王莎，李坤，王雅莉，等. 上游式尾矿库浸润线观测值与库水位关联作用机制研究[J].河南理工大学学报(自然科学版)，doi:10.16186/j.cnki.1673-9787.2025030069

WANG S，LI K，WANG Y, et al．Investigation of the correlation mechanism between phreatic line and reservoir water level observations in upstream-style tailings dams［J］.Journal of Henan Polytechnic University( Natural Science)，doi:10.16186/j.cnki.1673-9787. 2025030069

上游式尾矿库浸润线观测值与库水位关联作用机制研究(网络首发)

王莎^1,2，李坤^1,2，王雅莉^1,2，楚一帆^1,2，何斌斌^1,2

（1.矿冶科技集团有限公司，北京 100160；2.国家金属矿绿色开采国际联合研究中心，北京 100160）

摘要: [目的] 由于浸润线的空间形成机理为库水位沿堆积坝入渗导致的水头逐渐递减形成，因此，受到库水位变幅的影响，浸润线与库水位观测值呈现一定的相关关系，但目前其相关作用机制尚不完全明确。[方法] 基于上述关键问题，采集了典型上游式尾矿库浸润线和库水位的原型观测数据序列，建立了相关统计回归模型，开展了浸润线与库水位绝对值、水头、时效之间的关联关系研究，采用拟合优度判定回归效果；同时，基于上述建立的数学统计预测模型，采用灰色关联分析理论对比不同因子的重要性，获得不同因子与浸润线观测值的灰色关联系数结果，进而获得不同因子的重要性排序，揭示了浸润线主要影响因素。[结果] 根据灰色关联分析结果可知，统计模型中所构建的各因子的空间变化趋势与因变量存在较强的依赖关系。统计分析结果可知，浸润线测值与水头实测值H因子、水头平方H2因子、水头3次方H3因子、前5、10、15、20、25、30日水头平均值及前5、10、15、20、25、30日平均水头上升速率因子均有关，统计回归结果具备较高的精度。[结论] 不同安设位置处的浸润线观测值的统计回归模型效果不尽相同，应总体而言，与水头实测值及变化速率有关，并与时效体现一定的相关性，随着时间序列拉长，与时效的相关性逐渐减弱。

关键词: 浸润线；库水位；统计回归模型；灰色关联分析；时效

中图分类号：TD76

doi: 10.16186/j.cnki.1673-9787.2025030069

基金项目: 国家重点研发计划（2023YFC3012200），矿冶科技集团有限公司集团青年基金（编号04-2532），矿冶科技集团探索基金项目（02-2271）

收稿日期：2025-03-30

修回日期：2025-06-12

网络首发日期：2025-06-13

Investigation of the correlation mechanism between phreatic line and reservoir water level observations in upstream-style tailings dams

Wang Sha^1,2，Li Kun^1,2，Wang Yali^1,2，Chu Yifan^1,2，He Binbin^1,2

(1. Beijing General Research Institute of Mining & Metallurg Beijing 100160, China；2. National Centre for International Research on Green Metal Mining, Beijing 100160, China)

Abstract: [Objective] The spatial formation mechanism of the phreatic line results from the gradual decrease of the water head caused by the infiltration of the reservoir water level along the accumulation dam. Therefore, affected by the amplitude of the reservoir water level variation, there is a certain correlation between the phreatic line and the observed values of the reservoir water level. However, the relevant interaction mechanism remains not fully understood at present. [Methods] Based on the above key issues, this study collected the prototype observation data sequences of the phreatic line and the reservoir water level of a typical upstream tailings dam. A relevant statistical regression model was constructed to conduct an in-depth exploration of the relationships among the phreatic line, the absolute value of the reservoir water level, the water head, and the time effect. The goodness of fit was employed to evaluate the regression effect. Meanwhile, based on the established mathematical statistical prediction model, the grey correlation analysis theory was utilized to compare the importance of different factors. The grey correlation coefficient results between different factors and the observed values of the phreatic line were obtained, and then the importance ranking of different factors was determined, revealing the main influencing factors of the phreatic line. [Results] According to the results of grey correlation analysis, the spatial variation trends of various factors in the statistical model exhibit a strong dependence on the dependent variable. The statistical analysis results indicate that the measured values of the phreatic line are related to the measured water head value (H factor), the square of the water head (H⟡ factor), the cube of the water head (H³ factor), the average water head values of the previous 5, 10, 15, 20, 25, and 30 days, and the average water head rise rate factors of the previous 5, 10, 15, 20, 25, and 30 days. The statistical regression results demonstrate high precision. [Conclusion] The effects of the statistical regression models for the observed values of the phreatic line at different installation positions vary. Generally speaking, they are related to the measured values of the water head and its rate of change, and show a certain correlation with the time effect. As the time series lengthens, the correlation with the time effect gradually weakens.

Key words: phreatic line; reservoir water level; statistical regression model; grey correlation analysis; time effect

CLC: