| 注册
首页|期刊导航|水利学报|AI大模型在山洪灾害风险辨识与预警中的应用探讨

AI大模型在山洪灾害风险辨识与预警中的应用探讨

田济扬 章跃芬 严登华 徐晟昊 段嘉程 李建柱

水利学报2026,Vol.57Issue(5):675-690,16.
水利学报2026,Vol.57Issue(5):675-690,16.DOI:10.3724/j.slxb.20250412

AI大模型在山洪灾害风险辨识与预警中的应用探讨

Exploring the application of large AI models in flash flood risk identification and early warning

田济扬 1章跃芬 1严登华 1徐晟昊 2段嘉程 2李建柱3

作者信息

  • 1. 中国水利水电科学研究院,流域水循环与水安全全国重点实验室,北京 100038||水利部防洪抗旱减灾工程技术研究中心,北京 100038
  • 2. 水利部防洪抗旱减灾工程技术研究中心,北京 100038||天津大学 水利工程智能建设与运维全国重点实验室,天津 300350
  • 3. 天津大学 水利工程智能建设与运维全国重点实验室,天津 300350
  • 折叠

摘要

Abstract

Flash flood disasters are the main cause of fatalities among flood-related hazards in China.Risk identifica-tion and early warning represent crucial technologies for the proactive defense against flash floods.This paper system-atically reviews conventional techniques for flash flood risk identification and early warning both within China and abroad,providing a detailed analysis of the advantages,limitations and bottlenecks of various methods.By tracing the development of artificial intelligence(AI)and its role in transforming research paradigms in hydrological sci-ence,the study highlights the significant potential of AI in flash flood disaster prevention.It identifies six major chal-lenges confronting large AI models in the context of flash flood risk identification and early warning:data acquisition and quality,generalization and interpretability,balancing complexity with emergent capabilities,computational effi-ciency and parallel acceleration,and intelligent decision-making.Targeted solutions and future trends are discussed in response to these challenges.The paper proposes that future research should focus on big data governance,and the integration technology of AI and physical models.Additionally,efforts should focus on continuously improving hetero-geneous hybrid parallel computing framework and training strategies,advancing automated optimization of large model parameters and intelligent prediction,and enhancing learning capacity,generalization ability and risk predic-tion ability.The aim of these efforts is to provide both theoretical insights and practical guidance for the application of large AI models in flash flood risk identification and early warning.

关键词

山洪灾害/AI大模型/物理模型融合/智能预测/风险辨识与预警

Key words

flash flood disaster/large AI models/integration with physical models/intelligent prediction/risk iden-tification and early warning

分类

建筑与水利

引用本文复制引用

田济扬,章跃芬,严登华,徐晟昊,段嘉程,李建柱..AI大模型在山洪灾害风险辨识与预警中的应用探讨[J].水利学报,2026,57(5):675-690,16.

基金项目

国家重点研发计划项目(2024YFC3082200) (2024YFC3082200)

中国水利水电科学研究院五大人才计划项目(JZ0199A022021) (JZ0199A022021)

国家自然科学基金项目(52279022) (52279022)

水利学报

0559-9350

访问量0
|
下载量0
段落导航相关论文