, Page 0-0 (1)

Volume Title: ICASGE2023

DOI: ICASGE-GEO

¹Imaya Ariyarathna ; ²Katsuo Sasahara

¹Life Environment Conservation Science, Ehime University

²Natural Science Cluster, Research and Education Faculty, Kochi University, Japan

Time prediction methods based on monitoring surface displacement (SD) are effective for early warning against shallow landslides. However, the failure time prediction by Fukuzono’s original inverse-velocity method (INV) causes less accuracy due to the variation in inverse-velocity (1/v) caused by the noise of the measured SD, which amplifies the fluctuation of resultant 1/v. Therefore, the present study aims to introduce the pre-analysis to acquire better prediction by reducing the effect of noise on the measured SD. The data extraction (DE) and moving average (MA) methods were used to filter the measured SD for better smoothing of 1/v. Both reproducibility of measured SD and the scattering were assessed using root mean square error (RMSE) and determining factor (f), respectively, to select the optimum SD interval (∆x) for data extraction in the DE method. The data, treated by DE and MA methods, were utilized to predict the failure time based on the INV method and the relationship between velocity and acceleration on a logarithmic scale (VAA method). Accordingly, ∆x gives at the smallest sum of normalized RMSE and normalized (1-f), which offers a better prediction. When the SD at failure changes, ∆x is changed. The best prediction gives by the DE preprocessing with the VAA and the time prediction using processed data by the MA method shows poor prediction. In some cases, the prediction by VAA using MA data gives better prediction compared with the results of the INV method by MA data.

Shallow landslide; Failure time prediction; Model slope; Monitoring surface displacement

Geotechnical Engineering (GEO)