WSD-SVM在工作面底板破坏深度微震事件自动识别中的应用  

Application of WSD-SVM in micro-seismic events automatic recognition of the damage depth of working face floor

作　　者：樊鑫 赵晓光[1] 唐胜利[1] 解海军[1] 程建远 王云宏 王盼[1,2] FAN Xin;ZHAO Xiaoguang;TANG Shengli;XIE Haijun;CHENG Jianyuan;WANG Yunhong;WANG Pan(College of Geology and Environment,Xi’an University of Science and Technology,Xi’an 710054,China;Xi’an Research Institute(Group)Co.,Ltd.,China Coal Mine Technology and Engineering,Xi’an 710077,China)

机构地区：[1]西安科技大学地质与环境学院,陕西西安710054 [2]中煤科工西安研究院(集团)有限公司,陕西西安710077

出　　处：《西安科技大学学报》2023年第1期160-166,共7页Journal of Xi’an University of Science and Technology

基　　金：国家重点研发计划项目(2018YFC0807804)。

摘　　要：为解决煤矿微震事件识别中效率低、精度低、可靠性差的问题,将小波散射分解变换与支持向量机相结合,构建微震事件的WSD-SVM智能识别模型。首先,通过小波散射分解变换将微震监测数据分解成高、低频部分,并计算得到小波散射系数,构成散射特征矩阵;然后,选择70%的数据输入支持向量机模型进行训练,用得到的识别模型对其余30%的数据进行测试验证,获得识别结果。将山西保德煤矿某工作面微震监测时序数据作为实例,结果表明:WSD-SVM模型能够自动识别全部6个微震事件,用时1.651 s;而传统STA/LTA算法虽然仅用时0.731 s,但未能有效识别出其中的3个低信噪比事件,WSD-SVM模型的自动识别精度高于STA/LTA算法模型识别的精度,但需要较长的计算时长。小波散射分解变换方法的引入能够有效实现监测数据降维,大幅提高识别精度,为微震事件的自动识别提供了新思路。In order to solve the problems of low identification efficiency and accuracy of micro-seismic events and poor reliability in a coal mine, the wavelet scattering decomposition transform method and support vector machine are combined to construct a WSD-SVM auto recognition model.Firstly, the micro-seismic monitoring data are decomposed into high and low frequency parts through the wavelet scattering decomposition transform, and the wavelet scattering coefficients are calculated to achieve the scattering feature matrix;Then, 70% of the input data is selected for training SVM model, and the remaining 30% is tested and verified with the trained model to obtain the recognition results.The time sequence data of a working face of Baode coal mine in Shanxi Province is taken as an example.The results indicate that the model can automatically identify all the 6 microseismic events in the genuine signals, taking 1.651 s.The results indicate that the model can automatically identify all the 6 microseismic events in the genuine signals, taking 1.651 s.When the STA/LTA method is used, only 3 events can be recognized, taking 0.731 s, while the WSD-SVM model has a higher automatic recognition accuracy than the contrast model, requiring a larger time consumption.The introduction of the wavelet scattering decomposition transform method can effectively reduce the dimension of the monitoring data and greatly improve the identification accuracy, which provides a new idea for the automatic identification of micro-seismic events.

关 键 词：微震监测 小波散射分解 特征提取 支持向量机 自动识别 

分 类 号：TD163[矿业工程—矿山地质测量] TN911.7[电子电信—通信与信息系统]