Coal mine gas and coal dust explosion sound recognition method based on wavelet packet energy
-
摘要: 针对目前煤矿瓦斯和煤尘爆炸监测漏报率和误报率高，难以满足瓦斯和煤尘爆炸事故应急救援需求的问题，提出了一种基于小波包能量的煤矿瓦斯和煤尘爆炸声音识别方法。在煤矿井下重点监测区域安装矿用拾音器，实时采集煤矿井下设备工作声音及环境音等；通过小波包分解提取声音的小波包能量占比，构成表征声音信号的特征向量；将特征向量输入BP神经网络中，训练得到煤矿瓦斯和煤尘爆炸声音识别模型；提取待测声音信号的小波包能量占比，并构成特征向量输入模型中,识别待测声音信号的类型。根据特征向量和输出结果要求，建立了输入层、隐含层和输出层节点数分别为8，8，1的BP神经网络用于识别模型的训练；通过分析煤矿井下声音信号小波包分解结果，确立了采用Haar小波函数，选择小波包分解层数为3。实验结果表明：瓦斯和煤尘爆炸声音通过小波包分解后的能量占比与其他声音差异明显，且不同时长的同一声音信号的小波包能量占比分布稳定，因此小波包能量占比可有效表征声音信号特征，且具有较强的鲁棒性；BP神经网络训练速度快，仅需较少的训练迭代次数就能达到期望误差，且在煤矿井下众多干扰声音信号存在的情况下识别准确率达95%，与极限学习机模型、支持向量机模型相比，BP神经网络识别效果最优。Abstract: At present, it is difficult to meet the emergency rescue needs of gas and coal dust explosion accidents due to the rate of missing and false alarms in coal mine gas and coal dust explosion monitoring. In order to solve the above problems, a coal mine gas and coal dust explosion sound recognition method based on wavelet packet energy is proposed. This method installs mine-used pickups in the key monitoring areas of the coal mine to collect the working sound and environmental sound of the coal mine equipment in real-time. The wavelet packet energy ratio of sound is extracted through wavelet packet decomposition, and the feature vector characterizing the sound signal is formed. The feature vector is input into the BP neural network to obtain the sound recognition model of coal mine gas and coal dust explosion. The wavelet packet energy ratio of the sound signal to be measured is extracted and input into the model as the feature vector to recognize the type of sound signal to be measured. According to the requirements of feature vectors and output results, a BP neural network with 8, 8 and 1 nodes in the input layer, hidden layer and output layer is established to train the recognition model. By analyzing the results of wavelet packet decomposition of underground acoustic signals in coal mines, it is confirmed that the Haar wavelet function is used and the number of wavelet packet decomposition layers is chosen to be 3. The experimental results show that the energy proportion of gas and coal dust explosion sound decomposed by wavelet packet is obviously different from other sounds. The wavelet packet energy proportion distribution of the same sound signal with different time is stable. Therefore, the wavelet packet energy proportion can effectively represent the features of the sound signal and has strong robustness. BP neural network training speed is fast, and only a small number of training iterations can achieve the expected error. The recognition accuracy is up to 95% in the presence of many disturbing sound signals in the coal mine. BP neural network has the best recognition effect compared with the extreme learning machine model and support vector machine model.
-
图 1 基于小波包能量的煤矿瓦斯和煤尘爆炸声音识别方法原理
Figure 1. Principle of coal mine gas and coal dust explosion sound recognition method based on wavelet packet energy
图 2 经小波包分解的声音高频分量小波包能量占比差值分布
Figure 2. Distribution of wavelet packet energy proportion difference of high frequency sound component decomposed by wavelet packet
图 3 基于Haar小波函数的声音信号小波包分解结果及小波包系数分布
Figure 3. Wavelet packet decomposition results and wavelet packet coefficient distribution of sound signals based on Haar wavelet function
图 4 基于db4小波函数的声音信号小波包分解结果及小波包系数分布
Figure 4. Wavelet packet decomposition results and wavelet packet coefficient distribution of sound signals based on db4 wavelet function
图 5 不同时长下声音小波包能量占比分布
Figure 5. Wavelet packet energy proportion distribution of sound under different time
表 1 煤矿井下声音小波包能量占比分布
Table 1. Wavelet packet energy proportion distribution of sound in underground coal mine
% 声音 能量占比 d1 d2 d3 d4 d5 d6 d7 d8 瓦斯爆炸 87.280 8.081 1.013 2.581 0.230 0.539 0.048 0.228 煤尘爆炸 90.100 6.316 0.810 1.969 0.185 0.411 0.037 0.173 采煤机 99.798 0.151 0.036 0.004 0.009 0.001 0 0 刮板输送机 96.802 2.366 0.548 0.110 0.134 0.025 0.008 0.006 转载机 94.162 4.021 0.836 0.567 0.198 0.119 0.037 0.060 破碎机 97.182 1.961 0.403 0.260 0.096 0.053 0.016 0.029 乳化液泵 95.135 3.106 0.690 0.602 0.162 0.131 0.080 0.093 掘进机 94.002 3.944 0.792 0.749 0.181 0.149 0.077 0.105 锚杆机 57.115 9.755 8.843 7.940 2.054 2.898 7.320 4.075 风镐 37.388 21.513 10.640 13.989 1.537 3.236 6.919 4.778 馈电开关设备 99.801 0.140 0.036 0.007 0.009 0.003 0.003 0.002 高爆开关设备 98.624 1.029 0.251 0.022 0.062 0.006 0.003 0.002 移动变电站 99.199 0.604 0.147 0.009 0.037 0.002 0.001 0.001 通风机 85.903 9.983 2.096 1.112 0.507 0.237 0.055 0.108 水泵 91.574 5.867 1.235 0.733 0.287 0.152 0.069 0.083 胶带 90.865 6.463 1.290 0.780 0.310 0.165 0.046 0.079 胶轮车 98.605 1.053 0.255 0.017 0.063 0.004 0.001 0 
下载: 导出CSV
表 2 不同模型识别结果
Table 2. Recognition results of different models
% 模型 识别率 召回率 精确率 BP神经网络模型 95 75 100 SVM模型 91 69 100 ELM模型 84 20 100
下载: 导出CSV
-
[1] 孙继平. 煤矿瓦斯和煤尘爆炸感知报警与爆源判定方法研究[J]. 工矿自动化,2020,46(6):1-5,11. doi: 10.13272/j.issn.1671-251x.17617SUN Jiping. Research on method of coal mine gas and coal dust explosion perception alarm and explosion source judgment[J]. Industry and Mine Automation,2020,46(6):1-5,11. doi: 10.13272/j.issn.1671-251x.17617 [2] 孙继平,余星辰. 基于声音识别的煤矿重特大事故报警方法研究[J]. 工矿自动化,2021,47(2):1-5,44. doi: 10.13272/j.issn.1671-251x.17715SUN Jiping,YU Xingchen. Research on alarm method of coal mine extraordinary accidents based on sound recognition[J]. Industry and Mine Automation,2021,47(2):1-5,44. doi: 10.13272/j.issn.1671-251x.17715 [3] 孙继平,余星辰. 基于CEEMD分量样本熵与SVM分类的煤矿瓦斯和煤尘爆炸声音识别方法[J]. 采矿与安全工程学报,2022,39(5):1061-1070. doi: 10.13545/j.cnki.jmse.2022.0073SUN Jiping,YU Xingchen. Sound recognition method of coal mine gas and coal dust explosion based on CEEMD component sample entropy and SVM classification[J]. Journal of Mining & Safety Engineering,2022,39(5):1061-1070. doi: 10.13545/j.cnki.jmse.2022.0073 [4] 孙继平,余星辰. 基于声音特征的煤矿瓦斯和煤尘爆炸识别方法[J]. 中国矿业大学学报,2022,51(6):1096-1105. doi: 10.13247/j.cnki.jcumt.001451SUN Jiping,YU Xingchen. Recognition method of coal mine gas and coal dust explosion based on sound characteristics[J]. Journal of China University of Mining & Technology,2022,51(6):1096-1105. doi: 10.13247/j.cnki.jcumt.001451 [5] 彭佑多,谢伟华,郭迎福,等. 矿井掘进工作面粉尘对机器噪声衰减的影响[J]. 湖南科技大学学报(自然科学版),2012,27(1):23-29. doi: 10.3969/j.issn.1672-9102.2012.01.005PENG Youduo,XIE Weihua,GUO Yingfu,et al. Studies on the spread and attenuation of machine noise influenced by the heading face of mine roadway dust[J]. Journal of Hunan University of Science and Technology(Natural Science Edition),2012,27(1):23-29. doi: 10.3969/j.issn.1672-9102.2012.01.005 [6] 李端玲,成苈委,于功敬,等. 融合小波包和神经网络的脑电信号处理方法[J]. 北京邮电大学学报,2021,44(3):94-99.LI Duanling,CHENG Liwei,YU Gongjing,et al. An electroencephalogram signal processing method fusing wavelet packet and neural network[J]. Journal of Beijing University of Posts and Telecommunications,2021,44(3):94-99. [7] 李浩,毋文峰,蒲云,等. 基于小波包能量谱的装备悬臂梁结构损伤诊断[J]. 火力与指挥控制,2021,46(1):177-181. doi: 10.3969/j.issn.1002-0640.2021.01.031LI Hao,WU Wenfeng,PU Yun,et al. Cantilever beam structure damage diagnosis based on weaponry wavelet packet energy spectrum[J]. Fire Control & Command Control,2021,46(1):177-181. doi: 10.3969/j.issn.1002-0640.2021.01.031 [8] 郭健,钟昊荪. 基于实测数据的斜拉索振动分析与小波包能量占比研究[J]. 桥梁建设,2021,51(3):25-31. doi: 10.3969/j.issn.1003-4722.2021.03.004GUO Jian,ZHONG Haosun. Vibration analysis and wavelet packet energy ratio of stay cable based on measured data[J]. Bridge Construction,2021,51(3):25-31. doi: 10.3969/j.issn.1003-4722.2021.03.004 [9] 王伟,李兴华,陈作彬,等. 基于小波包变换的爆破振动信号能量熵特征分析[J]. 爆破器材,2019,48(6):19-23. doi: 10.3969/j.issn.1001-8352.2019.06.004WANG Wei,LI Xinghua,CHEN Zuobin,et al. Characteristic analysis of energy entropy of blasting vibration signal based on wavelet packet transform[J]. Explosive Materials,2019,48(6):19-23. doi: 10.3969/j.issn.1001-8352.2019.06.004 [10] 梁凯,韩庆邦. 小波包能量谱和BP神经网络在波纹管压浆超声检测中的应用[J]. 声学技术,2020,39(2):151-156.LIANG Kai,HAN Qingbang. Application of wavelet packet energy spectrum and BP neural network to ultrasonic detection of slurry in bellows[J]. Technical Acoustics,2020,39(2):151-156. [11] 郭飞,张培伟,张大海,等. 基于小波包能量特征向量的光纤布拉格光栅低速冲击定位[J]. 振动与冲击,2017,36(8):184-189. doi: 10.13465/j.cnki.jvs.2017.08.029GUO Fei,ZHANG Peiwei,ZHANG Dahai,et al. Localization of low-velocity impact by using fiber Bragg grating sensors based on wavelet packet energy eigenvector[J]. Journal of Vibration and Shock,2017,36(8):184-189. doi: 10.13465/j.cnki.jvs.2017.08.029 [12] 陈石,张兴敢. 基于小波包能量熵和随机森林的级联H桥多电平逆变器故障诊断[J]. 南京大学学报(自然科学),2020,56(2):284-289.CHEN Shi,ZHANG Xinggan. Fault diagnosis for cascaded H-bridge multilevel inverter based on wavelet packet energy entropy and random forest[J]. Journal of Nanjing University(Natural Science),2020,56(2):284-289. [13] 安春兰,甘方成,罗微,等. 提速道岔小波包能量熵故障诊断方法[J]. 铁道科学与工程学报,2015,12(2):269-274. doi: 10.3969/j.issn.1672-7029.2015.02.008AN Chunlan,GAN Fangcheng,LUO Wei,et al. Method of speed-up turnout fault diagnosis using wavelet packet energy entropy[J]. Journal of Railway Science and Engineering,2015,12(2):269-274. doi: 10.3969/j.issn.1672-7029.2015.02.008 [14] 陈琳,陈静,王惠民,等. 基于小波包能量熵的电池剩余寿命预测[J]. 电工技术学报,2020,35(8):1827-1835. doi: 10.19595/j.cnki.1000-6753.tces.190094CHEN Lin,CHEN Jing,WANG Huimin,et al. Prediction of battery remaining useful life based on wavelet packet energy entropy[J]. Transactions of China Electrotechnical Society,2020,35(8):1827-1835. doi: 10.19595/j.cnki.1000-6753.tces.190094 [15] 肖佳辰,卢超,林俊明,等. 基于小波包能量比变化率偏差的复合材料层板空气耦合超声概率损伤成像[J]. 复合材料学报,2021,38(8):2635-2645.XIAO Jiachen,LU Chao,LIN Junming,et al. Air coupled ultrasonic probabilistic damage imaging of composite laminates based on wavelet packet energy relative variation deviation[J]. Acta Materiae Compositae Sinica,2021,38(8):2635-2645. [16] 李伟,黄焱. 基于峰度检验和小波包分解的海洋平台脉冲噪声处理方法研究[J]. 振动与冲击,2021,40(6):220-226,242. doi: 10.13465/j.cnki.jvs.2021.06.029LI Wei,HUANG Yan. Impulse noise processing for an offshore platform based on kurtosis detection and wavelet packet decomposition[J]. Journal of Vibration and Shock,2021,40(6):220-226,242. doi: 10.13465/j.cnki.jvs.2021.06.029 [17] 杨婷婷,李岩,林雪琦. 基于车辆制动激励和小波包能量分析的连续梁桥基础冲刷识别方法[J]. 中国公路学报,2021,34(4):51-60. doi: 10.3969/j.issn.1001-7372.2021.04.004YANG Tingting,LI Yan,LIN Xueqi. Foundation scour identification method based on vehicle braking excitation and wavelet packet energy analysis for continuous beam bridges[J]. China Journal of Highway and Transport,2021,34(4):51-60. doi: 10.3969/j.issn.1001-7372.2021.04.004 [18] 齐添添,陈尧,何才厚,等. 损伤声发射信号小波包神经网络特征识别方法[J]. 北京邮电大学学报,2021,44(1):124-130. doi: 10.13190/j.jbupt.2020-118QI Tiantian,CHEN Yao,HE Caihou,et al. A wavelet packet neural network feature recognition method for damage acoustic emission signals[J]. Journal of Beijing University of Posts and Telecommunications,2021,44(1):124-130. doi: 10.13190/j.jbupt.2020-118 [19] LIU Yanbing,DHAKAL S,HAO Binyao,et al. Coal and rock interface identification based on wavelet packet decomposition and fuzzy neural network[J]. Journal of Intelligent & Fuzzy Systems,2020,38(4):3949-3959. [20] HE Qingbo. Vibration signal classification by wavelet packet energy flow manifold learning[J]. Journal of Sound and Vibration,2013,332(7):1881-1894. doi: 10.1016/j.jsv.2012.11.006 [21] 闫晓玲,董世运,徐滨士. 基于最优小波包Shannon熵的再制造电机转子缺陷诊断技术[J]. 机械工程学报,2016,52(4):7-12. doi: 10.3901/JME.2016.04.007YAN Xiaoling,DONG Shiyun,XU Binshi. Flaw diagnosis technology for remanufactured motor rotor based on optimal wavelet packet Shannon entropy[J]. Journal of Mechanical Engineering,2016,52(4):7-12. doi: 10.3901/JME.2016.04.007 -
点击查看大图
图(6) / 表(2)
计量
- 文章访问数: 179
- HTML全文浏览量: 39
- PDF下载量: 19
- 被引次数: 0
