土木工程智能科学与技术研究现状及展望

人工智能为土木工程学科和行业带来了新的发展机遇，土木工程智能科学技术的发展和创新与智能材料、3D打印、大数据、物联网、智能计算、机器人、元宇宙等数学、计算机科学和新兴产业技术进行深度融合，将深刻变革土木工程全寿期，包括城市规划、建筑设计、结构设计、工程建造、服役运维、防灾减灾等阶段。为此，从智能材料、智能设计、智能建造、智能运维、智能防灾等多个维度，总结当前土木工程智能科学与技术的研究现状，提出物理机器学习(基于数据范式与物理定律有机融合)的土木工程智能科学与技术特征；同时，探讨虚拟科学家在创造崭新土木工程智能科学与技术的前景。最后，展望了土木工程智能科学与技术发展目标，即从基于现有知识体系建立纯数据驱动和物理机器学习土木工程智能理论，到发展出具有自学习能力的智能体及其理论体系，最终创造出具有自我进化能力的虚拟科学家，发现新的数学方法与物理定律等。

State-of-the-art and prospect of intelligent science and technology in civil engineering

Artificial intelligence has brought new development opportunities and vitality to classical civil engineering. In the era of artificial intelligence, the development and innovation of civil engineering science and technology are bound to be deeply integrated with mathematics, computer science and emerging industrial technologies, such as intelligent materials, 3D printing, big data, Internet of things, intelligent computing, robotics, meta-universe and so on. Artificial intelligence techniques will also profoundly change the development of civil engineering in urban planning, architectural design, structural design, engineering construction, service operation and maintenance, disaster prevention and mitigation in the whole life cycle. The state-of-the-art research on intelligent science and technology in civil engineering from multiple dimensions, including intelligent materials, intelligent design, intelligent construction, intelligent operation and maintenance, and intelligent disaster prevention, were summarized. The naturally integrated characteristics of civil engineering intelligent science and technology on the basis of physical machine learning (merging data paradigm with physical laws) were presented. Meanwhile, the prospect of creating novel civil engineering intelligent science and technology by virtual scientists were discussed. Finally, the prospect of intelligent science and technology in civil engineering is outlined: establishing intelligence theory for civil engineering in pure data-driven and physical machine learning paradigms based on the existing knowledge system; developing an intelligent agent and its theoretical system with self-learning ability; eventually creating self-evolving virtual scientists to discover new mathematical methods and physical laws.