本研究改良數位雙生模型建構方法，並提出了一套解決虛實模型逼真度的一致性問題的方法。在此方法中，我們使用參數擬合、定量分析與誤差比較，結合逼真度指標來開發數位雙生技術。此方法能夠給予製造業在數位雙生發展的建議，讓未來在工業製造中面臨數位雙生一致性問題時，能夠參考本研究的建議，進行問題排除。在研究過程中，首先構建了物理實體層，接著建構數位雙生模型，包含虛擬模型、數據模型與知識模型。我們使用商業數位模擬軟體進行虛擬模型的建構，並針對數據模型調整參數結果與知識模型擬合結果進行定量分析、誤差與逼真度指標的比較，以觀察模擬軟體與實體機械手臂在運行加工過程中的工時誤差。最後，透過案例分析與討論，來展示本研究方法的可行性與有效性，並為未來的應用提供實際的參考。;In this study, we improve the digital twin model construction method and propose a methodology to solve the consistency problem of virtual model simulation. In this approach, we use parameter fitting, quantitative analysis and error comparison, combined with simulation metrics to develop digital twin technology. This methodology can give the manufacturing industry a suggestion for the development of digital twin, so that when facing digital twin consistency problems in industrial manufacturing in the future, they can refer to the suggestions in this study to troubleshoot the problems. In the research process, we first constructed a physical entity layer, and then constructed a digital twin model, including a virtual model, a data model, and a knowledge model. The virtual model was constructed using commercial digital simulation software, and the results of the data model tuning parameters were compared with those of the knowledge model for quantitative analysis, error and simulation metrics, in order to observe the work-time errors during the motion processing between the simulation software and the physical robotic arm. Finally, a case study and discussion are conducted to demonstrate the feasibility and effectiveness of this research methodology and to provide practical references for future applications.
