This study aims to improve the efficiency of construction efficiency to ensure the infrastructure needs of urban development. Targeting at realizing the intelligent manufacturing of buildings, the digital twins are applied to all stages of building construction, and the three-dimensional (3D) modeling is implemented. A digital twins framework in the construction field is constructed based on the building information modeling (BIM). From physical construction site entity modeling (PCSE modeling) and digital twins virtual body modeling (DTVB modeling) to virtual and real interaction modeling (VRI modeling), the digital twins process is fully applied to the various stages of building construction. The performance of the method proposed is further verified through the analysis of specific data. It is found that the model proposed shows better performance on the same platform and node compared with other models, and it shows the best scalability in the iterative calculation test on the Flink platform through efficient memory management and incremental iteration mechanism. In addition, the model has a delay of less than 100 milliseconds: in the Sleep scenario, the system delay of the latest model is about 5 times that of on the Flink, and the system interruption probability will firstly decrease and then increase when the time division factor (TDF) is increased. This study provides important reference value for the intelligent development of the construction industry and the high-quality development of buildings.

本研究旨在提高建设效率，保障城市发展的基础设施需求。以实现建筑智能制造为目标，将数字孪生应用于建筑施工的各个阶段，实现三维（3D）建模。基于建筑信息模型（BIM）构建建筑领域的数字孪生框架。从物理工地实体建模（PCSE建模）和数字孪生虚拟体建模（DTVB建模）到虚实交互建模（VRI建模），数字孪生过程充分应用于建筑施工的各个阶段。通过对具体数据的分析，进一步验证了所提出方法的性能。发现所提出的模型在相同平台和节点上相比其他模型表现出更好的性能，并且通过高效的内存管理和增量迭代机制在 Flink 平台的迭代计算测试中表现出最佳的可扩展性。另外，模型延迟小于100毫秒：在Sleep场景下，最新模型的系统延迟大约是Flink上的5倍，系统中断概率会随着时间的推移先下降后上升。分频因子 (TDF) 增加。本研究为建筑业智能化发展和建筑高质量发展提供重要参考价值。通过高效的内存管理和增量迭代机制，在 Flink 平台的迭代计算测试中展现出最佳的可扩展性。另外，模型延迟小于100毫秒：在Sleep场景下，最新模型的系统延迟大约是Flink上的5倍，系统中断概率会随着时间的推移先下降后上升。分频因子 (TDF) 增加。本研究为建筑业智能化发展和建筑高质量发展提供重要参考价值。通过高效的内存管理和增量迭代机制，在 Flink 平台的迭代计算测试中展现出最佳的可扩展性。另外，模型延迟小于100毫秒：在Sleep场景下，最新模型的系统延迟大约是Flink上的5倍，系统中断概率会随着时间的推移先下降后上升。分频因子 (TDF) 增加。本研究为建筑业智能化发展和建筑高质量发展提供重要参考价值。最新模型的系统延迟大约是 Flink 上的 5 倍，并且随着时分因子（TDF）的增加，系统中断概率会先减小后增大。本研究为建筑业智能化发展和建筑高质量发展提供重要参考价值。最新模型的系统延迟大约是 Flink 上的 5 倍，并且随着时分因子（TDF）的增加，系统中断概率会先减小后增大。本研究为建筑业智能化发展和建筑高质量发展提供重要参考价值。