Feature extraction and classification method of coal gangue acoustic signal during top coal caving
-
摘要: 针对综放工作面煤矸智能识别问题,设计了能感应尾梁动作并自动触发数据采集的放煤声信号采集装置,在山东能源古城煤矿3106综放工作面采集现场数据并进行人工标注,构建了放顶煤声信号分类样本库。研究了6种常用机器学习分类方法在时域、频域和时频域中的特征,以及在不同帧长、不同特征向量维度下的分类效果。结果表明:在不同帧长下,基于时频域特征的分类效果最稳定、准确率最高,随机森林、K近邻、决策树、多层感知器模型分类准确率均达到80 % 以上,其中基于小波包分解与随机森林算法的分类器性能最好,分类准确率为93.06 %。维度较高的时频域特征向量之间存在相关性,降维可以提取少量的综合特征并降低系统的运算量,利用主成分分析法将时频域特征向量降维至20后,分类准确率进一步提高至94.51 %。Abstract: To achieve the intelligent recognition of coal gangue in fully mechanized caving face, a coal caving acoustic signal collection device is designed, which can sense the movement of the tail beam and automatically trigger data collection.Field data is collected at the fully mechanized caving face 3106 of Gucheng Coal Mine, Shandong Energy, and it is manually labeled to construct a sample library of acoustic signal classification for top coal caving.Then, six machine learning classification methods are applied in the time domain, frequency domain and time-frequency domain, and the classification effect of them are evaluated by different frame lengths and different feature vector dimensions.The results show that: the classification effect based on time-frequency domain features is the most stable, and its accuracy rate is the highest by different frame lengths.The classification accuracy rate of random forest, K-nearest neighbor, decision tree and multi-layer perceptron model is above 80 %.Among them, the classifier performance based on wavelet packet decomposition and random forests are the best, and the classification accuracy is 93.06 %.There is a correlation between the time-frequency domain feature vectors and higher dimensions.Through dimensionality reduction, a small number of comprehensive features can be extracted and the amount of system calculations can be reduced.The principal component analysis is used to reduce the time-frequency domain feature vector to 20.Thus, the classification accuracy rate is further improved to 94.51 %.
-
表 1 放顶煤声信号的时域特征及其计算公式
Table 1. Time-domain characteristics of caving coal acoustic signal and its calculation formula
特征类型 计算公式 有效值 ${X_{{\rm{rms}}}} = \sqrt {\frac{1}{N}\sum\limits_{n = 1}^N x {{(n)}^2}} $ 标准差 $X_{\mathrm{std}}=\sqrt{\frac{1}{N} \sum\limits_{n=1}^{N}[x(n)-\bar{X}]^{2}}$ 偏度 $X_{\text {ske }}=\frac{1}{N} \sum\limits_{n=1}^{N}[x(n)-\bar{X}]^{3}$ 峭度 $X_{\mathrm{kur}}=\frac{1}{N} \sum\limits_{n=1}^{N}[x(n)-\bar{X}]^{4}$ 峰峰值 $X_{p-p}=\max [x(n)]-\min [x(n)] $ 波形因数 $C F=\frac{|\max [x(n)]|}{X_{\mathrm{rms}}}$ 脉冲因数 $SF = \frac{{{X_{{\rm{rms}}}}}}{{\frac{1}{N}\sum\limits_{n = 1}^N | x(n)|}}$ 峰值因数 $IF = \frac{{|\max [x(n)]|}}{{\frac{1}{N}\sum\limits_{n = 1}^N | x(n)|}}$ 表 2 放顶煤声信号的频域特征及其计算公式
Table 2. Frequency-domain characteristics of caving coal acoustic signal and its calculation formula
特征类型 计算公式 重心频率 ${{X_{{\rm{FC}}}} = \frac{{\sum\limits_{k = 1}^N k X(k)}}{{\sum\limits_{k = 1}^N X (k)}}}$ 均方频率 ${{X_{{\rm{MSF}}}} = \frac{{\sum\limits_{k = 1}^N {{k^2}} X(k)}}{{\sum\limits_{k = 1}^N X (k)}}}$ 频率方差 ${X_{{\rm{VF}}}} = \frac{{\sum\limits_{k = 1}^N {\left({k - {X_{{\rm{FC}}}}} \right)} X(k)}}{{\sum\limits_{k = 1}^N X (k)}}$ 表 3 多种分类模型的参数
Table 3. Parameters of multiple classification models
分类模型 SVM RF KNN NBC DT MLP 识别时间/ms 3.150 0 2.922 9 2.890 8 2.980 1 2.650 9 3.049 7 模型大小/MB 2.945 0 6.602 0 6.975 0 0.003 0 0.053 0 0.008 0 -
[1] 张守祥, 张学亮, 刘帅, 等. 智能化放顶煤开采的精确放煤控制技术[J]. 煤炭学报, 2020, 45(6): 2008-2020. https://www.cnki.com.cn/Article/CJFDTOTAL-MTXB202006010.htmZhang Shouxiang, Zhang Xueliang, Liu Shuai, et al. Intelligent precise control technology of fully mechanized top coal caving face[J]. Journal of China Coal Society, 2020, 45(6): 2008-2020. https://www.cnki.com.cn/Article/CJFDTOTAL-MTXB202006010.htm [2] 杨大鹏. 新形势下采矿工程专业人才培养对策研究[J]. 矿业科学学报, 2020, 5(3): 320-324. https://www.cnki.com.cn/Article/CJFDTOTAL-KYKX202003010.htmYang Dapeng. Research on countermeasures of mining engineering talent training in the new situation[J]. Journal of Mining Science and Technology, 2020, 5(3): 320-324. https://www.cnki.com.cn/Article/CJFDTOTAL-KYKX202003010.htm [3] 孙继平. 基于图像识别的煤岩界面识别方法研究[J]. 煤炭科学技术, 2011, 39(2): 77-79. https://www.cnki.com.cn/Article/CJFDTOTAL-MTKJ201102021.htmSun Jiping. Study on identified method of coal and rock interface based on image identification[J]. Coal Science and Technology, 2011, 39(2): 77-79. https://www.cnki.com.cn/Article/CJFDTOTAL-MTKJ201102021.htm [4] 王卫东, 张康辉, 吕子奇, 等. 基于深度学习的煤中异物机器视觉检测[J]. 矿业科学学报, 2021, 6(1): 115-123. https://www.cnki.com.cn/Article/CJFDTOTAL-KYKX202101013.htmWang Weidong, Zhang Kanghui, Lü Ziqi, et al. Machine vision detection of foreign objects in coal using deep learning[J]. Journal of Mining Science and Technology, 2021, 6(1): 115-123. https://www.cnki.com.cn/Article/CJFDTOTAL-KYKX202101013.htm [5] 张帆, 栾佳星, 崔东林, 等. 基于SSD-LeNet的矿井移动目标检测与识别方法[J]. 矿业科学学报, 2021, 6(1): 100-108. https://www.cnki.com.cn/Article/CJFDTOTAL-KYKX202101011.htmZhang Fan, Luan Jiaxing, Cui Donglin, et al. SSD-LeNet based method of mine moving target detection and recognition[J]. Journal of Mining Science and Technology, 2021, 6(1): 100-108. https://www.cnki.com.cn/Article/CJFDTOTAL-KYKX202101011.htm [6] 李亮, 王昕, 胡克想, 等. 探地雷达探测煤岩界面的方法与试验[J]. 工矿自动化, 2015, 41(9): 8-11. https://www.cnki.com.cn/Article/CJFDTOTAL-MKZD201509003.htmLi Liang, Wang Xin, Hu Kexiang, et al. Coal-rock interface detection method using ground penetrating radar and its experiment[J]. Industry and Mine Automation, 2015, 41(9): 8-11. https://www.cnki.com.cn/Article/CJFDTOTAL-MKZD201509003.htm [7] 王增才, 张秀娟, 张怀新, 等. 自然γ射线方法检测放顶煤开采中的煤矸混合度研究[J]. 传感技术学报, 2003, 16(4): 442-446. doi: 10.3969/j.issn.1004-1699.2003.04.015Wang Zengcai, Zhang Xiujuan, Zhang Huaixin, et al. The research on detection of rock content in coal rock mixture in top coal caving by natural gamma ray[J]. Journal of Transcluction Technology, 2003, 16(4): 442-446. doi: 10.3969/j.issn.1004-1699.2003.04.015 [8] 张宁, 任茂文, 刘萍. 基于主成分分析和BP神经网络的煤岩界面识别[J]. 工矿自动化, 2013, 39(4): 55-58. https://www.cnki.com.cn/Article/CJFDTOTAL-MKZD201304018.htmZhang Ning, Ren Maowen, Liu Ping. Recognition of coal-rock interface based on principal component analysis and BP neural network[J]. Industry and Mine Automation, 2013, 39(4): 55-58. https://www.cnki.com.cn/Article/CJFDTOTAL-MKZD201304018.htm [9] Liu W, He K, Liu C Y, et al. Coal-gangue interface detection based on Hilbert spectral analysis of vibrations due to rock impacts on a longwall mining machine[J]. Proceedings of the Institution of Mechanical Engineers, Part C: Journal of Mechanical Engineering Science, 2015, 229(8): 1523-1531. doi: 10.1177/0954406214543409 [10] 张艳丽, 张守祥. 基于EMD方法的煤岩界面识别研究[J]. 煤炭技术, 2007, 26(9): 49-51. doi: 10.3969/j.issn.1008-8725.2007.09.028Zhang Yanli, Zhang Shouxiang. Identification of coal and rock interface based on emepirical mode decomposition[J]. Coal Technology, 2007, 26(9): 49-51. doi: 10.3969/j.issn.1008-8725.2007.09.028 [11] 薛光辉, 柳二猛, 赵新赢, 等. 基于声压信号时域特征的综放工作面煤岩性状识别方法研究[J]. 煤炭工程, 2015, 47(6): 119-122. https://www.cnki.com.cn/Article/CJFDTOTAL-MKSJ201506039.htmXue Guanghui, Liu Ermeng, Zhao Xinying, et al. Research of coal-rock character recognition in fully mechanized caving face based on acoustic pressure data time domain feature[J]. Coal Engineering, 2015, 47(6): 119-122. https://www.cnki.com.cn/Article/CJFDTOTAL-MKSJ201506039.htm [12] 张艳丽, 张守祥, 矫林岳, 等. 独立分量分析法在综采煤岩界面识别中的应用[J]. 煤炭科学技术, 2007, 35(8): 22-24, 28. doi: 10.3969/j.issn.0253-2336.2007.08.005Zhang Yanli, Zhang Shouxiang, Jiao Linyue, et al. Application on ICA in coal and rock interface identification of the fully mechanized mining[J]. Coal Science and Technology, 2007, 35(8): 22-24, 28. doi: 10.3969/j.issn.0253-2336.2007.08.005 [13] 汪玉凤, 夏元涛, 王晓晨. 含噪超完备独立分量分析在综放煤岩识别中的应用[J]. 煤炭学报, 2011, 36(S1): 203-206. https://www.cnki.com.cn/Article/CJFDTOTAL-MTXB2011S1041.htmWang Yufeng, Xia Yuantao, Wang Xiaochen. Application on overcomplete ICA with noise in coal and rock identification of fully mechanized mining[J]. Journal of China Coal Society, 2011, 36(S1): 203-206. https://www.cnki.com.cn/Article/CJFDTOTAL-MTXB2011S1041.htm [14] 马瑞, 王增才, 王保平. 基于声波信号小波包变换的煤矸界面识别研究[J]. 煤矿机械, 2010, 31(5): 44-46. doi: 10.3969/j.issn.1003-0794.2010.05.018Ma Rui, Wang Zengcai, Wang Baoping. Coal-rock interface recognition based on wavelet packet transform of acoustic signal[J]. Coal Mine Machinery, 2010, 31(5): 44-46. doi: 10.3969/j.issn.1003-0794.2010.05.018 [15] 薛光辉, 赵新赢, 柳二猛, 等. 基于振动信号时域特征的综放工作面煤岩识别[J]. 煤炭科学技术, 2015, 43(12): 92-97. https://www.cnki.com.cn/Article/CJFDTOTAL-MTKJ201512019.htmXue Guanghui, Zhao Xinying, Liu Ermeng, et al. Time-domain characteristic extraction of coal and rock vibration signal in fully-mechanized top caving face[J]. Coal Science and Technology, 2015, 43(12): 92-97. https://www.cnki.com.cn/Article/CJFDTOTAL-MTKJ201512019.htm [16] Xu J, He Y, Zhang W, et al. Coal-rock interface recognition based on MFCC and neural network[J]. International Journal of Signal Processing, Image Processing & Pattern Recognition, 2013, 32(4): 586-590. http://connection.ebscohost.com/c/articles/90328052/coal-rock-interface-recognition-based-mfcc-neural-network [17] Zhang G X, Wang Z C, Zhao L, et al. Coal-rock recognition in top coal caving using bimodal deep learning and Hilbert-Huang transform[J]. Shock and Vibration, 2017, 2017: 1-13. http://www.onacademic.com/detail/journal_1000040468278110_cb8b.html [18] 陈喆, 王荣, 周文颖, 等. 非平稳信号度量方法综述[J]. 数据采集与处理, 2017, 32(4): 667-683. https://www.cnki.com.cn/Article/CJFDTOTAL-SJCJ201704003.htmChen Zhe, Wang Rong, Zhou Wenying, et al. Review on measurement parametrics and methods for nonstationary signal[J]. Journal of Data Acquisition and Processing, 2017, 32(4): 667-683. https://www.cnki.com.cn/Article/CJFDTOTAL-SJCJ201704003.htm [19] 程正兴. 小波分析算法与应用[M]. 西安: 西安交通大学出版社, 1998. [20] 杨国安, 钟秉林, 黄仁, 等. 机械故障信号小波包分解的时域特征提取方法研究[J]. 振动与冲击, 2001, 20(2): 25-28, 31. https://www.cnki.com.cn/Article/CJFDTOTAL-ZDCJ200102007.htmYang Guoan, Zhong Binglin, Huang Ren, et al. Research on the extraction method of time-domain symptoms based on wavelet packet decompositions of mechanical fault signal[J]. Journal of Vibration and Shock, 2001, 20(2): 25-28, 31. https://www.cnki.com.cn/Article/CJFDTOTAL-ZDCJ200102007.htm [21] Li Y P, Xie X J, Yang X H, et al. Diagnosis of early gastric cancer based on fluorescence hyperspectral imaging technology combined with partial-least-square discriminant analysis and support vector machine[J]. Journal of Biophotonics, 2019, 12(5): e201800324. http://www.onacademic.com/detail/journal_1000041572544999_9893.html [22] Dudani S A. The distance-weighted k-nearest-neighbor rule[J]. IEEE Transactions on Systems, Man, and Cybernetics, 1976, 6(4): 325-327. http://www.onacademic.com/detail/journal_1000035163913710_a7d6.html [23] Peng F C, Schuurmans D, Wang S J. Augmenting naive Bayes classifiers with statistical language models[J]. Information Retrieval, 2004, 7(3/4): 317-345. doi: 10.1023/B:INRT.0000011209.19643.e2 [24] Quinlan J R. Induction of decision trees[J]. Machine Learning, 1986, 1(1): 81-106. [25] 雷震. 随机森林及其在遥感影像处理中应用研究[D]. 上海: 上海交通大学, 2012. [26] 李彦冬, 郝宗波, 雷航. 卷积神经网络研究综述[J]. 计算机应用, 2016, 36(9): 2508-2515, 2565. https://www.cnki.com.cn/Article/CJFDTOTAL-JSJY201609029.htmLi Yandong, Hao Zongbo, Lei Hang. Survey of convolutional neural network[J]. Journal of Computer Applications, 2016, 36(9): 2508-2515, 2565. https://www.cnki.com.cn/Article/CJFDTOTAL-JSJY201609029.htm [27] Wold S, Esbensen K, Geladi P. Principal component analysis[J]. Chemometrics and Intelligent Laboratory Systems, 1987, 2(1/2/3): 37-52.