煤炭工程 ›› 2024, Vol. 56 ›› Issue (12): 133-139.doi: 10.11799/ce202412020

• 研究探讨 • 上一篇    下一篇

基于多元统计和混合模型的矿井涌水构成比例判别

曹思文,刘源,武亚遵,等   

  1. 1. 兖矿能源集团股份有限公司杨村煤矿
    2. 河南理工大学资源环境学院
  • 收稿日期:2024-05-21 修回日期:2024-08-12 出版日期:2024-12-20 发布日期:2025-01-08
  • 通讯作者: 刘源 E-mail:lY992884916@163.com

Proportional discrimination of mine water inrush composition based on multivariate statistics and mixed model

  • Received:2024-05-21 Revised:2024-08-12 Online:2024-12-20 Published:2025-01-08
  • Contact: Yuan Liu E-mail:lY992884916@163.com

摘要: 为了快速有效地判别涌(突)水水源,以杨村煤矿为研究实例,在分析矿井充水水源的基础上,在四个主要充水含水层中选定56个学习样本,采用Piper三线图剔除异常样本后保留52个典型样本,并选取HCO-3、SO2-4、Cl-、Na++K+、Mg2+和Ca2+六大常规离子作为判别指标,应用多元统计学中的Bayes逐步判别分析法建立矿井突水水源判别模型,Bayes判别模型对样本总体分类的准确率达到了91.06%,为有效地识别混合水水源构成比例奠定了基础,而后结合“M3”质量平衡理论建立了混合模型并确定了矿井涌水的构成比例。结果显示:受充水含水层特性及影响程度制约,前期矿井涌水构成中砂岩水含量较高,具体为奥灰水占15.2%,十下灰水占20.68%,3煤顶板砂岩水占48.28%,十三-十四灰水占15.85%。随时间的推移十三-十四灰水、奥灰水比例逐渐增大,其中奥灰水占23.65%,十下灰水占10.8%,3煤顶板砂岩水占26.73%,十三-十四灰水占38.83%。这与矿井前期主要回采五采区3号煤,后期逐步开采下组煤的情况相一致,表明该方法较符合生产实际。从整体上来看杨村煤矿矿井涌水中奥灰水占17.5%,十下灰水占19.86%,3煤顶板砂岩水占36.43%,十三-十四灰水占26.21%,表明当前矿井涌水以3号煤回采区的顶板砂岩水和下组煤回采区的十三-十四灰水为主,这为今后矿井防治水工作的开展提供了依据。

关键词: 涌水来源, Bayes模型, 混合模型, 比例判别, 质量平衡理论

Abstract: Rapid and effective identification of mine inrush water sources is a prerequisite for mine water control measures development, while the accurate quantification of different sources of water influx, and accordingly formulate targeted prevention and control measures is the key to ensure the safe production of the mine. Therefore, using Yangcun coal mine as a research example, and based on the analysis of mine water filling source, 56 study samples selected in four major recharge aquifers, 52 typical samples were retained after removing abnormal samples using Piper's trilinear plot. Selection of six conventional ions as discriminators which are HCO3-、SO42-、Cl-、Na++K+、Mg2+ and Ca2+, and Bayes stepwise discriminant analysis method in multivariate statistics was used to establish the discriminant model of mine water inrush source. The accuracy of Bayes discriminant model for the overall classification of samples reached 91.06 %. It lays a foundation for effective identification the proportion of mixed water source composition. Afterwards, a mixing model was developed with the "M3" mass balance theory and the composition of the mine water influx was determined. The results show that: restricted by the characteristics and influence degree of the water-filled aquifer, the sandstone water content in the early mine water inrush composition is higher, specifically, the Ordovician limestone water accounts for 15.2 %, the Shixia limestone water accounts for 20.68 %, the No.3 coal roof sandstone water accounts for 48.28 %, and the No.13-14 limestone water accounts for 15.85 %. With the passage of time, the proportion of 13-14 ash water and Ordovician ash water gradually increased, among which Ordovician ash water accounted for 23.65 %, 10 ash water accounted for 10.8 %, 3 coal roof sandstone water accounted for 26.73 %, 13-14 ash water was 38.83 %. This is consistent with the situation that the No.3 coal in the fifth mining area is mainly mined in the early stage of the mine, and the lower coal group is gradually mined in the later stage, indicating that the method is more in line with the actual production. On the whole, the Ordovician limestone water accounts for 17.5 %, the Shixia limestone water accounts for 19.86 %, the No.3 coal roof sandstone water accounts for 36.43 %, and the No.13-14 limestone water accounts for 26.21 %. It shows that the current mine inrush water is dominated by the roof sandstone water in the No.3 mining area and the No.13-14 limestone water in the lower group of mining area, which provides a basis for the future development of water control work in the mine.

中图分类号: