长江科学院院报 ›› 2019, Vol. 36 ›› Issue (10): 104-110.DOI: 10.11988/ckyyb.20190948

• 堤防工程安全运行与监测预警 • 上一篇    下一篇

基于人工智能的堤防工程大数据安全管理平台及其实现

饶小康1,2,3, 马瑞1,2,3, 张力1,2,3, 义崇政1,2,3   

  1. 1.长江勘测规划设计研究有限责任公司,武汉 430010;
    2.长江空间信息技术工程有限公司(武汉),武汉 430010;
    3.湖北省水利信息感知与大数据工程技术研究中心,武汉 430010
  • 收稿日期:2019-08-05 出版日期:2019-10-01 发布日期:2019-10-21
  • 作者简介:饶小康(1985-),男,湖北黄冈人,高级工程师,硕士,研究方向为水利水电工程施工数字化、数据挖掘。E-mail:283139246@qq.com
  • 基金资助:
    国家重点研发计划项目(2017YFC1502604,2018YFC0407904)

Big Data Safety Management Platform for Dyke Engineering Based onArtificial Intelligence: Research and Implementation

RAO Xiao-kang1,2,3, MA Rui1,2,3, ZHANG Li1,2,3, YI Chong-zheng1,2,3   

  1. 1.Changjiang Institute of Survey, Planning, Design and Research Co., Ltd., Wuhan 430010, China;
    2.Changjiang Spatial Information Technology Engineering Co., Ltd., Wuhan 430010, China;
    3.Hydronformation Perception and Big Data Engineering Technology Research Center ofHubei Province, Wuhan 430010, China
  • Received:2019-08-05 Online:2019-10-01 Published:2019-10-21

摘要: 针对堤防工程综合安全监测体系,基于物联网技术,建立堤防工程海量数据资源的采集、汇集、交换与共享云平台;同时针对堤防工程质量、险情演化、致溃机理,基于大数据和人工智能技术,构建堤防工程大数据开放平台和人工智能计算平台,提供海量、多源、异构数据的融合、共享,实现风险识别、评估、预警模型的构建、运算。针对堤防工程安全防护、加固中填筑石料粒径级配要求,利用堤防工程物联网监测平台进行数据采集汇集,在堤防工程人工智能计算平台构建砂石爆破开采级配预测模型,解决多层非线性问题,进行爆破开采优化设计和石料粒径级配控制。实例证明,基于人工智能的堤防工程大数据安全管理平台可将堤防填筑石料开采级配平均相对误差率控制在21%以内,满足开采级配设计要求,控制石料块度,从而保障了堤防工程质量,实现了堤防工程的安全管理。研究成果可为堤防工程填筑石料开采设计提供技术参考,保障工程质量。

关键词: 堤防工程, 安全管理平台, 大数据, 人工智能, 深度学习, 物联网, 开采级配设计要求

Abstract: A big data safety management platform for dyke engineering is presented in this paper. The internet of things (IoT) technology is adopted to build a cloud platform for the monitoring, collection, exchange, and sharing of massive data in the monitoring system. Meanwhile, big data and artificial intelligence technologies are employed for the fusion and sharing of massive, multi-source and heterogeneous data to identify and evaluate risks and construct early-warning model. Through the data acquisition and collection by the IoT monitoring platform, the model predicting the gradation after sandstone blasting is built to design the optimum blasting scheme and control the particle gradation of sandstone material. Engineering practice has demonstrated that the average relative error rate of controlling the gradation is within 21%, which meets the requirement and guarantees the dyke construction quality.

Key words: dyke engineering, safety management platform, big data, artificial intelligence, deep learning, IoT, design requirements of gradation

中图分类号: