论文部分内容阅读
岩体破裂产生的应力波触发微震事件,传统的互相关匹配难以识别微震事件深浅类型。本文利用谱矩心和线性度提取地震波的特征信息,并通过多层感知器网络进行分类识别,将微震事件分为浅源和深源。实验结果表明该方法的识别准确率为86.14%,谱矩心对微震事件分类精度高于线性度,且精度均高于传统的互相关方法。该方法不仅可以识别地震波形,也可为岩爆、滑坡等动力灾害监测提供预警信息。
The stress wave generated by the rupture of a rock mass triggers a microseismic event, which is difficult to identify by the traditional cross-correlation matching. In this paper, the characteristic information of seismic waves is extracted by using the centroid and linearity of the spectrum, and the multi-layer perceptron network is used for classification and recognition. The microseismic events are divided into shallow and deep sources. Experimental results show that the recognition accuracy of the proposed method is 86.14%, and the classification accuracy of spectral centroid for microseismic events is higher than that of linearity, and the accuracy is higher than the traditional cross-correlation method. The method can not only identify seismic waveforms, but also provide early warning information for dynamic disaster monitoring such as rock burst and landslide.