论文部分内容阅读
为对离子型稀土原地浸矿边坡进行稳定性预测,结合赣南离子型稀土矿山42个边坡实例,选取重度、黏聚力、内摩擦角、边坡角、孔隙压力比5个影响因子作为输入,边坡状态作为输出,通过随机森林算法建立边坡稳定性影响因素与边坡稳定状态之间的非线性关系。利用30组边坡稳定性数据作为随机森林预测模型的训练数据集,进行模型的学习训练;用另12组边坡稳定性数据作为预测模型的测试数据,通过训练好的边坡稳定性预测模型进行测试。结果表明,随机森林预测模型精度高,能够为离子型稀土原地浸矿边坡的灾害防治工作提供指导。
In order to predict the stability of in situ leaching slope of ion-rich rare earths and combine with 42 slope examples of Gannan ion-type rare-earth mines, five influencing factors, ie severe, cohesion, internal friction angle, slope angle and pore pressure ratio Factor as input and slope status as output, a nonlinear relationship between slope stability factors and slope stability status is established by stochastic forest algorithm. 30 groups of slope stability data were used as training data set of random forest prediction model to study the training of the model. Another 12 groups of slope stability data were used as the test data of the prediction model. Through the trained slope stability prediction model carry out testing. The results show that the random forest prediction model has high precision and can provide guidance for the prevention and control of the ion-type rare earth immersed mine slope.