102 / 2021-09-29 17:50:42

Seismic Monitoring and Seismic Data Clustering of mine safety pillar within EPOS PL+ project

seismic analysis,rock burst,Clustering,seismic hazard

Seismicity induced by mining exploitation is one of the most dangerous physical phenomena in underground mines and often leads to the destruction of mine workings. Local seismic systems installed in Polish underground coal mines are used routinely for monitoring seismic events with seismic energy range between 102J and 109J. These systems are able to locate accurately both low and high energy seismic tremors providing detailed information of seismicity distribution in space and time. The SOS - Seismic Observation System developed by Laboratory of Mining Geophysics at Central Mining Institute has also been installed in Marcel coal mine in the area of safety pillar. Seismic events that occurred in the area of this safety pillar between 2009-2021 were located and analyzed within the EPOS PL+ project to prepare seismic data catalogue as part of its research infrastructure. The recorded mining induced seismicity can be grouped to form spatial clusters of seismic events related to the geological and mining structures in mines, such as faults and pillars. However, typical seismicity in coal mines is mainly related to the current mining process and there is no distinct spatial or temporal group of seismic events, but rather they form a continuous cloud of events related to mining process. Therefore, for the clustering of such event clouds, a new time-space metric was introduced along with the Ward's minimal variance method, and a hierarchical induced seismicity cluster analysis was performed. The introduced time-space metric allowed for the simultaneous assessment of the seismic hazard in time and space, linking individual seismic clusters with a group of cracks dynamically formed during mining operations. The applied method of hierarchical clustering was used to study the seismicity induced in the Marcel Coal Mine with high seismic hazard in the area of the safety pillar. As a result, 20 separate clusters of seismic events correlated with the mining process were obtained. In each cluster, the energy distribution of seismic activity was calculated, showing, in some cases, a significant deviation from the Gutenberg-Richter distribution, which may be important in the seismic hazard analysis. As a measure of the accuracy of the location of each seismic cluster specially designed bootstrap algorithm was introduced. In this way, it was shown that errors in the location of seismic phenomena affect the location of the center of each cluster. The presented analysis of induced seismicity as groups of clusters relates them with groups of seismic sources and areas of high stress in the rock mass, which may potentially allow for determination of seismic and rock burst hazards.