compared with single-computer rendering under the same model scale scene,and the proposed method significantly increased the number of rendered frames in different model sizes,which was a significant improvement in rendering effect.

Key words: distributed rendering, twin model, multi-thread computing, dynamic synchronization

摘要: 针对生产线数字孪生场景模型数量庞大、逻辑关系复杂、运动模型状态多、动态信息量大,单机渲染方案难以满足逻辑计算与渲染输出的实时性要求,提出一种逻辑计算与实时渲染并行处理的分布式渲染方法。将场景数据分为逻辑数据和实时渲染数据,逻辑数据分布在控制端,渲染数据分布在渲染端,实现了多机并行渲染。以同步各渲染端动态模型的行为状态为切入点,设计了图形处理器多线程机制,实现了逻辑数据计算与运动实体同步判定计算;构建了包括车间状态数据、观察与投影矩阵等在内的同步数据包,结合主从同步数据传输与解析机制,实现了运动模型的快速精准同步。验证表明,在相同模型规模场景下,分布式渲染方法相比单机渲染可以将运行帧数由22.1提升至66.6,在不同模型规模下显著增加渲染帧数,明显提升渲染效果。

关键词: 分布式渲染, 孪生模型, 多线程计算, 动态同步

CLC Number: