• ▶ 2008-2024年被中国情报信息研究所评价中心评为“中国科技核心期刊”
  • ▶ 2019-2024年连续三届被中国科学院文献情报中心中国科学引文数据库CSCD(核心库)收录
  • ▶ 2021、2023年入编北京大学图书馆《中文核心期刊要目总览》
  • ▶ 2020-2024连续四年入选《科技期刊世界影响力指数(WJCI)报告》
张 晶, 葛 阳, 孙加龙, 平 扬, 张振洲, 刘智勇.考虑物理约束的结构化状态空间洪水预报模型研究水资源与水工程学报[J].,2025,36(5):93-101
考虑物理约束的结构化状态空间洪水预报模型研究
Structured state space flood forecasting model with physical constraints
  
DOI:10.11705/j.issn.1672-643X.2025.05.11
中文关键词:  洪水预报  物理约束  深度学习  结构化状态空间模型  Hydro-S4D模型  珠江流域
英文关键词:flood forecasting  physical constraint  deep learning  structured state space model  Hydro-S4D model  the Pearl River Basin
基金项目:国家重点研发计划项目(2023YFC3207500); 中国电力建设集团重点科技研发项目(DJ-ZDXM-2023-39)
作者单位
张 晶1, 葛 阳2, 孙加龙1, 平 扬1, 张振洲1, 刘智勇2 (1.中电建生态环境集团有限公司 广东 深圳 518101 2.中山大学 水资源与环境研究中心 广东 广州 510275) 
摘要点击次数: 512
全文下载次数: 138
中文摘要:
      针对现有深度学习洪水预报模型在处理长时序依赖时的性能衰减及缺乏物理机制约束的问题,提出一种融合结构化状态空间模型(S4D)与水文特定损失函数(HydroLoss)的洪水预报框架(Hydro-S4D)。该框架利用S4D模型高效捕捉长期依赖关系的能力,并设计了包含加权均方根误差、趋势一致性、峰值关注和相位校正的复合损失函数HydroLoss,将水文物理约束融入模型训练。以珠江三角洲典型水文站为例的应用结果表明:该框架显著提升了预报精度,1 d预见期纳什效率系数(NSE)达0.9,7 d预见期仍可维持在0.6左右;在洪峰拟合、过程形态保持以及不确定性量化方面,该框架均优于长短期记忆网络(LSTM)等对比模型,有效解决了其过度自信或不自信的问题。研究证实,将先进序列模型与蕴含领域知识的损失函数相结合,是提升中长期洪水预报精度与可靠性的有效路径,所提出的Hydro-S4D框架为数据驱动的水文预报模型赋予了更强的物理现实性和应用价值。
英文摘要:
      Existing deep learning models for flood forecasting face challenges of performance decay when dealing with long-term dependencies and a lack of physical constraints. To address these issues, this study proposes a flood forecasting framework, Hydro-S4D, which integrates a structured state space model (S4D) with a novel hydrology-specific loss function(HydroLoss). The framework leverages the S4D model’s capability to efficiently capture long-term dependencies, and proposes a customized composite loss function HydroLoss. This function integrates weighted MSE, trend consistency, peak attention, and phase correction to embed physical hydrological constraints into the model training process. Application to a typical hydrological station in the Pearl River Delta demonstrates that the framework significantly improves forecast accuracy, achieving a Nash-Sutcliffe efficiency (NSE) of 0.9 for 1-day lead time and maintaining it at approximately 0.6 for 7-day lead time. The proposed model outperforms benchmark models, such as the long short-term memory (LSTM) network, in fitting flood peaks, preserving hydrograph shapes, and providing balanced uncertainty quantification, thereby addressing their overconfidence or underconfidence issues. The results show that integrating advanced sequence models with domain-aware loss functions is an effective approach to enhancing the accuracy and reliability of medium- and long-term flood forecasts. The proposed Hydro-S4D framework imparts great physical realism and practical value to data-driven hydrological models.
查看全文  查看/发表评论  下载PDF阅读器
关闭