Analysis of the characteristics of seismic and acoustic signals produced by a dam failure and slope erosion test

Landslides and floods occur with a significant generation of energy release producing seismic and acoustic signals. The analysis of these signals aims to identify vital and specific characteristics of landslide and flood events, such as frequency content, spectral magnitude, energy variation, and timing of the events. This study examined the seismic and acoustic signals recorded during a large-scale physical dam model failure test. The tests were conducted on the streambed under the No. 2 Bridge in Huisun Forest Experimental Station in Nantou, Taiwan. A dam and a slope model were constructed, and dam failure and slope erosion tests were induced by releasing water to slowly accumulate at the upstream side of the dam and letting water overtop the dam, causing a breach. After dam overtopping, the lowering of the dam crest led to dam failure, causing a large volume of floodwater, resulting in erosion of the toe of the slope model, triggering landslides. Accelerometers and microphones were installed to collect the acoustic and seismic signals produced by the water flow and landslides generated. Unmanned aerial vehicles were used to continuously capture the topography of the slope to produce elevation models to calculate the volumes of the six landslides. Hilbert–Huang transform (HHT) was performed to evaluate the time–frequency spectra for the seismic and acoustic signals, and their spectral magnitudes were discussed. The results showed that the acoustic signals depicted and reacted to the failure event earlier than seismic signals; however, both types of signals were in agreement. The velocity of the flood energy was estimated to be 5.71 and 5.06 m/s by monitoring the seismic signals at different locations of the riverbank and through the analysis of spectral magnitude, respectively. Furthermore, the Arias intensity and the landslide volumes were found to be correlated through a quadratic relationship.