Scholarlink

<p>Ensuring the safety of water infrastructure during emergencies requires rapid, evidence-based decisions that often exceed the capacity of conventional approaches. This study proposes a novel intelligent decision support model CBR-KG that integrates Case-Based Reasoning (CBR) with a domain Knowledge Graph (KG) to improve automation and reliability in emergency management decision. The knowledge graph, constructed from 88 hydropower project, encodes project, environment, risk, and control attributes. Methodologically, the model enhances similarity calculation by incorporating neighbor-structure features, applies a game-theoretic COWA–G1 scheme to fuse subjective and objective weights, and uses graph convolutional networks to complete missing entities and strengthen inference. A case study of the Huangyang Pumped Storage Power Plant in China demonstrates the full workflow from case retrieval to prioritized safety recommendations. Results show that the CBR-KG achieves node-classification accuracy/F1 of 0.9536/0.9523 and link-prediction accuracy/AUC of 0.9516/0.9944. The model further retrieves highly similar precedents and consolidates validated countermeasures against dominant risks. These findings confirm that coupling CBR with KG reasoning significantly improves decision accuracy and timeliness under uncertainly, offering actionable and data-informed recommendations for enhancing the safety of large-scale hydraulic infrastructure during emergencies.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Case-Based Reasoning and Knowledge Graph-Driven Decision Support Model for Emergency Risk Management in Hydraulic Engineering Safety

  • 期刊论文

  • 发表日期：

    2026年2月16日

查看全文
  • Jiayuan Guo,
  • Yijie Bian,
  • Ming Li,
  • Yuanwen Zhang,
  • Jianbo Du,
  • Jiarui Wang

摘要

Ensuring the safety of water infrastructure during emergencies requires rapid, evidence-based decisions that often exceed the capacity of conventional approaches. This study proposes a novel intelligent decision support model CBR-KG that integrates Case-Based Reasoning (CBR) with a domain Knowledge Graph (KG) to improve automation and reliability in emergency management decision. The knowledge graph, constructed from 88 hydropower project, encodes project, environment, risk, and control attributes. Methodologically, the model enhances similarity calculation by incorporating neighbor-structure features, applies a game-theoretic COWA–G1 scheme to fuse subjective and objective weights, and uses graph convolutional networks to complete missing entities and strengthen inference. A case study of the Huangyang Pumped Storage Power Plant in China demonstrates the full workflow from case retrieval to prioritized safety recommendations. Results show that the CBR-KG achieves node-classification accuracy/F1 of 0.9536/0.9523 and link-prediction accuracy/AUC of 0.9516/0.9944. The model further retrieves highly similar precedents and consolidates validated countermeasures against dominant risks. These findings confirm that coupling CBR with KG reasoning significantly improves decision accuracy and timeliness under uncertainly, offering actionable and data-informed recommendations for enhancing the safety of large-scale hydraulic infrastructure during emergencies.