现代智慧城市中建筑群的抗灾能力对于城市的规划、建设和发展具有重要的意义。尽管单体建筑的可靠度研究已相当成熟，但针对建筑群抗灾可靠性的基础理论和模拟方法的研究仍相当薄弱。本项目通过学科交叉，引入大数据技术研究建筑群的抗灾可靠性预测与灾变模拟，结合多尺度物理-力学的精细化物理-数值模拟和结构物理-力学体征大数据、荷载作用与环境感控大数据，利用深度学习和大数据信息挖掘技术，建立城市建筑群荷载模型、灾害动力作用模型，性能退化模型，发展工程结构全寿命抗灾整体可靠性形成机制与可靠性指纹信息识别方法与类比推断理论，并进行建筑群灾变模拟技术的研究，实现城市建筑群全寿命抗灾可靠性的高效推断与区域建筑群灾害预测，在天河二号计算机上进行示范，为大数据驱动的智慧城市的防灾决策和评估提供技术支撑。

The ability of disaster prevention of group of buildings in modern smart city is crucial for the city’s planning, programming and developing. Although methodology for reliability analysis of individual building has been well developed, the basic theory and simulation method for reliability analysis of a group of buildings are very limited. To overcome this problem, this project is designed to, through disciplinary crossing and integration, study the reliability analysis and disaster simulation of group of buildings using the big data technology. To this purpose, database for live load, environmental load and disaster load will be established using unstructured data such as video, picture, various kinds of daily records and so on. With the help of technologies as information mining and deep learning, and based on the ideas of multi-scale physical-mechanical model and physical-numerical simulation, statistical models of live load, environmental load and material deterioration will be developed. Further, investigate the mechanism of global reliability of structures and the fingerprint identification method and analogical inference theory. Meanwhile, the collapse simulation technology for a group of buildings will be studied. Based on all the above, theory and simulation method for disaster prevention reliability analysis of group of buildings will be developed, which will be demonstrated on the Tianhe-2. The methodology developed in this project will provide strong support for big-data-driven disaster prevention decision for smart city.