中文摘要：

水下矿床是难采矿体的一个重要方面,针对某矿山的实际情况,深入分析了该矿的地质资料后,在查阅相关文献的基础上,建立了该矿导水裂隙带高度的粗糙集—神经网络预测模型.在比较了粗糙集—神经网络预测结果、地质详查报告提供的结果及采用经验公式计算的结果后,认为神经网络预测的结果较准确,其结论在该矿水下开采设计中可以采用.在预测的导水裂隙带高度基础上,参考采矿设计手册中的经验公式,并类比其它矿山水下开采的情况,计算出了防水安全岩柱的厚度.此项研究为该矿重新编制开采设计方案和安全专篇提供了依据,同时为矿山安全管理提供了参考.

英文摘要：

Underwater deposit is the important aspect of difficult-to-mine ore bodies,according to the actual situation of a mine,on the basis of deep analysis of the mine＇s geological information and related literatures,a rough set and artificial neural network（RS-ANN）prediction model about height of water flowing fractured zone has been established.The results of RS-ANN prediction are more accurate after careful comparison of the results by neural network prediction,the results of detailed geological surveys and the results by empirical formula,and the neural network predicting results can be taken in the mine＇s underwater mining design.Based on the predicted heights of water flowing fractured zone,combined with the empirical formula in mining design manual and referenced to other mines＇ underwater mining conditions,the safe waterproof rock pillars＇ heights have been calculated.This research provides the basis for re-establishment of mining design plan and safety articles,at the same time it provides reference for this mine＇s safety management.