基于半监督学习的煤层钻孔预抽瓦斯状态评价方法

晏立, 文虎, 王振平, 金永飞

晏立,文虎,王振平,等. 基于半监督学习的煤层钻孔预抽瓦斯状态评价方法[J]. 工矿自动化,2025,51(3):113-121. DOI: 10.13272/j.issn.1671-251x.2025020046
引用本文: 晏立,文虎,王振平,等. 基于半监督学习的煤层钻孔预抽瓦斯状态评价方法[J]. 工矿自动化,2025,51(3):113-121. DOI: 10.13272/j.issn.1671-251x.2025020046
YAN Li, WEN Hu, WANG Zhenping, et al. Evaluation method for gas pre-extraction status in coal seam boreholes based on semi-supervised learning[J]. Journal of Mine Automation,2025,51(3):113-121. DOI: 10.13272/j.issn.1671-251x.2025020046
Citation: YAN Li, WEN Hu, WANG Zhenping, et al. Evaluation method for gas pre-extraction status in coal seam boreholes based on semi-supervised learning[J]. Journal of Mine Automation,2025,51(3):113-121. DOI: 10.13272/j.issn.1671-251x.2025020046

基于半监督学习的煤层钻孔预抽瓦斯状态评价方法

基金项目: 

国家自然科学基金项目(52274227)。

详细信息
    作者简介:

    晏立(1997—),男,重庆人,讲师,博士,从事应急技术与管理工作,E-mail:yanli@xust.edu.cn

  • 中图分类号: TD712

Evaluation method for gas pre-extraction status in coal seam boreholes based on semi-supervised learning

  • 摘要:

    目前单一钻孔抽采状态评价方法通常依赖于瓦斯抽采浓度,而忽视了煤层瓦斯赋存的多样性。监督学习模型依赖于样本的特征标记,在样本量较大时,人工标注的成本较高;无监督学习模型缺乏样本标记,无法实现定性评价。针对上述问题,提出一种基于半监督学习的煤层钻孔预抽瓦斯状态评价方法。构建了包含甲烷浓度、抽采负压、环境温度等 8 项指标的多维度评价体系,采用层次分析法(AHP)与模糊评价法(FEM)结合的权重赋值方法,建立抽采效果等级划分标准。在此基础上,提出基于高斯混合模型(GMM)与 K−Means算法的半监督学习模型(SSGMM/SSK−Means),通过融合少量人工标注样本与大量未标注数据,实现单一钻孔抽采状态的动态分类。SSGMM聚集度更好,SSK−Means效率更高,形成“精度−效率”的互补关系。在陕西黄陵二号煤矿 215 工作面的应用结果表明:SSGMM和 SSK−Means的最大聚集度(MVCR)和修正 Rand 指数(ARI)分别达 82.64% 和 85.83%,显著优于传统聚类方法;通过动态反馈机制优化后,原等级为“差”的钻孔抽采效率提升 5.26%~5.80%,补差率达 100%。

    Abstract:

    Current evaluation methods for single-borehole gas extraction status typically rely on gas concentration, while overlooking the diversity of coal seam gas occurrence. Supervised learning models depend on labeled sample features, but manual labeling becomes costly when the sample size is large. Unsupervised learning models lack sample labeling, making qualitative evaluation infeasible. To address these issues, an evaluation method based on semi-supervised learning was proposed for the gas pre-extraction status evaluation of coal seam boreholes. A multi-dimensional evaluation system was established, incorporating eight indicators such as methane concentration, extraction negative pressure, and ambient temperature. The weighting method combining the analytic hierarchy process (AHP) and fuzzy evaluation method (FEM) was used to establish classification standards for extraction performance. Building on this, a semi-supervised learning model based on the Gaussian mixture model (GMM) and K-Means algorithm (SSGMM/SSK-Means) was developed. By integrating a small number of manually labeled samples and a large quantity of unlabeled data, the model enabled dynamic classification of single-borehole extraction status. The SSGMM demonstrated better clustering rate, while the SSK-Means achieved higher efficiency, developing a complementary "accuracy-efficiency" relationship. The application results from the 215 working face of the Huangling No. 2 Coal Mine in Shaanxi Province showed that the maximum validity clustering rate (MVCR) and adjusted rand index (ARI) of SSGMM and SSK-Means reached 82.64% and 85.83%, respectively, significantly outperforming conventional clustering methods. After optimization through a dynamic feedback mechanism, boreholes initially classified as "poor" showed an improvement of 5.26% to 5.80% in extraction efficiency, achieving a 100% remediation rate.

  • 随着煤矿智能化的快速发展,电缆在煤矿供电中的应用越来越广泛。但由于煤矿环境恶劣,电缆在运行过程中极易受到热应力、机械应力、电压应力等各种因素的影响,使电缆绝缘、护套等发生损坏,产生局部过热和漏电现象,不但会影响煤矿的正常开采工作,严重时还会引发火灾等安全事故。据统计,电缆故障造成的煤矿事故占比超过50%[1]。因此,准确检测矿用电缆的运行状态并及时排查安全隐患成为煤矿领域亟待解决的问题[2-4]

    目前,矿用电缆状态监测和故障诊断方法主要包括低压脉冲法、局放法、低频电流叠加法、直流分量法和直流叠加法等[5-9]。这些方法虽然对特定故障具有良好的诊断效果,但无法全面评估电缆各个部位的劣化状态,更无法对电缆的劣化和故障趋势进行预测。此外,这些诊断方法需要电缆停止运行后才能进行检测,严重影响煤矿的正常生产。

    20世纪20年代,有研究人员发现电气设备发生故障时会产生相应的谐波,催生了电气设备的谐波诊断技术[10-11]。近年来的研究发现电缆在受到热、电压、环境和机械应力时,会导致电缆介质磁束变化和介质振动,从而产生高次谐波,因此众多学者开始研究电缆的谐波诊断技术。文献[12]通过有限元法对电缆缺陷状态下电场和磁场的变化进行了仿真研究。文献[13]提出了一种基于损耗电流谐波的车载式电缆检测系统,该系统具有诊断速度快、准确性高的优点,但系统设备较为笨重,操作复杂,无法应用于狭窄的煤矿环境中。文献[14-16]提出了一种通过电流互感器采集谐波信号的诊断系统,通过Matlab编写的快速傅里叶变换(Fast Fourier Transform,FFT)程序对数据进行处理,得到谐波分量数据,该系统在测试中需要人工手动调节电桥平衡,工作效率低。文献[17]研制了一种测试装置并用于现场测试,该装置可通过连续检测形成电缆动态变化趋势,但检测精度有待进一步提高。

    针对现有电缆诊断系统存在的装置笨重、检测精确低、难以在煤矿应用的问题,提出一种基于电流谐波特征的矿用电缆劣化监测与故障诊断方法。首先,在线采集运行中的电缆谐波数据并进行小波变换处理,得到电缆中高次谐波的含量;然后,利用电缆故障特征向量对极限梯度提升树(XGBoost)模型进行训练;最后,通过构建的XGBoost模型对电缆劣化度进行实时监测和故障诊断。

    电缆运行过程中发生劣化后,其介质内部的磁偶极子会相应地发生改变,使得磁矩取向在电缆线芯电流磁场作用下重新排列,这种重新排列会在电流的高次谐波成分中体现出来。电缆异常状态下,介质内部磁束变化引起的涡电流是导致电缆电流中产生奇次谐波的主要原因,而机械振动等引起的涡电流是导致电缆电流中产生偶次谐波的主要原因,涡电流导致电缆发生局部过热现象,从而使电缆不同部位出现老化现象。谐波诊断技术根据上述原理对电流中的高次谐波成分进行分析,从而实现电缆运行状态监测和故障诊断[18]。电力电缆中的磁场$\varPhi $与电流$I $如图1所示。

    图  1  电力电缆中的磁场与电流
    Figure  1.  Magnetic field and current in cable

    XGBoost是一种使用提升框架合并模型的集成学习技术[19],其基础是梯度提升决策树(Gradient Boosting Decision Tree,GBDT)。与GBDT相比,XGBoost在目标函数上使用了2阶泰勒展开,可以保留更多的目标信息,提高了模型的准确性。对比其他回归预测模型,XGBoost模型在面对大量输入数据进行训练时,用时短,推理效率高,可以满足电缆故障实时诊断需求。

    遵循集成方法,XGBoost利用加法模型和前向分布算法,构建了一个具有多个分类和回归树(Classification and Regression Tree,CART)的集成树模型。对决策树进行评估,并选择最佳的决策树来预测目标值[20]

    设XGBoost由K个基模型组成,则有

    $$ {Y}_{i}=\sum _{k=1}^{K}{f}_{k}\left({d}_{i}\right) $$ (1)

    式中:Yi为第i个样本的预测值;fk为第k个基模型;di为第i个样本的故障特征。

    XGBoost的损失函数为

    $$ L=\sum _{i=1}^{n}l({y}_{i},{Y}_{i}) $$ (2)

    式中:n为样本总数;l为样本损失函数;yi为第 i个样本的真实值。

    XGBoost的目标函数为

    $$ O=L+\sum _{k=1}^{K}\varOmega \left({f}_{k}\right) $$ (3)

    式中:Ωfk)为正则项。

    $$ \varOmega \left({f}_{k}\right)=\gamma Q+\frac{1}{2}\lambda \sum _{j=1}^{Q}{\omega }_{j}^{2} $$ (4)

    式中:$\gamma $和λ为惩罚项;Q为决策树叶子的节点数目;ωj为节点j的权重。

    构建XGBoost模型[21],并对其进行训练和参数优化,构建流程如图2所示。

    图  2  XGBoost模型构建流程
    Figure  2.  XGBoost model construction process

    1) 提取电缆中的高次谐波含量信息,即故障特征向量信息。

    2) 对特征向量数据进行归一化处理。

    $$ x\left( {a,b} \right) = {\boldsymbol{X}}\left( {a,b} \right)/{\left| {\boldsymbol{X}}\left( {a,b} \right) \right|_{\max }} $$ (5)

    式中:$ x\left( {a,b} \right) $为归一化后的电缆高次谐波向量离散时间序列,$a$为谐波次数,$ b $为时间序列号;$ {\boldsymbol{X}}\left( {a,b} \right) $为电缆中的高次谐波向量;${\left| {\boldsymbol{X}}\left( {a,b} \right) \right|_{\max }} $为${\boldsymbol{X}}\left( {a,b} \right) $绝对值的最大值。

    3) 将归一化数据和已知的电缆故障劣化度数据导入XGBoost模型,形成训练样本集,进行模型训练。

    4) 根据模型评估函数优化XGBoost模型,得到最终的XGBoost模型[22]

    谐波信号采集电路结构如图3所示。谐波采集传感器进行电缆信号采集,然后对信号进行滤波、运放、AD转换和FFT处理,得到电流信号中的高次谐波成分[23]

    图  3  谐波信号采集电路结构
    Figure  3.  Structure of harmonic signal acquisition circuit

    高次谐波成分通过通信模块上传至故障诊断软件,对电缆的绝缘体、屏蔽层、保护层(简称主体部)和电缆接头(简称连接部)的劣化度进行计算,并与故障诊断专家数据库进行比较分析,最终获得电缆当前的运行状态。电缆故障诊断流程如图4所示。

    图  4  电缆故障诊断流程
    Figure  4.  Cable fault diagnosis process

    对采集的信号进行分解,得到2—10次谐波含量I2I10,计算总谐波失真率SS主要反映波形的畸变特性。

    $$ S = \frac{{\sqrt {I_2^2 + I_3^2 + \cdots + I_{10}^2} }}{{{I_1}}} \times 100{\text{%}} $$ (6)

    式中I1为基波。

    通过计算m次谐波含量Im与基波I1的比值,得到谐波含有率${H_m}$,再计算${H_m}$与$S$的比值,得到谐波指示值${Z_m}$。

    $$ {H_m} = \frac{{{I_m}}}{{{I_1}}} \times 100{\text{%}} $$ (7)
    $$ {Z_m} = \frac{{{H_m}}}{{S}} $$ (8)

    计算谐波指示值${Z_m}$与总谐波指示值Z0的比值,得到诊断计算值${C_m}$,将${C_m}$与m次谐波函数$ F({I_m}) $相乘,得到谐波判定值${P_m}$。

    $$ {C_m} = \frac{{{Z_m}}}{{{Z_0}}} $$ (9)
    $$ {Z_0} = {\sum _{m=2}^{10}}{{Z_m}}$$ (10)
    $$ {P_m} = {C_m} F({I_m}) $$ (11)

    当${P_m} \leqslant {Z_m}$时,说明m次谐波含量过大,对电缆正常运行产生了不利影响。计算m次谐波的故障贡献率:

    $$ {N_m} = \frac{{{H_m}}}{{{H_2} + {H_3} + \cdots + {H_{10}}}} \times 100{\text{%}} $$ (12)

    贡献率主要通过对前10次谐波进行主成分分析获得[24-26],见表1。将总谐波失真率与各次谐波贡献率等数据上传至专家系统,即可分析出电缆的劣化程度及劣化部位。

    表  1  矿用电缆劣化状态与高次谐波的关系
    Table  1.  Relationship between mining power cable degradation state and higher harmonics
    电力电
    缆部位
    劣化类型第一主成分
    谐波次数(贡献率)
    其他主成分
    谐波次数(贡献率)
    累计故障
    贡献率/%
    主体部绝缘体劣化初期劣化型3(41%),5(41%)4(6%),2(6%)94
    机械性损伤2(55%)4(16%),3(9%),5(6%)86
    电气性损伤5(59%)3(20%),4(8%),2(6%)93
    自然劣化型5(52%)3(28%),4(7%),2(6%)93
    屏蔽层劣化3(25%)5(24%),2(23%),4(18%)90
    保护层劣化2(39%)4(29%),3(10%),5(7%)85
    连接部发热7(53%)10(15%),9(11%),8(7%),6(5%)91
    污损8(35%)7(29%),9(13%),10(11%),6(7%)95
    龟裂9(33%)8(25%),7(21%),10(8%),6(5%)92
    变形10(30%)7(23%),8(17%),9(15%),6(6%)91
    下载: 导出CSV 
    | 显示表格

    选取30 000组相同功率电缆谐波诊断数据,将电缆主体部的2−5次谐波含量与其对应的贡献率相乘,得到4个谐波向量作为输入数据,通过XGBoost模型得出绝缘体、屏蔽层及保护层劣化度。将电缆连接部的7−10次谐波含量与其对应的贡献率相乘,得到4个谐波向量作为输入数据,通过XGBoost模型得出电缆接头劣化度。模型训练集部分主体部样本数据见表2,部分连接部样本数据见表3

    表  2  部分主体部样本数据
    Table  2.  Part of the main body sample data
    序号H2H3H4H5劣化度
    绝缘体屏蔽层保护层
    11.82.31.54.936.861.252.2
    22.42.11.45.337.854.147.7
    33.81.71.80.976.831.646.2
    43.42.11.42.463.049.949.9
    52.02.11.64.939.358.556.4
    62.91.31.22.375.043.867.6
    73.04.41.24.278.484.171.2
    83.06.01.02.178.295.754.0
    92.81.01.52.519.716.026.1
    102.81.51.00.569.041.448.7
    113.05.30.91.984.094.055.7
    123.01.41.75.949.742.970.2
    132.81.41.21.657.839.547.9
    142.41.11.11.944.638.146.8
    152.95.70.91.578.195.450.6
    下载: 导出CSV 
    | 显示表格
    表  3  部分连接部样本数据
    Table  3.  Part of the connection part sample data
    序号H7H8H9H10电缆接头
    劣化度
    11.20.40.40.482.6
    21.50.60.50.581.3
    31.20.40.50.778.8
    40.60.50.40.347.7
    50.70.50.40.446.2
    60.50.40.30.249.9
    70.70.40.30.256.4
    80.50.50.50.567.6
    90.80.40.40.471.2
    100.60.50.40.454.0
    110.50.40.40.246.1
    120.60.40.40.246.7
    130.60.40.40.355.7
    140.80.60.50.370.2
    150.70.40.40.347.9
    下载: 导出CSV 
    | 显示表格

    用电缆各次谐波与其对应的贡献率相乘后,计算各谐波向量的相对能量,最后得到影响电缆不同部位运行状态的谐波向量能量谱,如图5所示,各次谐波相对能量总和为1。

    图  5  电缆谐波向量能量谱
    Figure  5.  Energy spectrum of cable harmonic vector

    图5可看出,诊断电缆不同部位的运行状态时,谐波向量的相对能量明显不同:电缆绝缘体运行状态主要看2次谐波向量的变化;屏蔽层运行状态主要看2、3、5次谐波向量的变化;保护层运行状态主要看2、4次谐波向量的变化;电缆接头运行状态主要看7、8、9次谐波向量的变化。可以看出,得到的谐波向量完全表征了电缆不同部位的运行状态。

    取数据库中29940组数据对XGBoost模型进行训练,剩余60组数据作为测试集,最终电缆主体部和连接部的劣化度预测结果如图6图9所示。

    图  6  绝缘体劣化度预测结果
    Figure  6.  Prediction results of insulation degradation degree
    图  7  屏蔽层劣化度预测结果
    Figure  7.  Prediction results of shielding layer degradation
    图  8  保护层劣化度预测结果
    Figure  8.  Prediction results of degradation degree of protective layer
    图  9  电缆接头劣化度预测结果
    Figure  9.  Prediction results of cable joint deterioration

    选取决定系数R2为指标来反映模型的拟合优度,R2越接近1,表示其拟合的回归方程越优。选取均方误差(Mean-Square Error, MSE)、均方根误差(Root Mean Square Error,RMSE)、平均绝对百分比误差(Mean Absolute Percentage Error,MAPE)来评估模型预测精度,结果见表4

    表  4  电缆主体部和连接部预测精度评估参数
    Table  4.  Prediction accuracy evaluation parameters for cable main body and connection parts
    电缆R2${\rm{MSE}}$${\rm{MRSE}}$${\rm{MAPE}}$
    绝缘层0.93540.0018240.04220.0670
    屏蔽层0.92950.0007980.02820.0468
    保护层0.93850.0017360.04120.0607
    电缆接头0.95100.0009590.03100.0286
    下载: 导出CSV 
    | 显示表格

    表4可知,模型的拟合优度参数R2非常接近1,MSE、RMSE、MAPE均非常小,说明XGBoost模型的故障诊断准确性很高,具有较好的劣化趋势判断能力。

    为验证基于电流谐波特征的矿用电缆劣化监测与故障诊断方法的准确性及其在矿用电缆监测中的适用性,在淮南矿业集团潘东煤矿有限责任公司变电站内选取35根矿用电缆进行测试,电压等级为220 kV。部分高次谐波含有率见表5,电缆主体部运行状态实时数据、诊断报告、故障电缆如图10图12所示。

    表  5  部分高次谐波含有率
    Table  5.  Part of the high-order harmonic content
    序号H2H3H4H5H7H8H9H10时间
    11.51.11.20.90.50.40.40.32021−05−18
    21.41.31.41.10.50.40.30.42021−05−18
    31.61.11.50.90.70.60.40.32021−05−18
    41.51.21.41.30.60.30.30.22021−05−18
    53.81.51.51.20.80.50.20.12021−05−19
    62.41.41.41.30.60.50.40.32021−05−19
    73.81.81.81.20.70.40.30.22021−05−19
    83.42.11.41.00.90.50.50.32021−05−19
    93.32.11.61.10.60.30.40.22021−05−19
    下载: 导出CSV 
    | 显示表格
    图  10  电缆主体部运行状态实时数据
    Figure  10.  Real time data of the status of main body of the cable
    图  11  诊断报告
    Figure  11.  Diagnose report
    图  12  故障电缆
    Figure  12.  Faulty cable

    在监测35根电缆运行状态时,发现其中1根电缆B相的高次谐波含量异常,2次谐波含量较高,电缆主体部运行状态实时数据中绝缘体、屏蔽层、保护层的劣化度明显升高,而2次谐波含量的变化是导致绝缘体机械性劣化的主要参数指标,说明该电缆的绝缘体处于故障状态。经现场外观排查后,发现电缆的外护套有裂痕,验证了所提方法的准确性和实用性。

    在总结现有电缆谐波诊断技术不足的基础上,提出一种基于电流谐波特征的矿用电缆劣化监测与故障诊断方法。在线采集运行中的电缆谐波数据并进行小波变换处理,得到电缆中高次谐波的含量;利用电缆故障特征向量数据对XGBoost模型进行训练;通过构建的XGBoost模型对电缆劣化度进行实时监测和故障诊断。仿真结果表明:针对电缆不同部位提取的高次谐波向量的相对能量有明显不同,表明提取的高次谐波向量可表征电缆不同部位的运行状态;XGBoost模型的拟合优度参数R2高达 0.93,且误差较小。案例分析结果验证了基于电流谐波特征的矿用电缆劣化监测与故障诊断方法可对矿用电缆运行状态及劣化故障进行实时、准确的监测和诊断。

  • 图  1   煤层工作面瓦斯预抽钻孔精细化布控系统结构

    Figure  1.   Structure of refined control system for gas pre-extraction boreholes in coal seam working face

    图  2   One−hot编码

    Figure  2.   One-hot encoding

    图  3   半监督学习模型架构

    Figure  3.   Architecture of semi-supervised learning model

    图  4   SSGMM聚类结果

    Figure  4.   SSGMM clustering results

    图  5   SSK−Means聚类结果

    Figure  5.   SSK-Means clustering results

    图  6   SSGMM算法的钻孔评价结果

    Figure  6.   Evaluation results of boreholes using SSGMM algorithm

    图  7   SSK−Means算法的钻孔评价结果

    Figure  7.   Evaluation results of boreholes using SSK-Means algorithm

    图  8   试验现场

    Figure  8.   Test site

    图  9   SSGMM算法钻孔再评价结果

    Figure  9.   Re-evaluation results of boreholes using SSGMM algorithm

    图  10   SSK−Means算法钻孔再评价结果

    Figure  10.   Re-evaluation results of boreholes using SSK-Means algorithm

    表  1   钻孔抽采效果评价指标

    Table  1   Evaluation indicators for borehole and extraction performance

    因素名称 因素标号 单位 因素名称 因素标号 单位
    甲烷浓度 F1 % 抽采负压 F2 kPa
    环境温度 F3 抽采差压 F4 kPa
    环境压力 F5 kPa 工况流量 F6 m3/min
    瓦斯纯流量 F7 m3/min 前一天的瓦斯纯流量 F8 m3/min
    下载: 导出CSV

    表  2   AHP打分规则

    Table  2   AHP scoring rules

    x 意义 x 意义 x 意义
    1 相等重要 4 介于3,5中间 7 非常重要
    2 介于1,3中间 5 较为重要 8 介于7,9中间
    3 略微重要 6 介于5,7中间 9 最重要
    下载: 导出CSV

    表  3   评价指标的判断矩阵、权重及一致性检验

    Table  3   Judgment matrix, weights, and consistency test for evaluation indicators

    因素 F1 F2 F3 F4 F5 F6 F7 F8 $ {w_i} $ $ {\lambda _{{\text{max}}}} $ CR
    F1 1 3 8 4 5 2 1 1/3 0.191 8.433 0.043
    F2 1/3 1 5 2 3 1/3 1/3 1/5 0.057
    F3 1/8 1/5 1 1/4 1/3 1/6 1/8 1/9 0.008
    F4 1/4 1/2 4 1 2 1/3 1/5 1/7 0.043
    F5 1/5 1/3 3 1/2 1 1/4 1/5 1/6 0.028
    F6 1/2 3 6 3 4 1 1/3 1/5 0.131
    F7 1 3 8 5 5 3 1 1/2 0.192
    F8 3 5 9 7 6 5 2 1 0.251
    下载: 导出CSV

    表  4   随机一致性指标RI的取值

    Table  4   Value of random consistency index (RI)

    序号 1 2 3 4 5 6 7 8 9 10
    RI 0 0 0.58 0.9 1.12 1.24 1.32 1.41 1.45 1.49
    下载: 导出CSV

    表  5   聚类算法的评价指标及分类数

    Table  5   Evaluation indicators and classification numbers of clustering algorithms

    算法 MVCR/% ARI/% 分类个数
    簇1 簇2 簇3 簇4
    SSGMM 82.64 85.83 54 6 15 26
    SSK−Means 59. 10 71.27 69 6 7 19
    下载: 导出CSV

    表  6   煤层瓦斯预抽钻孔评价结果

    Table  6   Evaluation results of coal seam gas pre-extraction boreholes

    模型 钻孔编号
    SSGMM 425, 419, 414, 411, 410,
    400, 399, 391, 373,
    363, 361, 358, 359,
    357, 325
    424, 415, 408, 407, 406, 405, 404, 403, 402, 401, 397, 394, 389, 388, 387, 386, 384, 383, 382, 379, 378, 377, 375, 372, 371, 370, 368, 364, 362, 360, 356, 355, 354, 352, 351, 350, 349, 348, 347, 346, 345, 342, 341, 340, 339, 338, 337, 336, 335, 334, 333, 331, 330, 329 423, 422, 421, 420, 417, 416,
    413, 412, 409, 398, 393, 385,
    381, 380, 376, 374, 369, 367,
    366, 365, 353, 344, 343, 332,
    327, 326
    418, 396, 395, 392, 390, 328
    SSK−Means 425, 410, 400, 399,
    391, 363, 358
    424, 421, 419, 417, 415, 414, 411, 408, 407, 406, 405, 404, 403, 402, 401, 397, 394, 389, 388, 387, 386, 384, 383, 382, 379, 378, 377, 375, 373, 372, 371, 370, 368, 367, 364, 362, 361, 360, 359, 357, 356, 355, 354, 352, 351, 350, 349, 348, 347, 346, 345, 344, 343, 342, 341, 340, 339, 338, 337, 336, 335, 334, 333, 331, 330, 329, 327, 326, 325 423, 422, 420, 416, 413, 412,
    409, 398, 393, 385, 381, 380,
    376, 374, 369, 366, 365, 353,
    332
    418, 396, 395, 392, 390, 328
    下载: 导出CSV

    表  7   状态为“差”的钻孔缺陷

    Table  7   Borehole defects with the "poor" status

    孔号堵孔负压不合理漏气封孔不佳
    418
    396
    395
    392
    390
    328
    下载: 导出CSV
  • [1] 王德明,张伟,王和堂,等. 煤矿热动力重大灾害的不确定性风险特性研究[J]. 采矿与安全工程学报,2023,40(4):826-837.

    WANG Deming,ZHANG Wei,WANG Hetang,et al. Uncertain risk characteristics of major thermodynamic disasters in underground coal mines[J]. Journal of Mining & Safety Engineering,2023,40(4):826-837.

    [2] 林志军,李敏,贺珊,等. 基于博弈论−贝叶斯网络的煤矿瓦斯爆炸风险评估[J]. 煤炭学报,2024,49(8):3484-3497.

    LIN Zhijun,LI Min,HE Shan,et al. Risk assessment of gas explosion in coal mines based on game theory and Bayesian network[J]. Journal of China Coal Society,2024,49(8):3484-3497.

    [3]

    WEI Gaoming,WEN Hu,DENG Jun,et al. Liquid CO2 injection to enhance coalbed methane recovery:an experiment and in situ application test[J]. Fuel,2021,284. DOI: 10.1016/j.fuel.2020.119043.

    [4]

    MOHAMED T,MEHANA M. Coalbed methane characterization and modeling:review and outlook[J]. Energy Sources,Part A:Recovery,Utilization,and Environmental Effects,2025,47(1):2874-2896.

    [5]

    LI Tengteng,WU Bing,LEI Baiwei. Study on the optimization of a gas drainage borehole drainage horizon based on the evolution characteristics of mining fracture[J]. Energies,2019,12(23). DOI: 10.3390/en12234499.

    [6]

    TAO Shu,PAN Zhejun,TANG Shuling,et al. Current status and geological conditions for the applicability of CBM drilling technologies in China:a review[J]. International Journal of Coal Geology,2019,202:95-108. DOI: 10.1016/j.coal.2018.11.020

    [7]

    SONG Yawei,YANG Shengqiang,HU Xincheng,et al. Prediction of gas and coal spontaneous combustion coexisting disaster through the chaotic characteristic analysis of gas indexes in goaf gas extraction[J]. Process Safety and Environmental Protection,2019,129:8-16. DOI: 10.1016/j.psep.2019.06.013

    [8]

    ZHANG Yiwen,GUO Haishuai,LU Zhihui,et al. Distributed gas concentration prediction with intelligent edge devices in coal mine[J]. Engineering Applications of Artificial Intelligence,2020,92. DOI:10.1016/j.engappai. 2020.103643.

    [9]

    LIU Peng,FAN Long,LI Quangui,et al. Power ultrasound assisted coalbed methane enhancement recovery:field application and performance evaluation in underground coal mine[J]. Fuel,2022,324. DOI: 10.1016/j.fuel.2022.124575.

    [10]

    XIA Tongqiang,ZHOU Fubao,LIU Jishan,et al. Evaluation of the pre-drained coal seam gas quality[J]. Fuel,2014,130:296-305. DOI: 10.1016/j.fuel.2014.04.051

    [11]

    JIN Kan,CHENG Yuanping,WANG Wei,et al. Evaluation of the remote lower protective seam mining for coal mine gas control:a typical case study from the Zhuxianzhuang Coal Mine,Huaibei Coalfield,China[J]. Journal of Natural Gas Science and Engineering,2016,33:44-55. DOI: 10.1016/j.jngse.2016.05.004

    [12]

    XUE Yi,GAO Feng,GAO Yanan,et al. Quantitative evaluation of stress-relief and permeability-increasing effects of overlying coal seams for coal mine methane drainage in Wulan Coal Mine[J]. Journal of Natural Gas Science and Engineering,2016,32:122-137. DOI: 10.1016/j.jngse.2016.04.029

    [13]

    NI Xiaoming,TAN Xuebin,YANG Sen,et al. Evaluation of coal reservoir coalbed methane production potential in considering different coal structures in a coal seam section:a case study of the Shizhuang North Block in the Qinshui Basin[J]. Environmental Earth Sciences,2021,80. DOI: 10.1007/s12665-021-09979-8.

    [14]

    GUO Jianhong,ZHANG Zhansong,GUO Guangshan,et al. Evaluation of coalbed methane content by using kernel extreme learning machine and geophysical logging data[J]. Geofluids,2022(1). DOI: 10.1155/2022/3424367.

    [15]

    SONG Hongqing,DU Shuyi,YANG Jiaosheng,et al. Evaluation of hydraulic fracturing effect on coalbed methane reservoir based on deep learning method considering physical constraints[J]. Journal of Petroleum Science and Engineering,2022,212. DOI:10.1016/j.petrol. 2022.110360.

    [16]

    SOLEIMANI F,SI Guangyao,ROSHAN H,et al. Numerical modelling of gas outburst from coal:a review from control parameters to the initiation process[J]. International Journal of Coal Science & Technology,2023,10. DOI:10.1007/ s40789-023-00657-7.

    [17]

    LI Yunzhuo,SU Hetao,JI Huaijun,et al. Numerical simulation to determine the gas explosion risk in longwall goaf areas:a case study of Xutuan Colliery[J]. International Journal of Mining Science and Technology,2020,30(6):875-882. DOI: 10.1016/j.ijmst.2020.07.007

    [18]

    LIU Peng,FAN Jinyang,JIANG Deyi,et al. Evaluation of underground coal gas drainage performance:mine site measurements and parametric sensitivity analysis[J]. Process Safety and Environmental Protection,2021,148:711-723. DOI: 10.1016/j.psep.2021.01.054

    [19]

    LU Y Y,ZHANG H D,ZHOU Z,et al. Current status and effective suggestions for efficient exploitation of coalbed methane in China:a review[J]. Energy & Fuels,2021,35(11):9102-9123.

    [20]

    ZHOU Aitao,LI Jingwen,GONG Weili,et al. Theoretical and numerical study on the contribution of multi-hole arrangement to coalbed methane extraction[J]. Energy,2023,284. DOI: 10.1016/j.energy.2023.128598.

    [21]

    DU Feng,CUI Weilong,WANG Kai. Study on gas migration mechanism and multi-borehole spacing optimization in coal under negative pressure extraction[J]. Processes,2023,11(1). DOI: 10.3390/pr11010259.

    [22]

    LIU Jun,ZHANG Luwei,WANG Lin,et al. Negative pressure distribution of variable mass flow in coal mine drainage boreholes[J]. Energy Sources,Part A:Recovery,Utilization,and Environmental Effects,2025,47(1):1650-1667.

    [23]

    EL ISLEM KARABADJI N,AMARA KORBA A,ASSI A,et al. Accuracy and diversity-aware multi-objective approach for random forest construction[J]. Expert Systems with Applications,2023,225. DOI: 10.1016/j.eswa.2023.120138.

    [24]

    HUANG Qiuhao,WANG Chao,YE Ye,et al. Recognition of EEG based on improved black widow algorithm optimized SVM[J]. Biomedical Signal Processing and Control,2023,81. DOI: 10.1016/j.bspc.2022.104454.

    [25]

    NIE Feiping,CHEN Hong,XIANG Shiming,et al. On the equivalence of linear discriminant analysis and least squares regression[J]. IEEE Transactions on Neural Networks and Learning Systems,2024,35(4):5710-5720. DOI: 10.1109/TNNLS.2022.3208944

    [26] 马莲净,王颂,赵宝峰,等. 基于PCA−SSA−BPNN模型的矿井突水水源识别方法 [J/OL]. 采矿与安全工程学报:1-10[2024-06-18]. http://kns.cnki.net/kcms/detail/32.1760.TD.20231221.1014.002.html.

    MA Lianjing,WANG Song,ZHAO Baofeng,et al. A method for identifying the water inrush sources in mines based on PCA-SSA-BPNN model[J/OL]. Journal of Mining & Safety Engineering:1-10[2024-06-18]. http://kns.cnki.net/kcms/detail/32.1760.TD.20231221.1014.002.html.

    [27]

    AN Yiyao,ZHANG Ke,CHAI Yi,et al. Gaussian mixture variational-based transformer domain adaptation fault diagnosis method and its application in bearing fault diagnosis[J]. IEEE Transactions on Industrial Informatics,2024,20(1):615-625. DOI: 10.1109/TII.2023.3268750

    [28]

    NI Xiaoming,ZHAO Zheng,WANG Yanbin,et al. Optimisation and application of well types for ground development of coalbed methane from No. 3 coal seam in Shizhuang south block in Qinshui Basin,Shanxi Province,China[J]. Journal of Petroleum Science and Engineering,2020,193. DOI: 10.1016/j.petrol.2020.107453.

    [29]

    LOU Zhen,WANG Kai,KANG M,et al. Plugging methods for underground gas extraction boreholes in coal seams:a review of processes,challenges and strategies[J]. Gas Science and Engineering,2024,122. DOI:10.1016/j. jgsce.2024.205225.

    [30]

    YU Yang,ZENG Jun,ZHONG Lin,et al. Multi-views contrastive learning for dense text retrieval[J]. Knowledge-Based Systems,2023,274. DOI:10.1016/j.knosys.2023. 110624.

  • 期刊类型引用(3)

    1. 罗毅,黄毅文. 电气自动化系统中状态监测与故障诊断的实现方法研究. 中国新技术新产品. 2024(06): 31-34 . 百度学术
    2. 王楠,吕培庚,李怀政,陈志强,张勇,陶星宇,曹泰山,周勇军,叶炼. 深度残差网络下的电缆隧道低压故障检测. 人工智能科学与工程. 2024(02): 78-86 . 百度学术
    3. 杨晓超,郝慧良. 矿用电缆放电监测系统研究设计. 中国煤炭. 2024(S1): 406-410 . 百度学术

    其他类型引用(5)

图(10)  /  表(7)
计量
  • 文章访问数:  30
  • HTML全文浏览量:  2
  • PDF下载量:  4
  • 被引次数: 8
出版历程
  • 收稿日期:  2025-02-20
  • 修回日期:  2025-03-21
  • 网络出版日期:  2025-03-18
  • 刊出日期:  2025-03-14

目录

/

返回文章
返回