Landslide Dam Failure Analysis Using Imaging and Ranging Sensors

Landslide dam failure may be triggered by heavy rainfall or earthquake and may fail due to seepage or piping because of the asymmetric compaction. Hence, they have the potential to result in serious natural hazards. Rapid assessment of this phenomenon requires the application of investigation and monitoring techniques providing information on the ongoing failure process. To this aim, a downscaled model of a natural dam landslide was reconstructed in a simulation facility (the ‘Landslide Simulator’) located in the Lecco Campus of Politecnico di Milano university, Italy. The failure of the dam was induced by artificial rainfall. A sensor network was setup to record observations during the simulation experiment, including geotechnical, geophysical, and imaging/ranging sensors. This paper focuses on the analysis of deformation measurement and other changes over time, which were observed in the recorded image sequences and 3D point clouds to analyze and predict the failure of the dam. Results showed that water seepage may play a dominant role in the dam failure process, which is anticipated by a sharp increase of strain in the dam body. Furthermore, image processing techniques may help scientists to calibrate numerical models to improve their quality and reliability.