支持向量机模型在渗流监测中的应用

Application of Support Vector Machines Model in Seepage Monitoring

提出了基于一种支持向量机 (SVM)的渗流监测方法。该方法采用结构风险最小化原则 ,能够在对小样本学习的基础上 ,对其他样本进行快速、准确的拟合预测 ,具有更好的泛化性能和精度 ,减少了对经验的依赖。在算例中 ,基于 SVM的非线性特点 ,根据土石坝的实测资料建立了渗流监测模型 。

Because seepage monitoring is an uncertain, nonlinear system, it is difficult to describe such nonlinear characteristics of system by traditional methods, so the pitometer head could not be accurately forecasted. The authors presented a novel seepage monitoring method in which an improved Support Vector Machines (SVM) algorithm was applied and the principle of Structural Risk Minimization (SRM) was embedded into the SVM. Therefore, on the basis of learning by fewer samples the presented method could conduct fast and accurate head forecasting with other samples fitting head forecasting. The presented method is more generalized and its dependence on experience is weakened. As a example, based on the nonlinear characteristic of SVM, seepage monitoring model is built according to factual data of soil dam, which lays a foundation for head forecast and safety monitoring.