水利工程智能建造进展及关键技术  

Advances and key technologies of intelligent construction of hydraulic engineering

作　　者：刘耀儒[1] 侯少康 程立 黄跃群[1,3] LIU Yaoru;HOU Shaokang;CHENG Li;HUANG Yuequn(State Key Laboratory of Hydroscience and Engineering,Tsinghua University,Beijing 100084,China;China Renewable Energy Engineering Institute,Beijing 100120,China;Hunan Provincial Water Resources Development&Investment Co.,Ltd.,Changsha 410007,Hunan,China)

机构地区：[1]清华大学水沙科学与水利水电工程国家重点实验室,北京100084 [2]水电水利规划设计总院,北京100120 [3]湘水集团湖南省水利发展投资有限公司,湖南长沙410007

出　　处：《水利水电技术（中英文）》2022年第10期1-20,共20页Water Resources and Hydropower Engineering

基　　金：国家自然科学基金项目(52179105,41961134032)。

摘　　要：智能建造成为近年来工程建设领域的研究热点和发展趋势,通过集成人工智能、大数据和云计算等新一代信息技术,进一步推进了水利工程的安全、高效、高质、经济和绿色建设。简要回顾了水利工程智能建造的发展历程,从大坝智能建造的角度总结了智能碾压、智能温控、智能灌浆、智能安全监测和大坝智能施工机械5个方面的关键技术进展,并从隧洞智能建造的角度总结了围岩条件智能感知、地质灾害智能预测预警、TBM智能掘进、智能支护、隧洞智能施工机械5个方面的关键技术进展。目前,水利工程智能建造已逐渐由探索阶段向系统化发展阶段过渡,且已在一些工程中取得了良好的示范,但仍存在一些挑战和局限性。最后,从4个方面探讨了水利工程智能建造中的技术难点及展望,包括专业机理知识和数据驱动模型的融合、少样本条件下跨工程/工况的迁移学习、多源多模态异构互补数据信息的融合和基准数据集开发与数据标准化。Intelligent construction has become a research hotspot and development trend in the field of engineering construction in recent years. By integrating new generation information technologies, e.g., artificial intelligence, big data, block chain and cloud computing, it further promotes the safe, efficient, high-quality, economic and green construction of hydraulic engineering. The development process of intelligent construction in hydraulic engineering is briefly reviewed. The key technical progress of dam intelligent construction is summarized from five aspects: intelligent rolling, intelligent temperature control, intelligent grouting, intelligent safety monitoring and dam intelligent construction machine. The key technical progress of diversion tunnel intelligent construction is also summarized from five aspects: intelligent perception of surrounding rock conditions, intelligent prediction and early warning of geohazards, intelligent tunnelling of TBM, intelligent support and tunnel intelligent construction machine. Currently, intelligent construction of hydraulic engineering has gradually transitioned from the exploration stage to the systematization development stage, and has achieved good demonstration in some projects. However, there are still some challenges and limitations. Finally, the technical difficulties and prospects in the intelligent construction of hydraulic engineering are discussed from four aspects, including the fusion of professional mechanism knowledge and data-driven model, across-engineering/working conditions transfer learning under few-shot condition, the fusion of multi-source data, the development of benchmark data set and data standardization.

关 键 词：水利工程 智能建造 大坝 隧洞 机理-数据融合 迁移学习 多源数据 

分 类 号：TV52[水利工程—水利水电工程]