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CNN的煤岩瓦斯复合动力灾害预测

王凯 李康楠 杜锋 张翔 王衍海 周家旭

王凯, 李康楠, 杜锋, 张翔, 王衍海, 周家旭. 基于CNN的煤岩瓦斯复合动力灾害预测[J]. 矿业科学学报, 2023, 8(5): 613-622. doi: 10.19606/j.cnki.jmst.2023.05.003
引用本文: 王凯, 李康楠, 杜锋, 张翔, 王衍海, 周家旭. 基于CNN的煤岩瓦斯复合动力灾害预测[J]. 矿业科学学报, 2023, 8(5): 613-622. doi: 10.19606/j.cnki.jmst.2023.05.003
Wang Kai, Li Kangnan, Du Feng, Zhang Xiang, Wang Yanhai, Zhou Jiaxu. Prediction of coal-gas compound dynamic disaster based on convolutional neural network[J]. Journal of Mining Science and Technology, 2023, 8(5): 613-622. doi: 10.19606/j.cnki.jmst.2023.05.003
Citation: Wang Kai, Li Kangnan, Du Feng, Zhang Xiang, Wang Yanhai, Zhou Jiaxu. Prediction of coal-gas compound dynamic disaster based on convolutional neural network[J]. Journal of Mining Science and Technology, 2023, 8(5): 613-622. doi: 10.19606/j.cnki.jmst.2023.05.003

基于CNN的煤岩瓦斯复合动力灾害预测

doi: 10.19606/j.cnki.jmst.2023.05.003
基金项目: 

国家自然科学基金 52130409

国家自然科学基金 52004291

详细信息
    作者简介:

    王凯(1972—),男,河南遂平人,教授,博士生导师,主要从事安全工程与应急管理、矿井瓦斯及煤岩动力灾害防治、矿井通风等方面的教学与研究工作。Tel:13810850966,E-mail:kaiwang@cumtb.edu.cn

  • 中图分类号: TD713

Prediction of coal-gas compound dynamic disaster based on convolutional neural network

  • 摘要: 随着我国煤矿开采逐渐进入深部区域,煤岩瓦斯复合动力灾害日益严重,对煤矿的安全生产造成极大威胁。基于某矿现场数据,采用智能预测手段对煤岩瓦斯复合动力灾害进行研究。首先,依据大数据处理流程,应用箱型图分析法(Box-plot)与多重插补法(MI)进行数据清洗,结合灰色关联度分析法(GRA)建立煤岩瓦斯复合动力灾害指标体系;然后应用主成分分析法(PCA)进行数据降维,结合深度学习中的卷积神经网络(CNN)建立基于BMGP-CNN的煤岩瓦斯复合动力灾害预测模型;运用现场案例数据将此模型与BP模型、随机森林(RF)模型、支持向量机(SVM)模型及人工神经网络(ANN)模型进行对比验证,发现BMGP-CNN模型预测结果的准确率最高,且该模型的收敛速度较快,能够在数秒内完成预测。研究结果对于煤岩瓦斯复合动力灾害的预测和防控具有重要意义。
  • 图  1  煤岩瓦斯复合动力灾害预测模型建立流程

    Figure  1.  Prediction model of coal-gas compound dynamical disaster

    图  2  卷积神经网络结构示意图

    Figure  2.  Structure of convolutional neural network

    图  3  煤岩瓦斯复合动力灾害预测指标体系

    Figure  3.  Prediction index system of coal-gas compound dynamical disasters

    图  4  碎石图

    Figure  4.  Broken stone diagram

    图  5  卷积层参数优化

    Figure  5.  Parameter optimization of convolution layer

    图  6  其他参数优化

    Figure  6.  Other parameter optimization

    图  7  样本预测结果和实际结果

    Figure  7.  Sample prediction results

    表  1  部分完整的初始数据

    Table  1.   Initial data(part)

    序号 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15 X16 结果
    1 0.17 34.15 2.62 0 830 4.5 14.30 4 0.5 58.03 22.98 9 1.5 0.22 0 0 1
    2 0.50 28.10 0.20 1 434 4.5 2.22 2 0.5 59.87 8.67 10 0.6 0.07 0.47 0 0
    3 0.24 29.06 0.50 1 589 3.2 6.21 3 0.5 76.79 11.05 18 0.8 0.10 0 0 0
    4 0.44 29.06 0.78 1 588 2.9 6.98 3 0.5 43.44 9.29 19 1.2 0.47 0 0 2
    5 0.81 29.06 0.42 1 566 3.5 5.84 2 0.5 45.78 9.00 16 0 0 0 0 2
    6 0.17 34.15 2.65 0 840 4.5 14.10 4 0.5 35.05 23.52 11 1.2 0.19 0 0 1
    105 0.22 29.06 0.42 1 557 3.1 5.67 3 0.5 73.13 21.08 21 0 0 0.36 0.17 1
    下载: 导出CSV

    表  2  关联度与关联度排序

    Table  2.   Relevance and relevance ranking

    影响因素 关联度 关联度排名
    X1 0.879 48 1
    X2 0.879 45 2
    X3 0.874 15 3
    X4 0.872 95 4
    X5 0.872 83 5
    X6 0.869 74 6
    X7 0.867 43 7
    X8 0.867 35 8
    X9 0.849 27 9
    X10 0.849 00 10
    X11 0.836 547 11
    X12 0.821 76 12
    X13 0.788 63 13
    X14 0.786 68 14
    X15 0.776 99 15
    X16 0.762 00 16
    下载: 导出CSV

    表  3  成分矩阵

    Table  3.   Component matrix

    原始指标 成分
    1 2 3 4
    埋深 0.887 -0.276 0.073 -0.062
    软分层变化 0.001 0.468 0.729 0.045
    煤体破坏类型 0.603 0.402 -0.088 0.223
    煤厚 -0.029 0.108 0.142 0.837
    断层数量 0.689 0.449 -0.284 0.093
    坚固性系数 -0.483 -0.608 0.402 0.066
    瓦斯压力 0.580 -0.209 0.393 0.220
    瓦斯含量 0.810 -0.319 0.112 -0.213
    顶板抗压强度 -0.042 0.547 0.386 -0.474
    最大主应力 0.828 -0.237 0.095 -0.075
    下载: 导出CSV

    表  4  计算后的部分公因子数据

    Table  4.   Calculated common factor data(part)

    序号 Y1 Y2 Y3 Y4
    1 16.122 -1.956 1.166 -0.143
    2 -5.264 0.002 0.055 0.229
    3 1.120 0.789 0.024 -1.968
    4 0.906 -1.686 -0.191 -1.054
    5 -2.115 -3.156 0.760 -0.670
    6 16.336 -2.853 0.500 0.690
    105 1.487 -1.247 0.025 0.504
    下载: 导出CSV

    表  5  卷积神经网络预测模型参数

    Table  5.   Structure of CNN prediction model

    模型参数 取值
    卷积核尺寸 1×5
    卷积核数量 128
    步长 1
    卷积层数 1
    激活函数 ReLU
    池化层参数 1×5,1
    Dropout概率 0
    下载: 导出CSV

    表  6  各个模型的预测结果与准确率

    Table  6.   Prediction results and accuracy of each model

    序号 实际结果 BP RF SVM ANN
    1 0 0 1 0 0
    2 2 2 2 0 2
    3 1 1 0 2 1
    4 2 2 2 2 2
    5 0 0 0 0 1
    6 0 1 2 0 0
    7 2 2 2 2 2
    8 1 1 1 1 1
    9 0 0 0 0 2
    10 2 2 2 0 2
    11 1 1 1 1 1
    12 2 0 2 2 2
    13 1 2 1 1 2
    14 1 2 1 1 1
    15 2 1 2 1 2
    测试集准确率/% 66.7 80 73.3 80
    下载: 导出CSV
  • [1] Wang K, Zhou A T, Zhang J F, et al. Real-time numerical simulations and experimental research for the propagation characteristics of shock waves and gas flow during coal and gas outburst[J]. Safety Science, 2012, 50(4): 835-841. doi: 10.1016/j.ssci.2011.08.024
    [2] Wang K, Du F. Coal-gas compound dynamic disasters in China: a review[J]. Process Safety and Environmental Protection, 2020, 133: 1-17. doi: 10.1016/j.psep.2019.10.006
    [3] 潘一山. 煤与瓦斯突出、冲击地压复合动力灾害一体化研究[J]. 煤炭学报, 2016, 41(1): 105-112. https://www.cnki.com.cn/Article/CJFDTOTAL-MTXB201601016.htm

    Pan Yishan. Integrated study on compound dynamic disaster of coal-gas outburst and rockburst[J]. Journal of China Coal Society, 2016, 41(1): 105-112. https://www.cnki.com.cn/Article/CJFDTOTAL-MTXB201601016.htm
    [4] 王凯, 赵恩彪, 郭阳阳, 等. 中间主应力影响下含瓦斯复合煤岩体变形渗流及能量演化特征研究[J]. 矿业科学学报, 2023, 8(1): 74-82. doi: 10.19606/j.cnki.jmst.2023.01.007

    Wang Kai, Zhao Enbiao, Guo Yangyang, et al. Deformation, seepage and energy evolution characteristics of gas-bearing coal-rock under intermediate principal stress[J]. Journal of Mining Science and Technology, 2023, 8(1): 74-82. doi: 10.19606/j.cnki.jmst.2023.01.007
    [5] 朱丽媛, 潘一山, 李忠华, 等. 深部矿井冲击地压、瓦斯突出复合灾害发生机理[J]. 煤炭学报, 2018, 43(11): 3042-3050. https://www.cnki.com.cn/Article/CJFDTOTAL-MTXB201811011.htm

    Zhu Liyuan, Pan Yishan, Li Zhonghua, et al. Mechanisms of rockburst and outburst compound disaster in deep mine[J]. Journal of China Coal Society, 2018, 43(11): 3042-3050. https://www.cnki.com.cn/Article/CJFDTOTAL-MTXB201811011.htm
    [6] 尹光志, 李星, 鲁俊, 等. 深部开采动静载荷作用下复合动力灾害致灾机理研究[J]. 煤炭学报, 2017, 42(9): 2316-2326. https://www.cnki.com.cn/Article/CJFDTOTAL-MTXB201709015.htm

    Yin Guangzhi, Li Xing, Lu Jun, et al. Disaster-causing mechanism of compound dynamic disaster in deep mining under static and dynamic load conditions[J]. Journal of China Coal Society, 2017, 42(9): 2316-2326. https://www.cnki.com.cn/Article/CJFDTOTAL-MTXB201709015.htm
    [7] 齐庆新, 潘一山, 李海涛, 等. 煤矿深部开采煤岩动力灾害防控理论基础与关键技术[J]. 煤炭学报, 2020, 45(5): 1567-1584. https://www.cnki.com.cn/Article/CJFDTOTAL-MTXB202005003.htm

    Qi Qingxin, Pan Yishan, Li Haitao, et al. Theoretical basis and key technology of prevention and control of coal-rock dynamic disasters in deep coal mining[J]. Journal of China Coal Society, 2020, 45(5): 1567-1584. https://www.cnki.com.cn/Article/CJFDTOTAL-MTXB202005003.htm
    [8] 张庆贺, 袁亮, 杨科, 等. 深井煤岩动力灾害的连续卸压开采防治机理[J]. 采矿与安全工程学报, 2019, 36(1): 80-86, 102. https://www.cnki.com.cn/Article/CJFDTOTAL-KSYL201901012.htm

    Zhang Qinghe, Yuan Liang, Yang Ke, et al. Mechanism analysis on continuous stress-relief mining for preventing coal and rock dynamic disasters in deep coal mines[J]. Journal of Mining & Safety Engineering, 2019, 36(1): 80-86, 102. https://www.cnki.com.cn/Article/CJFDTOTAL-KSYL201901012.htm
    [9] 刘喜军. 深井煤岩瓦斯动力灾害防治研究[J]. 煤炭科学技术, 2018, 46(11): 69-75. https://www.cnki.com.cn/Article/CJFDTOTAL-MTKJ201811011.htm

    Liu Xijun. Study on coal and rock gas dynamics disaster prevention and control in deep mine[J]. Coal Science and Technology, 2018, 46(11): 69-75. https://www.cnki.com.cn/Article/CJFDTOTAL-MTKJ201811011.htm
    [10] 齐庆新, 潘一山, 舒龙勇, 等. 煤矿深部开采煤岩动力灾害多尺度分源防控理论与技术架构[J]. 煤炭学报, 2018, 43(7): 1801-1810. https://www.cnki.com.cn/Article/CJFDTOTAL-MTXB201807002.htm

    Qi Qingxin, Pan Yishan, Shu Longyong, et al. Theory and technical framework of prevention and control with different sources in multi-scales for coal and rock dynamic disasters in deep mining of coal mines[J]. Journal of China Coal Society, 2018, 43(7): 1801-1810. https://www.cnki.com.cn/Article/CJFDTOTAL-MTXB201807002.htm
    [11] 窦林名, 何学秋, Ren Ting, 等. 动静载叠加诱发煤岩瓦斯动力灾害原理及防治技术[J]. 中国矿业大学学报, 2018, 47(1): 48-59. https://www.cnki.com.cn/Article/CJFDTOTAL-ZGKD201801007.htm

    Dou Linming, He Xueqiu, Ren Ting, et al. Mechanism of coal-gas dynamic disasters caused by the superposition of static and dynamic loads and its control technology[J]. Journal of China University of Mining & Technology, 2018, 47(1): 48-59. https://www.cnki.com.cn/Article/CJFDTOTAL-ZGKD201801007.htm
    [12] 王佳信, 周宗红, 张继华, 等. 煤与瓦斯突出危险性预测的SαS-PNN模型及应用[J]. 传感技术学报, 2017, 30(7): 1112-1118. https://www.cnki.com.cn/Article/CJFDTOTAL-CGJS201707024.htm

    Wang Jiaxin, Zhou Zonghong, Zhang Jihua, et al. SαS-PNN model for forecast of coal and gas outburst risk and its application[J]. Chinese Journal of Sensors and Actuators, 2017, 30(7): 1112-1118. https://www.cnki.com.cn/Article/CJFDTOTAL-CGJS201707024.htm
    [13] 王雨虹, 刘璐璐, 付华, 等. 基于改进BP神经网络的煤矿冲击地压预测方法研究[J]. 煤炭科学技术, 2017, 45(10): 36-40. https://www.cnki.com.cn/Article/CJFDTOTAL-MTKJ201710006.htm

    Wang Yuhong, Liu Lulu, Fu Hua, et al. Study on predicted method of mine pressure bump based on improved BP neural network[J]. Coal Science and Technology, 2017, 45(10): 36-40. https://www.cnki.com.cn/Article/CJFDTOTAL-MTKJ201710006.htm
    [14] 孙玉峰, 李中才. 支持向量机法在煤与瓦斯突出分析中的应用研究[J]. 中国安全科学学报, 2010, 20(1): 25-30, 179. https://www.cnki.com.cn/Article/CJFDTOTAL-ZAQK201001006.htm

    Sun Yufeng, Li Zhongcai. Application study of SVM in analysis of coal and gas outburst[J]. China Safety Science Journal, 2010, 20(1): 25-30, 179. https://www.cnki.com.cn/Article/CJFDTOTAL-ZAQK201001006.htm
    [15] Pan Y M, Deng Y H, Zhang Q Z, et al. Dynamic prediction of gas emission based on wavelet neural network toolbox[J]. Journal of Coal Science and Engineering: China, 2013, 19(2): 174-181.
    [16] 史策, 高峰, 陈连城, 等. 煤矿冲击地压预测的PCA-GRNN方法[J]. 中国安全科学学报, 2016, 26(7): 119-124. https://www.cnki.com.cn/Article/CJFDTOTAL-ZAQK201607022.htm

    Shi Ce, Gao Feng, Chen Liancheng, et al. Prediction of pressure bump in coal mine by PCA-GRNN[J]. China Safety Science Journal, 2016, 26(7): 119-124. https://www.cnki.com.cn/Article/CJFDTOTAL-ZAQK201607022.htm
    [17] Spitzer M, Wildenhain J, Rappsilber J, et al. BoxPlotR: a web tool for generation of box plots[J]. Nature Methods, 2014, 11(2): 121-2.
    [18] 孙玲莉, 董世杰, 杨贵军. 常用多重插补法的插补重数选择[J]. 统计与决策, 2019, 35(23): 5-10. https://www.cnki.com.cn/Article/CJFDTOTAL-TJJC201923002.htm

    Sun Lingli, Dong Shijie, Yang Guijun. Selection of imputation multiplicity on multiple imputation methods[J]. Statistics & Decision, 2019, 35(23): 5-10. https://www.cnki.com.cn/Article/CJFDTOTAL-TJJC201923002.htm
    [19] 赵国飞, 康天合, 郭俊庆, 等. 基于区间值灰色关联度的煤层气区块生产潜力评价模型及应用[J]. 采矿与安全工程学报, 2020, 37(4): 794-803. https://www.cnki.com.cn/Article/CJFDTOTAL-KSYL202004018.htm

    Zhao Guofei, Kang Tianhe, Guo Junqing, et al. Application of evaluation model for the production potential of coalbed methane block based on interval value grey relational degree theory[J]. Journal of Mining & Safety Engineering, 2020, 37(4): 794-803. https://www.cnki.com.cn/Article/CJFDTOTAL-KSYL202004018.htm
    [20] 陈绍杰, 刘久潭, 汪锋, 等. 基于PCA-RA的滨海矿井水源识别技术研究[J]. 煤炭科学技术, 2021, 49(2): 217-225. https://www.cnki.com.cn/Article/CJFDTOTAL-MTKJ202102025.htm

    Chen Shaojie, Liu Jiutan, Wang Feng, et al. Technological research on water source identiftcation of coastal coalmines based on PCA-RA[J]. Coal Science and Technology, 2021, 49(2): 217-225. https://www.cnki.com.cn/Article/CJFDTOTAL-MTKJ202102025.htm
    [21] 朱志洁, 张宏伟, 韩军, 等. 基于PCA-BP神经网络的煤与瓦斯突出预测研究[J]. 中国安全科学学报, 2013, 23(4): 45-50. https://www.cnki.com.cn/Article/CJFDTOTAL-ZAQK201304009.htm

    Zhu Zhijie, Zhang Hongwei, Han Jun, et al. Prediction of coal and gas outburst based on PCA-BP neural network[J]. China Safety Science Journal, 2013, 23(4): 45-50. https://www.cnki.com.cn/Article/CJFDTOTAL-ZAQK201304009.htm
    [22] 王道元, 王俊, 孟志斌, 等. 煤矿安全风险智能分级管控与信息预警系统[J]. 煤炭科学技术, 2021, 49(10): 136-144. https://www.cnki.com.cn/Article/CJFDTOTAL-MTKJ202110019.htm

    Wang Daoyuan, Wang Jun, Meng Zhibin, et al. Intelligent hierarchical management & control and information pre-warning system of coal mine safety risk[J]. Coal Science and Technology, 2021, 49(10): 136-144. https://www.cnki.com.cn/Article/CJFDTOTAL-MTKJ202110019.htm
    [23] 刘慧敏, 徐方远, 刘宝举, 等. 基于CNN-LSTM的岩爆危险等级时序预测方法[J]. 中南大学学报: 自然科学版, 2021, 52(3): 659-670. https://www.cnki.com.cn/Article/CJFDTOTAL-ZNGD202103001.htm

    Liu Huimin, Xu Fangyuan, Liu Baoju, et al. Time-series prediction method for risk level of rockburst disaster based on CNN-LSTM[J]. Journal of Central South University: Science and Technology, 2021, 52(3): 659-670. https://www.cnki.com.cn/Article/CJFDTOTAL-ZNGD202103001.htm
    [24] 陆继翔, 张琪培, 杨志宏, 等. 基于CNN-LSTM混合神经网络模型的短期负荷预测方法[J]. 电力系统自动化, 2019, 43(8): 131-137. https://www.cnki.com.cn/Article/CJFDTOTAL-DLXT201908018.htm

    Lu Jixiang, Zhang Qipei, Yang Zhihong, et al. Short-term load forecasting method based on CNN-LSTM hybrid neural network model[J]. Automation of Electric Power Systems, 2019, 43(8): 131-137. https://www.cnki.com.cn/Article/CJFDTOTAL-DLXT201908018.htm
    [25] 孙鑫, 徐杨, 林柏泉, 等. 煤与瓦斯突出影响因素评价分析的模糊层次分析方法[J]. 中国安全科学学报, 2009, 19(10): 145-149, 177. https://www.cnki.com.cn/Article/CJFDTOTAL-ZAQK200910028.htm

    Sun Xin, Xu Yang, Lin Baiquan, et al. Evaluation and analysis on influential factors of coal and gas outburst based on fuzzy analytic hierarchy process[J]. China Safety Science Journal, 2009, 19(10): 145-149, 177. https://www.cnki.com.cn/Article/CJFDTOTAL-ZAQK200910028.htm
    [26] 赵红泽, 王宇新, 李淋, 等. 基于灰色关联分析与GA-BP神经网络的拉斗铲生产能力预测[J]. 矿业科学学报, 2020, 5(1): 58-66. http://kykxxb.cumtb.edu.cn/article/id/265

    Zhao Hongze, Wang Xinyu, Li Lin, et al. Production capacity prediction of dragline based on grey correlation analysis and GA-BP neural network[J]. Journal of Mining Science and Technology, 2020, 5(1): 58-66. http://kykxxb.cumtb.edu.cn/article/id/265
    [27] Shah A D, Bartlett J W, James C, et al. Comparison of random forest and parametric imputation models for imputing missing data using MICE: A CALIBER study[J]. American Journal of Epidemiology, 2014(6): 764-774.
    [28] 温廷新, 张波, 邵良杉. 煤与瓦斯突出预测的随机森林模型[J]. 计算机工程与应用, 2014, 50(10): 233-237. https://www.cnki.com.cn/Article/CJFDTOTAL-JSGG201410050.htm

    Wen Tingxin, Zhang Bo, Shao Liangshan. Prediction of coal and gas outburst based on random forest model[J]. Computer Engineering and Applications, 2014, 50(10): 233-237. https://www.cnki.com.cn/Article/CJFDTOTAL-JSGG201410050.htm
    [29] 赵学军, 李育珍, 武文斌. BP神经网络改进TSVM的矿产资源评价模型研究[J]. 矿业科学学报, 2016, 1(2): 188-195. http://kykxxb.cumtb.edu.cn/article/id/26

    Zhao Xuejun, Li Yuzhen, Wu Wenbin. Mineral resources evaluation model research based on BP neural network and TSVM algorithm[J]. Journal of Mining Science and Technology, 2016, 1(2): 188-195. http://kykxxb.cumtb.edu.cn/article/id/26
    [30] 由伟, 刘亚秀, 李永, 等. 用人工神经网络预测煤与瓦斯突出[J]. 煤炭学报, 2007, 32(3): 285-287. https://www.cnki.com.cn/Article/CJFDTOTAL-MTXB200703013.htm

    You Wei, Liu Yaxiu, Li Yong, et al. Predicting the coal and gas outburst using artificial neural network[J]. Journal of China Coal Society, 2007, 32(3): 285-287. https://www.cnki.com.cn/Article/CJFDTOTAL-MTXB200703013.htm
    [31] 牟全斌. 工作面煤与瓦斯突出区域预测模型探讨[J]. 煤炭科学技术, 2009, 37(9): 44-47. https://www.cnki.com.cn/Article/CJFDTOTAL-MTKJ200909015.htm

    Mu Quanbin. Discussion on prediction model of coal and gas outburst area in coal mining face[J]. Coal Science and Technology, 2009, 37(9): 44-47. https://www.cnki.com.cn/Article/CJFDTOTAL-MTKJ200909015.htm
  • 加载中
图(7) / 表(6)
计量
  • 文章访问数:  395
  • HTML全文浏览量:  95
  • PDF下载量:  97
  • 被引次数: 0
出版历程
  • 收稿日期:  2023-02-17
  • 修回日期:  2023-04-03
  • 刊出日期:  2023-10-31

目录

    /

    返回文章
    返回