随机森林模型在边坡安全系数预测中的应用

Application of Random Forest Model in Prediction of Slope Safety Factor

水对边坡的稳定性影响较大,过多的水容易引起边坡失稳,特别是原地注液过程中,坡体内土体的强度容易下降,边坡容易出现滑坡,使得企业的安全生产及环境受到影响。本文采用随机森林(Random Forest,RF)对该类型边坡进行稳定性预测,根据已有的边坡数据进行模型的训练及测试,表明发现随机森林模型可以满足工程预测需要;要获得较高的精度,全风化层的内摩擦角取值的精度一定要控制好,其对预测精度影响较大;决定边坡安全与否的重要元素是全风化土的内摩擦角,其次为注液强度,注液强度作为可控要素,应加强重视。

Water has a great influence on the stability of the slope.Excessive water is easy to cause slope instability,especially in the process of in-situ injection,the strength of the soil in the slope is easy to decline,and the slope is prone to landslides,which affects the safety production and environment of enterprises.In this paper,random forest(RF)is used to predict the stability of this type of slope.According to the existing slope data,the model is trained and tested,indicating that the random forest model can meet the needs of engineering prediction.To obtain higher accuracy,the accuracy of the internal friction angle of the whole weathering layer must be controlled well,which has a great influence on the prediction accuracy.The important factor determining the safety of slope is the internal friction angle of fully weathered soil,followed by the injection intensity.As a controllable element,the injection intensity should be paid more attention.