考虑位移滞后效应的降雨型滑坡SSA-DELM位移预测模型研究

Rainfall Induced Landslide Displacement Prediction Model Based on SSA-DELM Considering Hysteresis Effect of Displacement

针对东南丘陵山地降雨型滑坡变形发展特征,现有滑坡预测模型应用存在局限,结合滑坡变形特点研究基于智能算法的滑坡预测模型。以福建安溪尧山滑坡为例,选取2019年9月至2022年6月滑坡监测数据进行研究,采用集对分析、灰关联法、麻雀搜索算法及深度极限学习机对滑坡位移进行预测,提出了一种考虑滑坡位移滞后时间基于深度学习的滑坡位移预测模型。结果表明:SSA-DELM模型的MAE、MAPE、RMSE相较于已有的BP神经网络、SVM模型均更小,同时模型结合了滑坡影响因子以及水位-位移滞后特征,具有明确的物理意义,位移预测效果较好且精度较高,可推广应用于类似的滑坡位移预测中。

This paper proposes a new landslide displacement prediction model.The model is based on deep learning which considers the hysteresis effect of landslide displacement.There are limitations in the application of existing landslide prediction models,due to the deformation and development characteristics of rainfall-induced landslide in southeastern China,in which there are plenty of mountains and hills.Therefore,it is of great theoretical and practical significance to study the landslide prediction model for disaster prevention and reduction.The model is based on intelligent algorithm,combining with the characteristics of landslide deformation.Research selected the landslide displacement monitoring data from September 2019 to June 2022,in Mount Yao,Anxi,Fujian Province.The model adopted gray relational analysis,set pair analysis and optimization of deep extreme learning machine based on sparrow search algorithm.Results show that the mean absolute error,the mean absolute percentage error and the root mean square error(RMSE)of the proposed SSA-DELM are all lower than existing BP,SVM model.Moreover,model integrates with influence factors of landslide and hysteresis effect of water level displacement.