杏山铁矿冲击地压预测的优化GM(1,1)模型  

Optimized GM(1,1) Model for Prediction of Rock Burst in Xingshan Iron Mine

作　　者：孙世国[1] 苗子臻 龚之淇 魏帅颖 田波[1] 

机构地区：[1]北方工业大学土木工程学院,北京100144

出　　处：《金属矿山》2017年第1期165-168,共4页Metal Mine

基　　金：国家自然科学基金项目(编号:41172250);"十二五"国家科技支撑计划项目(编号:2012BAK09B06);北京市创新团队提升计划项目(编号:IDHT20140501);北京市科研基地建设-科研创新平台;科研专项(编号:XN083)

摘　　要：冲击地压作为矿山开采中最严重的自然灾害之一,给矿山生产安全及作业人员生命安全带来了严重威胁。为对冲击地压进行实时有效的预测及防治,首先在非等间隔序列GM(1,1)模型的基础上,通过残差修正的方法对模型进行了优化,然后基于残差优化模型,结合现场监测数据,进行了杏山铁矿冲击地压预测的研究。研究结果表明:相比修正前模型,残差优化模型预测精度更高、预测效果更优;对于杏山铁矿-143 m水平工作面,随时间推移,围岩应力缓慢增加,但近期仍不会达到预警临界值,无冲击地压危险。Rock burst is one of the most serious natural disasters in mine, which brings a serious threat to the safety of mine production and the workers. To realize the prediction and prevention of rock burst in real time effectively, firstly, the non- equal interval sequence GM （ 1,1 ） model was optimized by residual error correction method. Then, based on the residual opti- mization model, and combined with monitoring data, the rock burst prediction of Xingshan Iron Mine was researched. The re- suits showed that, compared with the initial model, residual optimization models have higher prediction accuracy and better pre- diction effect;For the -143 m level working face of Xingshan Iron Mine,with the passage of time, the surrounding rock stress increased slowly, but it will not reach the critical value of early-warning in the near future. So there is no danger of rock burst.

关 键 词：非等间隔序列 GM(1 1)模型 残差修正 冲击地压 

分 类 号：TD822[矿业工程—煤矿开采]