论文部分内容阅读
由于复杂工程地质条件和环境因素的综合影响,边坡变形呈现复杂非线性演变特征。针对位移时间序列未能完全考虑环境因素对边坡变形的影响,故将影响边坡变形的有效降雨量加入监测位移时序,组成多因素位移时间序列。引入粒子群算法(PSO)对支持向量机(SVM)的模型参数寻优,结合滚动预测方法,建立了适合边坡变形预测的多因素位移时间序列PSO-SVM模型。以华光潭一级厂房后边坡表面观测位移为例进行预测分析,研究表明,新模型预测结果科学可靠,有效弥补了传统PSO-SVM后期预测泛化能力的不足,提高了模型的预测精度。新模型在边坡位移时序预测中具有一定的工程应用价值。
Due to the complicated influence of complex engineering geological conditions and environmental factors, the deformation of slope shows complicated nonlinear evolution characteristics. In view of the fact that the displacement time series can not completely consider the influence of environmental factors on the slope deformation, effective rainfall, which affects the slope deformation, is added to the monitoring displacement sequence to form a multi-factor displacement time series. The Particle Swarm Optimization (PSO) is introduced to optimize the model parameters of Support Vector Machine (SVM). Combined with rolling forecasting method, a multi-factor displacement time series PSO-SVM model is established for slope deformation prediction. The case study shows that the prediction results of the new model are scientific and reliable, which effectively make up for the lack of generalization ability of the traditional PSO-SVM post-prediction and improve the prediction accuracy of the model. The new model has some engineering application value in slope displacement prediction.