Novel rockfall models allow the specification of general polyhedral block shapes and incorporate the scarring behavior of the surface substrate of an impacting rock. Despite these promising developments, accurate determination of important parameters governing the rockground interaction pose a challenge.
Here, we report the methods and first results of real scale field experiments that were designed to calibrate the RAMMS::ROCKFALL simulation model. The three-dimensional trajectory reconstruction with fully temporal information of induced single block rockfall experiments is presented. Experimental rocks were instrumented with in-situ accelerometers and gyroscopes. Unmanned aerial systems mapping techniques, videogrammetry and radar measurements were additionally employed as external measurement techniques.
This setup allows capturing rock velocities, jump heights and jump lengths, as well as scaring duration, extents and impact accelerations. The cross-validation of the results obtained via different experimental techniques exhibit promisingly small differences. The experimental data provides novel detailed insights into the rock-ground interaction. The presented complete kinematic description of rock trajectories represents the first systematic experimental data set available for calibration purposes - overcoming the limitation of solely being dependent on case study back calculations.
Nets and dams are widely used mitigation measures to protect roads, railways or settlements from rockfall. To dimension such protection infrastructure, engineers need tools that can accurately determine kinetic energies, run out distances, jump heights and lateral dispersion of rockfalls. Three dimensional rockfall simulation models are a commonly used method to determine these parameters. For this purpose, there are numerous different rockfall models on the market (e.g. Guzzetti et al., 2002; Agliardi & Crosta, 2003; Lan et al., 2007; Woltjer et al., 2008; Leine et al., 2014; Dorren, 2016). Often the rock shape is reduced to a single point (lumped mass models) or simplified to a sphere. Another drawback of most models is the implementation of the rock-ground interaction model: they assume, that the impact happens on a single point and the energy dissipation of the ground is simplified as a two overall, tangential and normal coefficient of restitution (COR), which is usually determined by case study back calculations. Thanks to their fast computational performance and ability to incorporate stochastic processes, COR models are widespread. Additionally, the lack of available data regarding rock-soil interaction during a real scale rockfall event, prevented a model description of meter long scars after rockfall events as seen in the field. It was so far best practice to cope with the phenomena as - until now - there has been no detailed data regarding the rock-soil interaction available. Even in case studies based on scar mapping (e.g. Saroglou et al., 2018), the rock trajectories are ambiguous, as the temporal information is missing, which allows still several different parabolas to fit.
