Metal products are susceptible to factors such as cutting force, clamping force and heat in the machining process, resulting in product quality problems, such as geometric deformation and surface defects. The real-time observation and control of product quality are integral to optimizing machining process. Digital twin technologies can be used to monitor and control the quality of products via multi-scale based quality analysis. However, previous research on digital twin lacks a fine-grained expression and generation method for product multi-scale quality, making it impossible to carry out an in-depth analysis of product quality. Aiming at addressing this challenge, we study the multi-scale evolution mechanism of the digital twin model and explore the knowledge generation method of the digital twin data. The proposed method constructed the digital twin quality knowledge model from the macro, meso, and micro levels by utilizing the data of the digital twin mimic model. These multi-scale quality knowledge models can express product quality in a fine-grained way and provide data support for digital twin-based decision-making. Finally, we tested the method in monitoring and controlling the machining quality of an air rudder to verify the feasibility of the proposed method.

金属产品在加工过程中易受切削力，夹紧力和热量等因素的影响，从而导致产品质量问题，例如几何变形和表面缺陷。实时观察和控制产品质量是优化加工过程所不可或缺的。通过基于多尺度的质量分析，数字孪生技术可用于监视和控制产品质量。但是，以往对数字孪生的研究缺乏产品多尺度质量的细粒度表达和生成方法，无法对产品质量进行深入的分析。为了应对这一挑战，我们研究了数字孪生模型的多尺度演化机制，并探索了数字孪生数据的知识生成方法。所提出的方法利用数字双胞胎模拟模型的数据，从宏观，中观和微观层面构建了数字双胞胎质量知识模型。这些多尺度的质量知识模型可以细粒度地表达产品质量，并为基于双胞胎的数字决策提供数据支持。最后，我们测试了该方法在监测和控制空气舵的加工质量中的可行性，从而验证了该方法的可行性。