基于Logistic回归的大采深厚煤层冲击地压预警

Rock burst early-warning for thick coal seam in deep mining based on Logistic regressio

  • 摘要: 针对大采深厚煤层冲击地压监测预警难题,分别建立了基于常规指标(支护阻力、钻孔应力和瓦斯浓度)及常规指标(支护阻力、钻孔应力和瓦斯浓度)与地球物理指标(微震平均震源距、微震日能量、微震日频次、电磁辐射强度)综合的Logistic回归冲击地压预警模型,得到了冲击地压危险性(发生概率)与监测指标之间的定量表达式,并结合千秋煤矿现场实测数据对冲击地压预警模型进行验证。研究结果表明,基于综合指标Logistic回归的冲击地压预警模型预测准确率达89.2%,其拟合优度和预测准确率均优于基于常规指标Logistic回归的冲击地压预警模型。

     

    Abstract: For difficult rock burst monitoring and early-warning for thick coal seam in deep mining, a rock burst early-warning model based on Logistic regression was established by use of conventional indexes (support resistance, borehole stress and gas concentration) as well as the one by use of comprehensive indexes including the conventional indexes (support resistance, borehole stress and gas concentration) and geophysical indexes (average focal distance, daily pulses and daily energy of microseism, and intensity of electromagnetic radiation). Quantitative expression between probability of rock burst occurrence and comprehensive indexes was obtained. Finally, the rock burst early-warning models were tested by use of measured data in Qianqiu Coal Mine. The research results show that forecasting accuracy rate of the rock burst early-warning model based on Logistic regression by use of comprehensive indexes achieves 89.2%, whose goodness of fit and forecasting accuracy rate is higher than the ones of the model by use of the conventional indexes.

     

/

返回文章
返回