摘要
数十年间,地震学家及地震工程学家通力合作,为包括地震构造特征、地震活动性、震源特性、地震动预测模型及场地效应等多个关键问题的解决提供了支撑,形成了地球科学与工程科学交叉融合的具有独特性的工程地震学科,并取得了系统的应用性研究成果。作为工程地震学重要分支的强震动地震学得到了迅猛发展,为地震区划和工程抗震研究奠定了坚实基础,为城乡建设和核电、交通、能源等多类型行业的发展提供了地震安全保障。近年来,随着算力、算法及算料(数据)等人工智能关键要素的大力发展,进一步实现强震动地震学与信息学科交叉成为可能,也迅速成为本领域的热点问题。本文首先分析了强震动地震学研究进展与关键问题,探讨了其与人工智能交叉的框架。而后从知识嵌入、数据-知识融合及知识发现3个层面,综述了行业研究成果,重点介绍:(1)地震波动相关的控制方程与边界、初始条件物理嵌入理论与求解方法;(2)数据与物理机制联合驱动的人工智能地震动模型构建理论与方法;(3)强震动人工智能生成模型等。最后,讨论了目前强震动地震学与人工智能研究亟须解决的关键问题,并对未来的发展方向进行了展望。
Over the past few decades,seismologists and earthquake engineers have worked in close collaboration,jointly addressing a range of critical issues,including seismotectonic characteristics,seismicity,source properties,ground-motion modeling,and site effects.These efforts have led to the establishment and maturation of engineering seismology,a distinctive interdisciplinary field bridging Earth sciences and engineering sciences,and have produced a series of systematic and impactful applied research outcomes.Among its subdisciplines,strong-motion seismology has developed particularly rapidly,providing a solid scientific foundation for seismic zonation and earthquake-resistant engineering,and playing a crucial role in ensuring seismic safety for urban and rural construction as well as for key sectors such as nuclear power,transportation,and energy.In recent years,rapid advances in artificial intelligence—especially in computing power,algorithms,and data availability—have created new opportunities for deep integration between strong ground motion seismology and information science,making this interdisciplinary direction a major focus of current research.This paper first reviews recent progress and outstanding challenges in strong ground motion seismology and outlines a conceptual framework for its integration with artificial intelligence.It then surveys representative studies from three perspectives:knowledge embedding,data–knowledge fusion,and knowledge discovery,with emphasis on the following topics:(1)theoretical foundations and solution strategies for physically embedding governing equations,boundary conditions,and initial conditions associated with seismic wave propagation;(2)theories and methodologies for developing artificial-intelligence-based ground-motion models driven by the joint constraints of observational data and physical mechanisms;and(3)artificial-intelligence-based models for strong-motion simulation and generation.Finally,the paper discusses key issues that urgently need to be addressed in current interdisciplinary research between strong-motion seismology and artificial intelligence,and provides perspectives on future development directions.
作者
陈苏
傅磊
丁毅
刘献伟
李多为
胡晓虎
李小军
Chen Su;Fu Lei;Ding Yi;Liu Xianwei;Li Duowei;Hu Xiaohu;Li Xiaojun(Key Laboratory of Urban Security and Disaster Engineering of the Ministry of Education,Beijing University of Technology,Beijing 100124,China;Institute of Geophysics,China Earthquake Administration,Beijing 100081,China;School of Civil and Environmental Engineering,Nanyang Technological University,Singapore 639798)
出处
《震灾防御技术》
北大核心
2025年第4期697-716,共20页
Technology for Earthquake Disaster Prevention
基金
国家重点研发计划项目(2023YFC3007403)
国家自然科学基金(52192675、51878626)。
关键词
强震动地震学
人工智能
学科交叉
知识嵌入
知识发现
Engineering seismology
Artificial intelligence
Interdisciplinary studies
Knowledge embedding
Knowledge discovery
