摘要
建筑结构设计随着计算机技术进步迈入新的发展阶段,基于模拟与基于人工智能的设计为两个重要发展方向。基于模拟的设计为超出设计规范的特殊结构或特殊工况设计提供了关键手段,其关键难题是发展可以准确、高效模拟结构在地震、火、风等灾害作用下受力行为的模拟方法。基于人工智能的设计则是通过学习既有设计案例,从而高效生成满足工程经验和设计规范要求的结构设计图,可以有效提升结构设计效率,其关键难题在于对工程经验和结构特征的学习和表达。该文介绍了作者团队开展的基于模拟的抗震、防火、抗风设计的典型案例,以及基于生成对抗网络和图神经网络的人工智能结构设计方法,提出了L0~L5不同层级结构智能设计的划分建议,并展望了未来将基于模拟的设计和基于人工智能的设计相融合的发展方向。
With the advancement of computer technology,building structural design has entered a new stage of development,and simulation-based and artificial intelligence(AI)-based design are two significant development avenues.Simulation-based design is an essential tool for designing complex and unique structures or under special loading conditions beyond the design specifications.The primary obstacle is to develop a simulation approach that can accurately and effectively represent the mechanical behavior of structures under earthquake,fire,wind,and other disasters.Moreover,by learning from existing design cases,AI-based design can efficiently develop structural design drawings that fulfill the needs of engineering experience and design standards,thereby improving the efficiency of structural design.The critical difficulty lies in the learning and expression of engineering knowledge and structural features.This paper describes the typical cases of simulation-based seismic,fire-resistant,and wind-resistant designs conducted by the author’s team,along with the intelligent structural design techniques based on generative adversarial networks and graph neural networks.Different intelligent structural design levels(L0-L5)are proposed.In addition,the paper proposes the future development direction of the combination of simulation-based design with AI-based design.
作者
陆新征
廖文杰
顾栋炼
许镇
郑哲
LU Xin-zheng;LIAO Wen-jie;GU Dong-lian;XU Zhen;ZHENG Zhe(Key Laboratory of Civil Engineering Safety and Durability of China Education Ministry,Department of Civil Engineering,Tsinghua University,Beijing 100084,China;Research Institute of Urbanization and Urban Safety,School of Civil and Resource Engineering,University of Science and Technology Beijing,Beijing 100083,China)
出处
《工程力学》
北大核心
2025年第3期1-17,共17页
Engineering Mechanics
基金
住房和城乡建设部科学技术计划项目(2022-K-073)
中国博士后科学基金项目(2022M721879)
国家自然科学基金项目(52238011)
腾讯基金会项目(科学探索奖)
清华大学“水木学者”计划项目(2022SM005)。
关键词
建筑结构设计
结构模拟方法
基于模拟的设计
生成式网络模型
结构智能设计
building structural design
structural simulation method
simulation-based design
generative neural networks
AI-based design
