Shear strain prediction of reservoir landslide based on FBG monitoring and bagging-MLP algorithm

Abstract

Landslides pose significant threats to human lives and infrastructure, and precise understanding and prediction are necessary for effective disaster mitigation. Traditional monitoring methods primarily focus on surface displacement monitoring, which has limitations in understanding the complex evolution of sliding surfaces. This also restricts the improvement in the accuracy and timeliness of deformation prediction models. This study takes the Xinpu landslide in the Three Gorges Reservoir area as an example, utilizing fiber Bragg grating (FBG) technology to monitor the shear strain and shallow soil moisture content during the landslide deformation process. Combining geotechnical and hydrological parameters, a shear strain prediction method considering deformation lag effect is proposed based on machine learning methods. Our findings demonstrate the effectiveness of FBG technology for accurate shear strain monitoring. The integration of hydrological and geotechnical parameters enhances strain prediction accuracy, reflecting the complex interplay of factors influencing landslide deformations. This study presents a shear strain prediction model for shallow sliding surface, contributing to early warning systems and landslide disaster management.