Research on fault diagnosis method of ventilation network based on machine learning
-
摘要: 机器学习算法通过对已知数据的学习来预测未知数据，现有通风系统故障诊断方法大多针对1种机器学习算法进行研究，无法保证所选算法为最优。针对该问题，对8种机器学习算法进行比较，并选择支持向量机(SVM)、随机森林和神经网络3种算法进行通风网络故障诊断研究。根据矿井通风系统实际布局，按照几何相似、运动相似、动力相似准则构建通风网络管道模型，得到由管道网络分支和管道网络节点组成的通风网络，通过实验获取风量数据，并采用标准化方法对数据进行预处理；通过交叉验证和网格搜索对基于SVM、随机森林、神经网络的通风网络故障诊断模型进行参数寻优。实验及现场测试结果表明，基于SVM、随机森林、神经网络的通风网络故障诊断模型在实验平台测试集上的准确率分别为0.89，0.88和0.95，在煤矿现场测试集上的准确率分别为0.86，0.90和0.96，神经网络模型的故障诊断效果均为最佳。将煤矿现场收集的120组新风量数据输入神经网络模型进行预测，故障诊断准确率达0.98，验证了基于神经网络的通风网络故障诊断模型的可行性和准确性。Abstract: The machine learning algorithm predicts unknown data by learning known data. Most of the existing fault diagnosis methods of ventilation system focus on a machine learning algorithm, which can not guarantee the selected algorithm to be optimal. In order to solve this problem, eight machine learning algorithms are compared, and three algorithms, support vector machine ( SVM), random forest and neural network, are selected to study the fault diagnosis of ventilation network. According to the actual layout of the mine ventilation system, a ventilation network pipeline model is constructed according to the criteria of geometric similarity, motion similarity and dynamic similarity. A ventilation network consisting of pipeline network branches and pipeline network nodes is obtained, and air volume data is obtained through experiments, and the data is preprocessed by a standardized method. Through cross-validation and grid search, the parameters of ventilation network fault diagnosis model based on SVM, random forest and neural network are optimized. The results of experiment and field test show that the accuracy of ventilation network fault diagnosis model based on SVM, random forest and neural network are 0.89, 0.88 and 0.95 respectively on the test set of experimental platform, and 0.86, 0.90 and 0.96 respectively on the test set of coal mine field. The neural network model has the best fault diagnosis effect. 120 sets of fresh air volume data collected in coal mine field are input into neural network model for prediction, and the fault diagnosis accuracy rate reaches 0.98, which verifies the feasibility and accuracy of the ventilation network fault diagnosis model based on neural network.
-
Keywords:
- mine ventilation /
- fault diagnosis /
- machine learning /
- support vector machine /
- random forest /
- neural network /
- cross-validation /
- grid search
-
-
![]()
图 1 基于机器学习的通风网络故障诊断方法流程
Figure 1. Flow of fault diagnosis method of ventilation network based on machine learning
![]()
图 2 通风网络故障诊断实验平台
Figure 2. Experimental platform of fault diagnosis of ventilation network
![]()
图 6 SVM模型交叉验证平均分数热力图
Figure 6. The heat map of cross-validation average score of SVM model
![]()
图 7 随机森林模型交叉验证平均分数热力图
Figure 7. The heat map of cross-validation average score of random forest model
![]()
图 8 神经网络模型交叉验证平均分数热力图
Figure 8. The heat map of cross-validation average score of neural network model
![]()
图 9 故障诊断模型在各分支上的预测准确率
Figure 9. Prediction accuracy of fault diagnosis model on each branch
表 1 8种机器学习算法比较
Table 1 Comparison of eight machine learning algorithms
机器学习算法 优点 缺点 最近邻 适用于小型低维空间数据集，容易解释 用于大型数据集时表现不佳 线性模型 训练和预测速度快 用于低维空间分类时受限 朴素贝叶斯 适用于不确定性问题 精度低于线性模型 决策树 速度快，不需要进行数据缩放 容易过拟合 随机森林 可降低过拟合，不需要进行数据缩放 用于高维稀疏数据时表现不佳 梯度提升决策树 不需要进行数据缩放 用于高维稀疏数据时表现不佳，且训练速度慢 SVM 用于中等数据集时性能强 需要进行数据缩放，对参数敏感 神经网络 可构建非常复杂的模型，预测能力强 对数据缩放和参数选取敏感 
下载: 导出CSV
表 2 部分风量数据
Table 2 Part of the air volume data
m3/min 序号 故障分支 风量1 风量2 风量3 ··· 风量16 风量17 风量18 1 e8 639.54 634.95 288.16 ··· 259.48 257.34 1315.92 2 e10 672.61 667.72 557.35 ··· 166.60 165.23 1376.53 3 e15 660.15 655.44 229.89 ··· 207.10 205.30 1353.86 4 e16 666.42 661.72 208.44 ··· 187.69 186.16 1365.37 5 e55 655.63 650.81 244.23 ··· 219.92 218.11 1345.45 298 e10 652.34 647.71 226.33 ··· 228.46 226.58 1339.56 299 e15 658.53 653.84 235.05 ··· 211.65 209.91 1350.90 300 e55 660.93 655.31 228.81 ··· 206.05 204.35 1354.46
下载: 导出CSV
表 3 故障诊断模型准确率比较
Table 3 Comparison of accuracy of fault diagnosis models
故障诊断模型 最优参数 准确率 训练集 测试集 SVM C=104
γ=10−10.99 0.89 随机
森林p=15
q=40.96 0.88 神经
网络t=14
α=10−50.96 0.95
下载: 导出CSV
表 4 3种故障诊断模型准确率
Table 4 Accuracy of three fault diagnosis models
故障诊断
模型准确率 训练集 测试集 SVM 0.97 0.86 随机森林 0.93 0.90 神经网络 0.98 0.96
下载: 导出CSV
表 5 故障位置诊断结果统计
Table 5 Statistics of fault location diagnosis results
故障位置 样本
个数正确
个数错误
个数准确率 FC−2−2−001 40 39 1 0.98 FC−2−2−002 40 40 0 1.00 FC−2−2−003 40 39 1 0.98
下载: 导出CSV
-
[1] 张瑞新,毛善君,赵红泽,等. 智慧露天矿山建设基本框架及体系设计[J]. 煤炭科学技术,2019,47(10):1-23. ZHANG Ruixin,MAO Shanjun,ZHAO Hongze,et al. Framework and structure design of system construction for intelligent open-pit mine[J]. Coal Science and Technology,2019,47(10):1-23.
[2] 王国法,杜毅博. 智慧煤矿与智能化开采技术的发展方向[J]. 煤炭科学技术,2019,47(1):1-10. WANG Guofa,DU Yibo. Development direction of intelligent coal mine and intelligent mining technology[J]. Coal Science and Technology,2019,47(1):1-10.
[3] 周福宝,魏连江,夏同强,等. 矿井智能通风原理、关键技术及其初步实现[J]. 煤炭学报,2020,45(6):2225-2235. ZHOU Fubao,WEI Lianjiang,XIA Tongqiang,et al. Principle,key technology and preliminary realization of mine intelligent ventilation[J]. Journal of China Coal Society,2020,45(6):2225-2235.
[4] 王国法,王虹,任怀伟,等. 智慧煤矿2025情景目标和发展路径[J]. 煤炭学报,2018,43(2):295-305. WANG Guofa,WANG Hong,REN Huaiwei,et al. 2025 scenarios and development path of intelligent coal mine[J]. Journal of China Coal Society,2018,43(2):295-305.
[5] 张庆华,姚亚虎,赵吉玉. 我国矿井通风技术现状及智能化发展展望[J]. 煤炭科学技术,2020,48(2):97-103. ZHANG Qinghua,YAO Yahu,ZHAO Jiyu. Status of mine ventilation technology in China and prospects for intelligent development[J]. Coal Science and Technology,2020,48(2):97-103.
[6] 宁剑,任怡睿,林济铿,等. 基于人工智能及信息融合的电力系统故障诊断方法[J]. 电网技术,2021,45(8):2925-2936. NING Jian,REN Yirui,LIN Jikeng,et al. Power system fault diagnosis based on artificial intelligence and information fusion[J]. Power System Technology,2021,45(8):2925-2936.
[7] 陈瑞, 杨春曦, 翟持, 等. 特征加权的高斯加权K近邻−支持向量机的水泵故障诊断方法[J/OL]. 机械科学与技术: 1-8[2022-03-08]. DOI: 10.13433/j.cnki.1003-8728.20200358. CHEN Rui, YANG Chunxi, ZHAI Chi, et al. A weighted GWKNN-SVM algorithm for fault diagnosis of water pumps[J/OL]. Mechanical Science and Technology for Aerospace Engineering: 1-8[2022-03-08]. DOI: 10.13433/j.cnki.1003-8728.20200358.
[8] 程晓之,王凯,郝海清,等. 矿井局部通风智能调控系统及关键技术研究[J]. 工矿自动化,2021,47(9):18-24. CHENG Xiaozhi,WANG Kai,HAO Haiqing,et al. Research on intelligent regulation and control system and key technology of mine local ventilation[J]. Industry and Mine Automation,2021,47(9):18-24.
[9] 熊中杰,邱颖宁,冯延晖,等. 基于机器学习的风电机组变桨系统故障研究[J]. 太阳能学报,2020,41(5):85-90. XIONG Zhongjie,QIU Yingning,FENG Yanhui,et al. Fault analysis of wind turbine pitch system based on machine learning[J]. Acta Energiae Solaris Sinica,2020,41(5):85-90.
[10] 刘剑,郭欣,邓立军,等. 基于风量特征的矿井通风系统阻变型单故障源诊断[J]. 煤炭学报,2018,43(1):143-149. LIU Jian,GUO Xin,DENG Lijun,et al. Resistance variant single fault source diagnosis of mine ventilation system based on air volume characteristic[J]. Journal of China Coal Society,2018,43(1):143-149.
[11] 刘剑,刘丽,黄德,等. 基于风量−风压复合特征的通风系统阻变型故障诊断[J]. 中国安全生产科学技术,2020,16(1):85-91. LIU Jian,LIU Li,HUANG De,et al. Resistance variant fault diagnosis of ventilation system based on composite features of air volume and air pressure[J]. Journal of Safety Science and Technology,2020,16(1):85-91.
[12] 刘剑,蒋清华,刘丽,等. 矿井通风系统阻变型故障诊断及风速传感器位置优化[J]. 煤炭学报,2021,46(6):1907-1914. LIU Jian,JIANG Qinghua,LIU Li,et al. Resistance variant fault diagnosis of mine ventilation system and position optimization of wind speed sensor[J]. Journal of China Coal Society,2021,46(6):1907-1914.
[13] 周启超,刘剑,刘丽,等. 基于SVM的通风系统故障诊断惩罚系数与核函数系数优化研究[J]. 中国安全生产科学技术,2019,15(4):45-51. ZHOU Qichao,LIU Jian,LIU Li,et al. Research on fault fiagnosis penalty coefficient and kernel function coefficient optimization of ventilation system based on SVM[J]. Journal of Safety Science and Technology,2019,15(4):45-51.
[14] 黄德,刘剑,刘永,等. 矿井通风阻变故障观测特征组合选择试验研究[J]. 煤炭学报,2021,46(12):3922-3933. HUANG De,LIU Jian,LIU Yong,et al. Experimental research on combination selection of observation feature of resistance variation fault in mine ventilation[J]. Journal of China Coal Society,2021,46(12):3922-3933.
[15] 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.
[16] 汤志立,徐千军. 基于9种机器学习算法的岩爆预测研究[J]. 岩石力学与工程学报,2020,39(4):773-781. TANG Zhili,XU Qianjun. Rockburst prediction based on nine machine learning algorithms[J]. Chinese Journal of Rock Mechanics and Engineering,2020,39(4):773-781.
-
期刊类型引用(15)
1. 李冠华，屈世甲，冯黎莉. 瓦斯爆炸冲击波作用下人工坝体动力响应特征. 煤矿安全. 2025(02): 40-46 . 
百度学术
2. 邱常青. 余吾煤业岗位风险辨识研究与应用. 煤. 2024(03): 55-58+87 . 百度学术
3. 刘波，吴远巍，边昊. 煤炭企业安全风险动态管控体系质量评价模型. 中国煤炭. 2024(08): 43-49 . 百度学术
4. 穆熙军. 山西柳林县煤矿安全生产风险防控策略分析. 煤矿现代化. 2023(02): 22-25 . 百度学术
5. 李光宇，李守军，缪燕子. 基于机器视觉和灰色模型的矿井外因火灾辨识与定位方法. 矿业安全与环保. 2023(02): 82-87 . 百度学术
6. 王玲. “目视化”管理在煤矿安全风险管控中的应用. 江西煤炭科技. 2023(03): 250-253 . 百度学术
7. 白春玉. 基于风险管控的煤矿安全综合防控体系建设研究. 内蒙古煤炭经济. 2023(13): 91-93 . 百度学术
8. 刘光. 基于精益生产的采煤机综合效率提升研究. 价值工程. 2023(25): 55-58 . 百度学术
9. 王必强. 煤矿安全风险管控机制应用分析. 露天采矿技术. 2023(06): 116-119 . 百度学术
10. 李宏，马金魁，田坤云，杨皓. 瓦斯抽采钻孔封孔作业风险辨识与评估. 中国安全科学学报. 2023(12): 85-91 . 百度学术
11. 陈小林. 采煤工作面火灾风险阶段性多指标综合评价方法. 工矿自动化. 2022(07): 90-95+104 . 本站查看
12. 陈小林. 智能化矿山建设背景下的煤矿监管监察模式. 煤矿安全. 2022(08): 237-241 . 百度学术
13. 姜红年，陈小林，王超，胡然，张羽. 基于风险监测预警的非煤矿山安全监管监察模式. 工矿自动化. 2022(08): 114-121 . 本站查看
14. 邵常雄，高强，高珂. 基于“互联网+智能”的多信息融合综合管控平台的建设. 电子技术与软件工程. 2022(20): 23-27 . 百度学术
15. 何璐翔. 煤矿风险监测预警体系构建及系统应用. 山西焦煤科技. 2022(11): 41-44 . 百度学术
其他类型引用(2)
计量
- 文章访问数: 347
- HTML全文浏览量: 83
- PDF下载量: 42
- 被引次数: 17
