基于回归分析的采煤工作面支架初步选型探究

Preliminary selection of coal mining face support based on regression analysis

液压支架的合理选型是采场空间安全的重要保证。针对采用经验估算法预测支架支护阻力进行支架选型的方法存在预测不准确的问题,提出了基于回归算法的采煤工作面支架选型方法,通过收集100个不同条件下工作面的煤层赋存条件、工作面基本参数和“三机”配套情况为数据集,分别对支架工作阻力、支撑高度和可缩量建立基于决策树、随机森林、支持向量机的回归模型,并通过基于PVD的回归算法进行有效性检验,以均方误差(MSE)、平均绝对误差(MAE)和预测准确率(R~2)作为性能评估指标对于模型的预测效果进行分析。结果表明,决策树和随机森林回归对于支架选型3个重要参数的预测准确最高率达到83.5%、90.7%以及89%,通过PVD降维后建立的回归模型准确率达到88.4%、87.2%和83.8%,说明回归算法在支架初步选型中有不错的效果。该方法能够应用于工作面支架选型,为煤矿工作面未采时进行支架初步选型提供新思路。

Reasonable selection of hydraulic support is an important guarantee of stope space safety.In view of the experience estimation method was used to predict bracket support resistance,then support selection method had the problem of inaccurate,put forward the coal mining face support selection method based on the regression algorithm,by collecting 100 coal seam occurrence conditions,basic parameters of the working face,and the matching situation of the"three machines"under different conditions,the dataset was compiled.Regression models based on decision tree,random forest and support vector machine were established respectively for the working resistance,support height and shrinkage of the support,and the validity was tested based on PVD-regression algorithm.The mean square error(MSE),mean absolute error(MAE)and prediction accuracy(R 2)were used as performance evaluation indexes to analyze the prediction effect of the model.The results showed that the highest prediction accuracy of decision tree and random forest regression for the three important parameters of support selection reached 83.5%,90.7%and 89%,and the accuracy of regression model established after dimension reduction by PVD reached 88.4%,87.2%and 83.8%,indicating that the regression algorithm had a good effect in the preliminary selection of support.This method can be applied to the selection of the working face support and provides a new idea for the preliminary selection of the support when the working face is not mined.