| 摘要: |
| 目的 为了解决复杂装备零件装配中效率和精度低的问题，开发一种基于数字孪生技术与改进动态目标检测算法的复杂高端装备动态交互装配方法。方法 构建面向数字孪生的人机交互框架并优化目标检测算法，实现复杂装备装配过程的实时监控。操作人员通过可视化界面与系统进行动态交互，实现装配工艺的实时优化。首先，通过多源传感器实时采集装备的实时数据;其次，基于实时数据构建装备虚拟模型，生成与实际装配场景高度一致的视频并通过改进后的YOLOv8n算法实现对工业零件的精准检测与分类;最后，通过人机交互过程对装配过程进行动态优化并通过设计交互界面实时显示检测结果，为装备装配提供智能化支持。结果 提出方法能够在复杂高端装备的装配过程中实现精准的检测与实时动态响应。改进后动态目标检测结果在保持高精度的同时，较未改进版本检测速度提升24.9%。结论 不仅为复杂高端装备人机交互装配提供了创新解决方案，而且为数字孪生技术在工业领域的进一步应用与发展提供了有力支持。 |
| 关键词: 数字孪生 人机交互 工业装配 目标检测 YOLOv8n |
| DOI：10.19554/j.cnki.1001-3563.2025.08.005 |
| 分类号: |
| 基金项目:浙江省重点研发计划(2025C01023，2024C01029);贵州大学开放基金(No. PBD2024-0515) |
|
| Dynamic Interactive Assembly Method for Complex High-end Equipment Based on Digital Twin |
|
MA Xubeng1, FENG Yixiong1,2, CHENG Dinghao2
|
|
(1. College of Mechanical Engineering, Guizhou University, Guiyang 550025, China;2. State Key Laboratory of Fluid Power and Mechatronic Systems, Zhejiang University, Hangzhou 310027, China)
|
| Abstract: |
| The work aims to develop a dynamic interactive assembly method for complex high-end equipment based on digital twin technology and an improved dynamic target detection algorithm to address the issues of low efficiency and poor precision in the assembly of complex equipment parts. A human-machine interaction framework towards digital twins was constructed and the target detection algorithm was optimized to achieve real-time monitoring of the complex equipment assembly process. Operators could interact dynamically with the system through a visual interface, thereby achieving real-time optimization of the assembly process. Initially, real-time data of the equipment was collected through multi-source sensors. Subsequently, a virtual model of the equipment was constructed based on real-time data, generating videos highly consistent with the actual assembly scene, and precise detection and classification of industrial parts were achieved through the optimized YOLOv8n algorithm. Then, the assembly process was dynamically optimized through the human-machine interaction process, and the detection results were displayed in real-time through the designed interactive interface, providing intelligent support for equipment assembly. The method proposed in this study could achieve precise detection and real-time dynamic response during the assembly process of complex high-end equipment. The improved dynamic target detection results not only maintained high precision but also increased the detection speed by 24.9% compared with the unimproved version. This study not only provides an innovative solution for human-machine interactive assembly of complex high-end equipment but also offers strong support for the further application and development of digital twin technology in the industrial field. |
| Key words: digital twin human-machine interaction industrial assembly object detection YOLOv8n |
