留言板

尊敬的读者、作者、审稿人, 关于本刊的投稿、审稿、编辑和出版的任何问题, 您可以本页添加留言。我们将尽快给您答复。谢谢您的支持!

姓名
邮箱
手机号码
标题
留言内容
验证码

基于GCN−GRU的瓦斯浓度时空分布预测

秦嘉欣 葛淑伟 龙凤琪 张永茜 李雪

秦嘉欣,葛淑伟,龙凤琪,等. 基于GCN−GRU的瓦斯浓度时空分布预测[J]. 工矿自动化,2023,49(5):82-89, 111.  doi: 10.13272/j.issn.1671-251x.2022060105
引用本文: 秦嘉欣,葛淑伟,龙凤琪,等. 基于GCN−GRU的瓦斯浓度时空分布预测[J]. 工矿自动化,2023,49(5):82-89, 111.  doi: 10.13272/j.issn.1671-251x.2022060105
QIN Jiaxin, GE Shuwei, LONG Fengqi, et al. Spatiotemporal distribution prediction of gas concentration based on GCN-GRU[J]. Journal of Mine Automation,2023,49(5):82-89, 111.  doi: 10.13272/j.issn.1671-251x.2022060105
Citation: QIN Jiaxin, GE Shuwei, LONG Fengqi, et al. Spatiotemporal distribution prediction of gas concentration based on GCN-GRU[J]. Journal of Mine Automation,2023,49(5):82-89, 111.  doi: 10.13272/j.issn.1671-251x.2022060105

基于GCN−GRU的瓦斯浓度时空分布预测

doi: 10.13272/j.issn.1671-251x.2022060105
基金项目: 国家自然科学基金重点资助项目(61936008)。
详细信息
    作者简介:

    秦嘉欣(1997—),女,陕西宝鸡人,硕士研究生,研究方向为机器学习、深度学习,E-mail:durta_qin@163.com

  • 中图分类号: TD712

Spatiotemporal distribution prediction of gas concentration based on GCN-GRU

  • 摘要: 在煤矿井下复杂环境下,传统瓦斯浓度预测模型的预测精度较低,虽然通过引入各种优化算法对传统瓦斯浓度预测模型进行优化,提高了瓦斯浓度预测精度,但仅从时间维度进行建模,忽略了瓦斯浓度的空间特性,易导致重要先验知识丢失,影响预测效果。针对上述问题,提出一种基于图卷积神经网络(GCN)和门控循环单元(GRU)的瓦斯浓度时空分布预测模型。首先,对瓦斯浓度历史数据进行预处理,根据各采集节点间的空间距离,构建瓦斯浓度空间节点图,用于对节点间复杂的依赖关系进行建模。然后,在每个采样时间点,将瓦斯浓度和节点间的距离权重参数作为输入,获得瓦斯的空间节点图结构后,通过GCN进行空间特征自适应学习和图卷积运算,得到瓦斯浓度的空间特征,再将瓦斯浓度的空间特征信息转化为序列数据,输入到GRU。最后,GRU对时间序列下各时刻组成的瓦斯空间特征信息进行处理,通过基于序列到序列模型和自动编码器,生成模型预测结果。试验结果表明:① GCN−GRU模型能够较为准确地预测瓦斯浓度的总体变化趋势,预测结果与实际数据的拟合度优于历史平均(HA)模型和支持向量回归(SVR)模型。② GCN−GRU模型的均方根误差较HA模型、SVR模型、移动平均自回归(ARIMA)模型分别降低了0.5%,71.4%,37.9%,平均绝对误差分别降低了10.5%,82.4%,82.4%,准确率分别提高了0.06%,17.7%,13.8%,表明GCN−GRU模型具有较强的鲁棒性,且泛化性能较好。③ GCN−GRU模型较HA模型、SVR模型、ARIMA模型更能关注到前序重要特征的影响。这主要是由于GRU的2个门关注了数据的时间特征,GRU在保留门控功能的基础上,减少训练参数,在一定程度上提高了模型训练效率,降低了训练时长。

     

  • 图  1  采掘工作面瓦斯浓度变化

    Figure  1.  Change of gas concentration of mining working face

    图  2  瓦斯运移扩散运动模型

    Figure  2.  Gas diffusion movement model

    图  3  瓦斯传感器现场布置

    Figure  3.  Field layout of gas sensors

    图  4  图结构的瓦斯浓度数据

    Figure  4.  Gas concentration data of graph structure

    图  5  瓦斯浓度预测模型整体框架

    Figure  5.  The overall framework of gas prediction model

    图  6  图信号传递过程

    Figure  6.  Graph signal transmission process

    图  7  GRU模型结构

    Figure  7.  GRU model structure

    图  8  序列到序列模型内部结构

    Figure  8.  Internal structure of the sequence-to-sequence model

    图  9  模型参数对预测结果的影响

    Figure  9.  Influence of model parameters on prediction results

    图  10  本文模型的损失函数曲线

    Figure  10.  Loss function curve of the proposed model

    图  11  不同模型预测结果对比

    Figure  11.  Comparison of prediction results of different models

    表  1  各模型性能指标

    Table  1.   Performance indexes of each model

    模型均方根误差平均绝对误差准确率/%R2
    HA0.039 30.023 279.890.794 7
    SVR0.067 00.061 765.760.405 1
    ARIMA0.053 90.038 368.91*
    本文模型0.039 10.021 079.940.698 1
    下载: 导出CSV
  • [1] 秦玉金,苏伟伟,姜文忠,等. 我国矿井瓦斯涌出量预测技术研究进展及发展方向[J]. 煤矿安全,2020,51(10):52-59.

    QIN Yujin,SU Weiwei,JIANG Wenzhong,et al. Research progress and development direction of mine gas emission forecast technology in China[J]. Safety in Coal Mines,2020,51(10):52-59.
    [2] 王雨虹,王淑月,王志中,等. 基于改进蝗虫算法优化长短时记忆神经网络的多参数瓦斯浓度预测模型研究[J]. 传感技术学报,2021,34(9):1196-1203.

    WANG Yuhong,WANG Shuyue,WANG Zhizhong,et al. Multi-Parameter gas concentration prediction model based on improved locust algorithm to optimize long and short time memory neural network[J]. Chinese Journal of Sensors and Actuators,2021,34(9):1196-1203.
    [3] 戚昱. 基于信息融合和GA−BP的煤矿瓦斯浓度预测方法研究[J]. 煤炭技术,2022,41(6):159-161.

    QI Yu. Research on coal mine gas concentration prediction method based on information fusion and GA-BP[J]. Coal Technology,2022,41(6):159-161.
    [4] 冉啟华,吴何碧,丁力生,等. 基于A−GRU的瓦斯浓度序列预测研究[J]. 软件导刊,2022,21(5):38-42.

    RAN Qihua,WU Hebi,DING Lisheng,et al. Research on gas concentration sequence prediction based on A-GRU[J]. Software Guide,2022,21(5):38-42.
    [5] 梁运培,栗小雨,李全贵,等. 基于CS−LSTM的工作面瓦斯浓度智能预测研究[J]. 矿业安全与环保,2022,49(4):80-86.

    LIANG Yunpei,LI Xiaoyu,LI Quangui,et al. Research on intelligent prediction of gas concentration in working face based on CS-LSTM[J]. Mining Safety & Environmental Protection,2022,49(4):80-86.
    [6] 付华,刘雨竹,徐楠,等. 基于多传感器−深度长短时记忆网络融合的瓦斯浓度预测研究[J]. 传感技术学报,2021,34(6):784-790. doi: 10.3969/j.issn.1004-1699.2021.06.011

    FU Hua,LIU Yuzhu,XU Nan,et al. Research on gas concentration prediction based on multi-sensor-deep long short-term memory network fusion[J]. Chinese Journal of Sensors and Actuators,2021,34(6):784-790. doi: 10.3969/j.issn.1004-1699.2021.06.011
    [7] 吴响. 煤矿瓦斯场分布演化规律及其时空建模研究[D]. 徐州: 中国矿业大学, 2014.

    WU Xiang. Study on the evolution law of coal mine gas distribution field and its spatio-temporal modeling[D]. Xuzhou: China University of Mining and Technology, 2014.
    [8] SCARSELLI F,GORI M,TSOI A C,et al. The graph neural network model[J]. IEEE Transactions on Neural Networks,2009,20(1):61-80. doi: 10.1109/TNN.2008.2005605
    [9] 梁宏涛,刘硕,杜军威,等. 深度学习应用于时序预测研究综述[J]. 计算机科学与探索,2023(5):1-21.

    LIANG Hongtao,LIU Shuo,DU Junwei,et al. Review of deep learning applied to time series prediction[J]. Journal of Frontiers of Computer Science and Technolog,2023(5):1-21.
    [10] 李欢. 基于时空序列的瓦斯浓度预测方法研究[D]. 徐州: 中国矿业大学, 2019.

    LI Huan. The research of gas concentration prediction method based on space-time sequence[D]. Xuzhou: China University of Mining and Technology, 2019.
    [11] AQ 1029−2019 煤矿安全监控系统及检测仪器使用管理规范[S].

    AQ 1029-2019 Application and management standard for coal mine safety monitoring system and detector[S].
    [12] CUI Zhiyong,HENRICKSON K,KE Ruimin,et al. Traffic graph convolutional recurrent neural network:a deep learning framework for network-scale traffic learning and forecasting[J]. IEEE Transactions on Intelligent Transportation Systems,2020,21(11):4883-4894. doi: 10.1109/TITS.2019.2950416
    [13] 徐冰冰,岑科廷,黄俊杰,等. 图卷积神经网络综述[J]. 计算机学报,2020,43(5):755-780.

    XU Bingbing,CEN Keting,HUANG Junjie,et al. A survey on graph convolutional neural network[J]. Chinese Journal of Computers,2020,43(5):755-780.
    [14] FEY M, LENSSEN J E. Fast graph representation learning with PyTorch geometric[C]. International Conference on Learning Representations, New Orleans, 2019.
    [15] 丁遥,张志利,赵晓枫,等. 半监督局部特征保留图卷积高光谱图像分类[J]. 北京航空航天大学学报,2023(5):1-12.

    DING Yao,ZHANG Zhili,ZHAO Xiaofeng,et al. Semi-supervised locality preserving dense graph convolution for hyperspectral image classification[J]. Journal of Beijing University of Aeronautics and Astronautics,2023(5):1-12.
    [16] YU Bing, YIN Haoteng, ZHU Zhanxing. Spatio-temporal graph convolutional networks: a deep learning framework for traffic forecasting[C]. International Joint Conference on Artificial Intelligence, Stockholm, 2018: 3634-3640.
    [17] 张世琨,谢睿,叶蔚,等. 基于关键词的代码自动摘要[J]. 计算机研究与发展,2020,57(9):1987-2000.

    ZHANG Shikun,XIE Rui,YE Wei,et al. Keyword-based source code summarization[J]. Journal of Computer Research and Development,2020,57(9):1987-2000.
    [18] 金季豪,阮彤,高大启,等. 语义图驱动的面向复杂逻辑关系的自然语言问答[J]. 中文信息学报,2021,35(12):122-132.

    JIN Jihao,RUAN Tong,GAO Daqi,et al. Semantic graph driven question answering towards complex logical rrelationships[J]. Journal of Chinese Information Processing,2021,35(12):122-132.
    [19] WEISS R J, CHOROWSKI J, JAITLY N, et al. Sequence-to-sequence models can directly transcribe foreign speech[C]. Annual Conference of the International Speech Communication Association (InterSpeech), Stockholm, 2017.
    [20] MAZUMDAR S, KUMAR A S. Forecasting data center resource usage: an experimental comparison with time-series methods[M]. Cham: Springer International Publishing, 2018.
  • 加载中
图(11) / 表(1)
计量
  • 文章访问数:  92
  • HTML全文浏览量:  6
  • PDF下载量:  4
  • 被引次数: 0
出版历程
  • 收稿日期:  2022-06-29
  • 修回日期:  2023-05-09
  • 网络出版日期:  2022-12-01

目录

    /

    返回文章
    返回