多因素影响下的充填体强度预测与分析  

Prediction and Analysis of Backfill Strength Under the Influence of Multiple Factors

作　　者：刘明利[1] 申康[1] 刘钟涛[1] LIU Mingli;SHEN Kang;LIU Zhongtao(Teaching Center for Computer Experiment,Henan University of Economics and Law,Zhengzhou,Henan 450046,China)

机构地区：[1]河南财经政法大学计算机实验教学中心,河南郑州450046

出　　处：《矿业研究与开发》2023年第7期5-11,共7页Mining Research and Development

基　　金：河南省科技攻关项目(222102210334,222102210252).

摘　　要：为分析和预测多因素影响下的充填体强度,开展了充填配合比试验,并以支持向量机回归(SVR)模型为基础,结合灰狼优化算法(GWO)建立了一种新型充填体强度预测模型。结果表明,充填体强度随水泥掺量、料浆质量浓度的增大而增大,随粗细骨料比的增大先增大后减小。采用GWO对SVR中的惩罚因子与核函数参数进行迭代寻优,成功建立了以料浆质量浓度、水泥掺量、人工砂尾砂比和养护时间作为输入变量,以充填体的单轴抗压强度作为输出变量的强度预测模型。模型测试集的均方根误差为0.187,决定系数为0.993。与原始SVR和PSO-SVR相比,GWO-SVR模型的预测精度和可靠性有较大提高,成功实现了多因素影响下充填体强度的高精度预测。In order to analyze and predict the backfill strength under the influence of multiple factors,the filling mix proportion test was carried out.Based on the support vector regression(SVR)model and combined with the grey wolf optimization algorithm(GWO),a new backfill strength prediction model was established.The results show that the backfill strength increases with the increase of cement content and slurry mass concentration,and increases first and then decreases with the increase of coarse-fine aggregate ratio.GWO was used to iteratively optimize the penalty factor and kernel function parameters in SVR.The strength prediction model with slurry mass concentration,cement content,artificial sand-tailings ratio and curing time as input variables and uniaxial compressive backfill strength as output variable was successfully established.The root mean square error of the model test set was 0.187,and the coefficient of determination was 0.993.Compared with the original SVR and PSO-SVR,the prediction accuracy and reliability of the GWO-SVR model were greatly improved,and the high-precision prediction of the backfill strength under the influence of multiple factors was successfully realized.

关 键 词：充填体强度 预测模型 配合比试验 灰狼优化算法 支持向量回归 

分 类 号：TD853.391.2[矿业工程—金属矿开采] TD315[矿业工程—矿山开采]