Knowledge graph and function block based Digital Twin modeling for robotic machining of large-scale components

Xuexin Zhang,Lianyu Zheng,Wei Fan,Wei Ji,Lingjun Mao,Lihui Wang

DOI: https://doi.org/10.1016/j.rcim.2023.102609

IF: 10.103

2023-07-14

Robotics and Computer-Integrated Manufacturing

Abstract:Robotic machining is a potential method for machining large-scale components (LSCs) due to its low cost and high flexibility. However, the low stiffness of robots and complex machining process of LSCs result in a lack of alignment between the physical process and digital models, making it difficult to realize the robotic machining of LSCs. The recent Digital Twin (DT) concept shows potential in terms of representing and modeling physical processes. Therefore, this study proposes a robotic machining DT for LSCs. However, the current DT is not capable of knowledge representation, multi-source data integration, optimization algorithm implementation, and real-time control. To address these issues, Knowledge Graph (KG) and Function Block (FB) are employed in the proposed robotic machining DT. Here, robotic machining related information, such as the machining parameters and errors, is represented in the virtual space by building the KG, whereas the FBs are responsible for integrating and applying the algorithms for process execution and optimization based on real-world events. Moreover, a novel adaptive process adjustment strategy is proposed to improve the efficiency of the process execution. Finally, a prototype system of the robotic machining DT is developed and validated by an experiment on robotic milling of the assembly interface for an LSC. The results demonstrate that the robotic machining is successfully optimized and improved by the proposed method.

robotics,computer science, interdisciplinary applications,engineering, manufacturing

What problem does this paper attempt to address?

The problem that this paper attempts to solve is the low - precision and low - efficiency problems existing in the robotic machining process of large - scale components (LSCs). Specifically, due to the low stiffness of industrial robots and the complexity of the LSCs machining process, there is a lack of consistency between the actual machining process and the digital model, which makes it difficult to realize the robotic machining of LSCs. Although the Digital Twin (DT) concept shows potential in representing and modeling physical processes, the current DT does not yet have the capabilities of knowledge representation, multi - source data integration, optimization algorithm implementation, and real - time control. Therefore, this research proposes a DT model for robotic machining based on Knowledge Graph (KG) and Function Block (FB), aiming to solve the above problems in the following ways:

1. **Knowledge Graph (KG)**: It is used to represent all information related to robotic machining in the virtual space, such as machining parameters and errors, so as to realize the knowledge representation of the machining process.
2. **Function Block (FB)**: It is responsible for integrating and applying algorithms according to real - world events to support the execution and optimization of the machining process.
3. **Adaptive process adjustment strategy**: A new adaptive process adjustment strategy is proposed to improve the execution efficiency of the machining process.

Finally, the application effect of the proposed DT model in the robotic milling of LSC assembly interfaces is verified through experiments, and the results show that this method can successfully optimize and improve the robotic machining process.