中文摘要：

针对思山岭铁矿深埋厚大矿体所采用的阶段矿房嗣后充填采矿法的结构参数优化问题,基于正交试验方法设计了16种模拟方案,运用FLAC3D进行数值计算,建立了表征采场结构参数的5个因素与表征采场稳定性的8项指标之间的对应关系。将得到的16组对应关系作为SOM模型的输入数据集,经过无监督的竞争学习,建立SOM模型。基于所建立的SOM模型,完成了扩展方案的正向预测、结构参数反向预测以及影响采场稳定性的主因素分析等工作。研究结果表明：采场长度对采场稳定性影响最大,采场高度次之,其他因素影响相对较小。适合思山岭深埋矿体采矿方法的最佳采场参数为：段高100 m,采场长50 m,矿房和矿柱宽度15 m,尾砂充填。

英文摘要：

Taking the deep large-sized orebody of Sishanling iron mine as an example, the after-mining block caving filling method has been employed to obtain the optimal structural parameters. With the help of orthogonal experiment, 16 simulation schemes have been designed. Using the FLAC3 Dto conduct calculation, the interrelated correspondence has been achieved between 5 factors in stope structural parameters and 8 indexes for stop stability. Adopting the unsupervised and competitive study, 16 groups of scheme data have been applied to SOM model for data inputting. Therefore, based on SOM model, the analysis of the prediction for expansion schemes, reversal analysis for structural parameters and the stability-related factors have been done. The results have shown that the length of stope serves as the most influential factor to stability, then the height of stope less influential, while the other factors may not arouse much consideration. The optimal stope parameters for the deep and large-sized orebody in Sishanling iron ore are as follows：100 m level height, 50 m stope length, 15 m room and pillar width and using tailings for backfilling.