基于POD与机器学习的综掘工作面流场快速预测算法

金兵; 张浪; 李伟; 郑义; 刘彦青; 张逸斌

doi:10.13272/j.issn.1671-251x.2024080090

基于POD与机器学习的综掘工作面流场快速预测算法

doi: 10.13272/j.issn.1671-251x.2024080090

金兵^{1, 2, 3, 4,},
张浪^{1, 2, 3, 4},
李伟^{1, 2, 3, 4},
郑义^{1, 2, 3, 4},
刘彦青^{1, 2, 3, 4},
张逸斌^{1, 2, 3, 4}

1.
煤炭科学技术研究院有限公司，北京　100013
2.
煤科通安（北京）智控科技有限公司，北京　100013
3.
北京市煤矿安全工程技术研究中心，北京　100013
4.
煤炭智能开采与岩层控制全国重点实验室，北京　100013

基金项目: 国家重点研发计划项目（2023YFC3009001-8）；天地科技科技创新专项重点项目(2024-TD-ZD011-01)；煤科院技术创新基金项目（2021CX-II-15）。

详细信息

作者简介:
金兵（1994—），男，辽宁抚顺人，助理研究员，硕士，主要从事矿井通风与防尘技术研究工作，E-mail：1029347049@qq.com

中图分类号: TD714.4
计量
- 文章访问数: 48
- HTML全文浏览量: 8
- PDF下载量: 3
- 被引次数: 0
出版历程
- 收稿日期: 2024-08-30
- 修回日期: 2024-10-29
- 网络出版日期: 2024-09-29

Rapid prediction algorithm for flow field in fully mechanized excavation face based on POD and machine learning

JIN Bing^{1, 2, 3, 4
,},
ZHANG Lang^{1, 2, 3, 4},
LI Wei^{1, 2, 3, 4},
ZHENG Yi^{1, 2, 3, 4},
LIU Yanqing^{1, 2, 3, 4},
ZHANG Yibin^{1, 2, 3, 4}

1.
China Coal Research Institute, Beijing 100013, China
2.
CCRI Tong'an (Beijing) Intelligent Control Technology Co., Ltd., Beijing 100013, China
3.
Beijing Engineering and Research Center of Mine Safe, Beijing 100013, China
4.
State Key Laboratory of Intelligent Coal Mining and Stratum Control, Beijing 100013, China

摘要

摘要: 针对综掘工作面降尘措施难以合理利用的问题，提出了一种基于本征正交分解（POD）与机器学习的综掘工作面流场快速预测算法。利用计算流体力学（CFD）技术对多种工况下的综掘工作面风流场和粉尘浓度场进行模拟，获得高维度流场数据；利用POD对高维度流场数据进行降维，提取能够反映流场主要特性的核心模态，得到流场工况的基函数模态与模态系数；通过机器学习方法预测不同工况下占总能量达到90%以上的模态系数，从而对未知工况的模态系数进行预测，利用预测的模态系数与基函数模态进行重构，得到未知工况的风流场数据或粉尘浓度场。研究结果表明：综掘工作面数值模拟模型的相对误差在3%以内，能够准确反映实际的风流和粉尘分布状况；风流场选择前5阶模态，粉尘浓度场选择前7阶模态，即可兼顾POD重构精度与效率；支持向量机（SVM）模型对模态系数的预测能力优于随机森林及神经网络模型，针对60种不同工况，基于POD和SVM预测的风流速度、粉尘浓度与数值模拟结果之间的相对误差分别为0.36 m/s，86.24 mg/m³，风流场及粉尘浓度场平均预测耗时为73 s，实现了对矿井综掘工作面风流场和粉尘浓度场的高精度快速预测。
- 综掘工作面 /
- 粉尘防治 /
- 风流场预测 /
- 粉尘浓度场预测 /
- 本征正交分解 /
- 机器学习 /
- 支持向量机
Abstract: To effectively utilize dust suppression measures in fully mechanized excavation faces, this study proposed a rapid prediction algorithm for the flow field based on proper orthogonal decomposition (POD) and machine learning. First, computational fluid dynamics (CFD) technology was used to simulate the air flow field and dust concentration field under various conditions, generating high-dimensional flow field data. Then, the POD method was applied to reduce the dimensionality of this data, extracting core modes that captured the main characteristics of the flow field and producing basis function modes and mode coefficients. Machine learning techniques were subsequently used to predict the mode coefficients that accounted for over 90% of the total energy under different conditions, enabling predictions of mode coefficients for unknown conditions. Finally, by reconstructing the flow or dust concentration field data using the predicted mode coefficients and basis function modes, rapid and accurate predictions for the flow field in excavation faces were achieved. The results showed that the numerical simulation model for the excavation face had a relative error within 3%, accurately reflecting the actual air flow and dust distribution. Selecting the first five modes for the flow field and the first seven modes for the dust concentration field balanced the accuracy and efficiency of POD reconstruction. The support vector machine (SVM) model outperformed the Random Forest and Neural Network models in predicting mode coefficients. For 60 different conditions, the relative errors between the POD and SVM-predicted flow velocity and dust concentration, and the CFD results, were 0.36 m/s and 86.24 mg/m³, respectively. The average prediction time for the flow and dust concentration fields was 73 seconds, achieving high-precision, rapid predictions for airflow and dust concentration in mine excavation faces.
- fully mechanized excavation face /
- dust prevention and control /
- airflow field prediction /
- dust concentration field prediction /
- proper orthogonal decomposition /
- machine learning /
- support vector machine

HTML全文

图 1 基于POD与机器学习的综掘工作面流场快速预测流程

Figure 1. Fast prediction flow for fullymechanized excavation face based on proper orthogonal decomposition (POD) and machine learning

下载: 全尺寸图片幻灯片

图 2 综掘工作面数值模拟模型

1—工作面；2—掘进机；3—压风筒；4—输送带；5—抽风筒；6—巷道壁；7—出风口。

Figure 2. Numerical simulation model of fully mechanized excavation face

下载: 全尺寸图片幻灯片

图 3 网格独立性验证

Figure 3. Grid independence validation

下载: 全尺寸图片幻灯片

图 4 风流场和粉尘浓度场系统各阶模态能量贡献率及累计贡献率

Figure 4. Energy contribution rate and cumulative contribution rate of each mode in air flow field and dust concentration field system

下载: 全尺寸图片幻灯片

图 5 重构后风流场与粉尘浓度场及原流场MAE

Figure 5. Mean absolute error （MAE） of reconstructed air flow field, dust concentration field and original flow field

下载: 全尺寸图片幻灯片

图 6 展示平面

Figure 6. Display plane

下载: 全尺寸图片幻灯片

图 7 数值模拟结果与模态重构结果对比

Figure 7. Comparison of numerical simulation and mode reconstruction results

下载: 全尺寸图片幻灯片

图 8 数值模拟结果与预测结果对比

Figure 8. Comparison of numerical simulation and prediction results

下载: 全尺寸图片幻灯片

表 1 离散相参数设定

Table 1. Discrete phase parameter settings

设置选项	设定
相间耦合	开启
耦合频率	20
粉尘材料	coal−mv
喷射方式	表面喷射
质量流率/（kg·s⁻¹）	0.008，0.012
最大颗粒直径/m	5×10⁻⁵，1×10⁻⁴，2×10⁻⁴
最小颗粒直径/m	5×10⁻⁷，1×10⁻⁶
积分尺度	0.15
粒径数量/个	10
湍流扩散模型	离散随机游走模型

下载: 导出CSV

表 2 边界条件设定

Table 2. Boundary condition setting

部件	边界条件	设定
压风筒	入口类型	速度入口
	入口速度/（m·s⁻¹）	8，9，10，11，12
	出风口类型	压力出口
抽风筒	入口类型	速度入口
抽风筒	入口速度/（m·s⁻¹）	10，11，12，13，14
巷道	出风口类型	压力出口
巷道	壁面类型	石壁
掘进机	壁面类型	金属壁
输送带	壁面类型	金属壁

下载: 导出CSV

表 3 风速及粉尘浓度模拟值与实测值对比

Table 3. Comparison of simulated and measured values of wind speed and dust concentration

参数	风速/（m·s⁻¹）	粉尘浓度/（mg·m⁻³）
模拟值	3.42	864.37
实测值	3.59	842.34
相对误差/%	4.74	2.62

下载: 导出CSV

表 4 3种机器学习模型性能评估指标

Table 4. Performance evaluation indicators for three machine learning models

模型	RMSE		R²
模型	风流场/（m·s⁻¹）	粉尘浓度场/（mg·m⁻³）	风流场	粉尘浓度场
SVM	0.29	425.47	0.90	0.86
RF	0.78	889.23	0.76	0.64
NN	1.98	1043.61	0.62	0.51

下载: 导出CSV

表 5 3种机器学习模型的交叉验证结果

Table 5. Cross-validation results of three machine learning models

流场类型	模型	平均 RMSE	标准差	95%置信区间
风流场/（m·s⁻¹）	SVM	0.31	0.40	[0.15, 0.47]
	RF	0.82	0.24	[0.36, 1.28]
	NN	2.05	0.22	[1.62, 2.48]
粉尘浓度场/（mg·m⁻³）	SVM	431.78	47.23	[341.42, 522.14]
	RF	905.67	78.34	[756.81, 1055.22]
	NN	1 048.92	96.75	[859.21, 1238.63]

下载: 导出CSV

参考文献(23)

[1]	袁亮. 煤矿粉尘防控与职业安全健康科学构想[J]. 煤炭学报,2020,45(1):1-7. YUAN Liang. Scientific conception of coal mine dust control and occupational safety[J]. Journal of China Coal Society,2020,45(1):1-7.
[2]	中华人民共和国国家卫生健康委员会规划发展与信息化司. 2018年我国卫生健康事业发展统计公报[EB/OL]. [2024-08-22]. http://www.Nhc.gov.cn/ guihuaxxs/s10748/201905/9b8d52727cf346049de8acce25ffcbd0.shtml. Planning,Development and Information Technology Department of National Health Commission of the People’s Republic of China. 2018 China's health development statistical bulletin[EB/OL]. [2024-08-22]. http://www.Nhc.gov.cn/guihuaxxs/s10748/201905/9b8d52727cf346049de8acce25ffcbd0.shtml.
[3]	赵春双,刘剑,田瑞祥,等. 压入式掘进通风流场PIV实验研究[J]. 矿业安全与环保,2017,44(2):21-25,30. doi: 10.3969/j.issn.1008-4495.2017.02.005 ZHAO Chunshuang,LIU Jian,TIAN Ruixiang,et al. PIV experiment study on flow field of forced ventilation system for roadway heading[J]. Mining Safety & Environmental Protection,2017,44(2):21-25,30. doi: 10.3969/j.issn.1008-4495.2017.02.005
[4]	韩敏,王建国,王康. 多抽风筒对综掘面除尘的影响研究[J]. 矿业安全与环保,2022,49(5):114-118. HAN Min,WANG Jianguo,WANG Kang. Study on the influence of multiple exhaust pipes on dust removal in fully mechanized excavation face[J]. Mining Safety & Environmental Protection,2022,49(5):114-118.
[5]	李天宇,陈曦,钟文琪. 基于CFD与POD的煤粉锅炉三维速度场快速预测[J]. 东南大学学报(自然科学版),2022,52(4):641-649. LI Tianyu,CHEN Xi,ZHONG Wenqi. Rapid prediction of three-dimensional velocity field of pulverized coal boiler based on CFD and POD[J]. Journal of Southeast University (Natural Science Edition),2022,52(4):641-649.
[6]	陈刚,李跃明. 非定常流场降阶模型及其应用研究进展与展望[J]. 力学进展,2011,41(6):686-701. doi: 10.6052/1000-0992-2011-6-lxjzJ2011-009 CHEN Gang,LI Yueming. Advances and prospects of the reduced order model for unsteady flow and its application[J]. Advances in Mechanics,2011,41(6):686-701. doi: 10.6052/1000-0992-2011-6-lxjzJ2011-009
[7]	傅奇星,张之豪,余秋阳,等. 基于POD模态拟合的汽车尾流场重构[C]. 中国汽车工程学会汽车空气动力学分会学术年会,上海,2022. FU Qixing,ZHANG Zhihao,YU Qiuyang,et al. Reconstruction of automobile wake based on POD modes fitting[C]. Annual Conference of the Automotive Aerodynamics Branch of the Chinese Society of Automotive Engineers,Shanghai,2022.
[8]	王磊,高丽敏,茅晓晨,等. 基于POD方法的对转压气机叶顶非定常流场分析[J/OL]. 航空动力学报:1-16[2024-08-22]. https://doi.org/10.13224/j.cnki.jasp.20220896. WANG Lei,GAO Limin,MAO Xiaochen,et al. Analysis of tip unsteady flow field in a counter-rotating compressor based on POD method[J/OL]. Journal of Aerospace Power:1-16 [2024-08-22]. https://doi.org/10.13224/j.cnki.jasp.20220896.
[9]	孙翀,田甜,竺晓程,等. 风力机翼型非定常流场POD和EPOD分析[J]. 上海交通大学学报,2022,56(1):45-52. SUN Chong,TIAN Tian,ZHU Xiaocheng,et al. Analysis of POD and EPOD for unsteady flow field of wind turbine airfoil[J]. Journal of Shanghai Jiao Tong University,2022,56(1):45-52.
[10]	贾续毅,龚春林,李春娜. 基于POD和BPNN的流场快速计算方法[J]. 西北工业大学学报,2021,39(6):1212-1221. doi: 10.3969/j.issn.1000-2758.2021.06.006 JIA Xuyi,GONG Chunlin,LI Chunna. Fast flow simulation method based on POD and BPNN[J]. Journal of Northwestern Polytechnical University,2021,39(6):1212-1221. doi: 10.3969/j.issn.1000-2758.2021.06.006
[11]	肖颖,肖翔域,段壮,等. 采用本征正交分解和长短期记忆网络模型的离心泵流场预测[J/OL]. 西安交通大学学报:1-11[2024-08-22]. http://kns.cnki.net/kcms/detail/61.1069.T.20240703.1631.002.html. XIAO Ying,XIAO Xiangyu,DUAN Zhuang,et al. Flow field prediction in centrifugal pump based on the proper orthogonal decomposition-radial basis function model[J/OL]. Journal of Xi'an Jiaotong University:1-11 [2024-08-22]. http://kns.cnki.net/kcms/detail/61.1069.T.20240703.1631.002.html.
[12]	LIU Qiang,NIE Wen,HUA Yun,et al. The effects of the installation position of a multi-radial swirling air-curtain generator on dust diffusion and pollution rules in a fully-mechanized excavation face:a case study[J]. Powder Technology,2018,329:371-385. doi: 10.1016/j.powtec.2018.01.064
[13]	AMIRI Z,MOVAHEDIRAD S. Bubble-induced particle mixing in a 2-D gas-solid fluidized bed with different bed aspect ratios:a CFD-DPM study[J]. Powder Technology,2017,320:637-645. doi: 10.1016/j.powtec.2017.07.097
[14]	HU Shengyong,LIAO Qi,FENG Guorui,et al. Numerical study of gas-solid two-phase flow around road-header drivers in a fully mechanized excavation face[J]. Powder Technology,2019,344:959-969. doi: 10.1016/j.powtec.2018.12.076
[15]	于欣,陈连军,刘国明. 喷浆作业粉尘分布影响因素的数值模拟[J]. 矿业研究与开发,2017,37(2):97-101. YU Xin,CHEN Lianjun,LIU Guoming. Numerical simulation of influencing factors on dust distribution during shotcreting[J]. Mining Research and Development,2017,37(2):97-101.
[16]	周刚,张琦,白若男,等. 大采高综采面风流−呼尘耦合运移规律CFD数值模拟[J]. 中国矿业大学学报,2016,45(4):684-693. ZHOU Gang,ZHANG Qi,BAI Ruonan,et al. CFD simulation of air-respirable dust coupling migration law at fully mechanized mining face with large mining height[J]. Journal of China University of Mining & Technology,2016,45(4):684-693.
[17]	CHENG Weimin,YU Haiming,ZHOU Gang,et al. The diffusion and pollution mechanisms of airborne dusts in fully-mechanized excavation face at mesoscopic scale based on CFD-DEM[J]. Process Safety and Environmental Protection,2016,104:240-253. doi: 10.1016/j.psep.2016.09.004
[18]	ZHAO Kai,JANUTOLO M,BARLA G. A completely 3D model for the simulation of mechanized tunnel excavation[J]. Rock Mechanics and Rock Engineering,2012,45(4):475-497. doi: 10.1007/s00603-012-0224-3
[19]	ROWLEY C W,COLONIUS T,MURRAY R M. Model reduction for compressible flows using POD and Galerkin projection[J]. Physica D:Nonlinear Phenomena,2004,189(1/2):115-129.
[20]	李鑫灵,袁梅,董洪,等. PSO−SVM模型在掘进工作面突出预警系统中的应用[J]. 煤矿安全,2021,52(9):90-95. LI Xinling,YUAN Mei,DONG Hong,et al. Application of PSO-SVM model in outburst warning system of heading face[J]. Safety in Coal Mines,2021,52(9):90-95.
[21]	成小雨,周爱桃,郭焱振,等. 基于随机森林与支持向量机的回采工作面瓦斯涌出量预测方法[J]. 煤矿安全,2022,53(10):205-211. CHENG Xiaoyu,ZHOU Aitao,GUO Yanzhen,et al. Prediction method of gas emission based on random forest and support vector machine[J]. Safety in Coal Mines,2022,53(10):205-211.
[22]	张浪,张迎辉,张逸斌,等. 基于机器学习的通风网络故障诊断方法研究[J]. 工矿自动化,2022,48(3):91-98. ZHANG Lang,ZHANG Yinghui,ZHANG Yibin,et al. Research on fault diagnosis method of ventilation network based on machine learning[J]. Journal of Mine Automation,2022,48(3):91-98.
[23]	刘彦青. 基于巷道摩擦阻力系数BP神经网络预测模型的矿井风网风量预测研究[J]. 矿业安全与环保,2021,48(2):101-106. LIU Yanqing. Study on the air quantity of mine ventilation network based on BP neural network prediction model of friction resistance coefficient in roadway[J]. Mining Safety & Environmental Protection,2021,48(2):101-106.