论文部分内容阅读
根据回采巷道稳定性的影响因素,选取围岩强度、埋深、节理裂隙发育程度、巷道跨度、直接顶与煤层厚度之比和松动圈厚度6个指标作为巷道稳定性识别的样本变量。通过搜集部分矿井35条回采巷道相关数据,采用随机森林建立回采巷道稳定性分类模型,并将该模型的预测效果与决策树、BP神经网络和支持向量机模型进行对比。研究结果表明:采用随机森林模型误判率低,具有较高的预测精度,能够相对有效地对回采巷道的稳定性进行判定。
According to the influencing factors of roadway stability, six indexes of surrounding rock strength, depth, development degree of joint and fissure, roadway span, the ratio of direct roof to coal seam thickness and loose ring thickness are selected as sample variables for roadway stability identification. By collecting the relevant data of 35 mining tunnels in some mines, a stochastic forest was used to establish the classification model of mining roadway stability, and the prediction results of the model were compared with the decision tree, BP neural network and support vector machine model. The results show that the stochastic forest model has low false positive rate and high prediction accuracy, which can determine the stability of mining gateway relatively effectively.