Abstract: As core facility of hydropower project, the surface condition of water- retaining structure would directly affect project safety and operational efficiency. Traditional treatment methods rely on manual inspections and empirical judgment, which suffer from issues such as low efficiency and insufficient accuracy. This paper proposes an intelligent solution based on large models, which achieves efficient reuse of experience by constructing a specialized knowledge base for addressing surface hazards on water-retaining structures. The paper details the knowledge base construction process—including data collection, cleaning, knowledge extraction, and storage—and utilizes large models to generate targeted treatment recommendations. Application validation demonstrates that this method can significantly improve treatment efficiency and accuracy, providing robust support for the safe operation and maintenance of water-retaining structures.

Key words: water-retaining structure, surface defect, large model, knowledge base, experience reuse

摘要： 水电站挡水建筑物作为水利工程的核心设施，其表面状态直接关系到工程安全与运行效率。传统治理方式依赖人工巡检与经验判断，存在效率低、准确性不足等问题。笔者提出基于大模型的智能化解决方案，通过构建挡水建筑物表面隐患治理专业知识库，实现经验的高效复用。笔者详细阐述了知识库构建流程，包括数据收集、清洗、知识抽取与存储，并利用大模型生成针对性治理建议。应用验证表明，该方法显著提升了治理效率与精准度，可为挡水建筑物安全运维提供有力支撑。

关键词: 挡水建筑物, 表面隐患, 大模型, 知识库, 经验复用