基于GA-SVM的边坡稳定性预测模型研究
STUDY ON SLOPE STABILITY PREDICTION MODEL BASED ON GA-SVM
贾荣谷 1王学祥 2李育红 1高连通 1殷诗茜3
作者信息
- 1. 云南建投第一勘察设计有限公司,昆明 650102
- 2. 云南宾鹤高速公路投资开发有限公司,鹤庆 671599
- 3. 昆明理工大学,昆明 650500
- 折叠
摘要
边坡失稳每年在全球造成重大的经济损失,为了更快速精确地判断边坡的稳定状态,保障边坡工程的安全.针对此问题,本文提出采用GA-SVM算法构建边坡稳定性预测模型.选取6个典型边坡参数:容重、黏聚力、内摩擦角、坡角、坡高、孔隙水压力比作为输入端要素,边坡状态作为输出端要素,结合收集到740个工程实际案例数据构建完成模型训练样本集.结果表明:GA-SVM模型受试者工作特征曲线下面积(AUC)值为0.942,准确率为0.976,召回率为0.976,精确率0.979,F1 score为0.976,反映出该模型在识别边坡失稳状态时的预测精度高、泛化能力强、预测结果可靠.再结合工程实例验证反映出边坡状态的评估结果与实际情况一致.说明可将GA-SVM模型应用于实际的边坡稳定性预测,可为边坡的设计施工提供依据,在实际的工程应用中具有良好的应用前景.
Abstract
Slope instability causes significant economic losses worldwide every year.In order to judge the stability state of slopes more quickly and accurately and to guarantee the safety of slope engineering.To address this issue,this paper proposes to construct a slope stability prediction model using GA-SVM algorithm.Six typical slope parameters are selected:capacity,cohesion,internal friction angle,slope angle,slope height and pore water pressure ratio as input side elements and slope state as output side elements,combined with 740 actual engineering case data collected to build a completed model training sample set.The results show that the AUC value of GA-SVM model is 0.942,the accuracy is 0.976,the recall is 0.976,the accuracy is 0.979,and the F1 score is 0.976,reflecting that the model has high prediction accuracy,strong generalization ability and reliable prediction results in identifying the slope instability state.Then,combined with the engineering example verification reflects that the assessment results of slope state are consistent with the actual situation.It indicates that the GA-SVM model can be applied to the actual slope stability prediction,which can provide a basis for the design and construction of slopes and has good application prospects in the actual engineering application.
关键词
边坡稳定性/机器学习/GA-SVM模型/稳定性预测Key words
slope stability/machine learning/GA-SVM model/stability prediction引用本文复制引用
出版年
2024
NETL
NSTL
万方数据