关键词: 微震监测, 岩爆, 深部开采, 模式识别, 时间非线性

Abstract: The value of ground stress was measured by the method of surface strain during deep excavation in Hongtoushan Copper Mine to investigate the impact acting on the rockmass in the depths of the mine. A 3-D geological model was developed for the No.13 stope at the elevation of -767 m, where the sensors were arranged properly to form a microseismic monitoring system so as to obtain continuously the rule of the activity of deep ground pressure and seize the information microseismic activities due to ground pressure, thus locating the microseismic source. Then, the interfering signals were rejected in the seized information by the pattern recognition based on neural network, such as the signals of bursting, rock drilling and power supply. The efficient locating result by microseismic monitoring is thus shown, and the cause of different microseismic events can be distinguished with the magnitudes of released microseismic energies. Furthermore, it is probable to predict the rockburst in accordance to the number of microseismic events within a unit time that has been set via the time-dependent nonlinear analysis of the data of microseismic events monitored continuously, which is of actual importance to the prediction of dynamic hazard/disaster in the mine.