School of Engineering and Architecture, Chongqing University of Science and Technology, Chongqing 401331, China.
Comput Intell Neurosci. 2022 Mar 24;2022:4452128. doi: 10.1155/2022/4452128. eCollection 2022.
Digital twins (DTs) can realize the integration of information and entities. It is widely used because of its simulation characteristics and virtual reality (VR) mapping. Its application to industrial product management and control is explored. First, the concept and the functions in different stages of DTs are expounded. Second, the Workench simulation platform and SolidWorks software are applied in the design of the aluminum alloy flange according to DTs in the design stage of industrial product management and control. Third, the role of DTs in industrial product management and control is confirmed through a comparative experiment. Finally, an intelligent algorithm for the automatic identification of internal defects is designed based on lightweight deep learning to improve the efficiency of ultrasonic detection. The results show that the accuracy of the lightweight convolution neural network (CNN) is 94.1%; the model size is 2.9 MB; the network is more lightweight and has an excellent performance in ultrasonic defect detection; the nonlinear finite element analysis results and the test results are consistent. Therefore, it is proved that the finite element analysis method is reliable and helps to improve the efficiency and shorten the design cycle. The emergence of DTs provides a technical scheme for product management and control under the three-dimensional model.
数字孪生 (DT) 可以实现信息与实体的集成。由于其具有仿真特性和虚拟现实 (VR) 映射，因此被广泛应用。本文探讨了其在工业产品管理和控制中的应用。首先，阐述了 DT 的概念和不同阶段的功能。其次，根据工业产品管理和控制设计阶段的 DT，应用 Workench 仿真平台和 SolidWorks 软件进行铝合金法兰的设计。然后，通过对比实验确定了 DT 在工业产品管理和控制中的作用。最后，设计了一种基于轻量级深度学习的自动识别内部缺陷的智能算法，以提高超声检测的效率。结果表明，轻量级卷积神经网络 (CNN) 的准确率为 94.1%；模型大小为 2.9MB；该网络更加轻量级，在超声缺陷检测方面具有出色的性能；非线性有限元分析结果与测试结果一致。因此，证明了有限元分析方法是可靠的，有助于提高效率和缩短设计周期。DT 的出现为三维模型下的产品管理和控制提供了技术方案。
