International Society for Rock Mechanics and Rock Engineering ISRM International Symposium

Norwegian Group for Rock Mechanics Eurock 2020 – Hard Rock Engineering
Trondheim, Norway, 14-19 June

Full Scale Testing of Rockfall Nets under Realistic

Conditions

A. Caviezel, M. A. Sanchez, G. Lu, M. Christen & P. Bartelt

WSL Institute for Snow and Avalanche Research SLF, 7260 Davos Dorf, Switzerland
caviezel@slf.ch

C. Wendeler & A. Lanter

Geobrugg AG, 8590 Romanshorn, Switzerland

Abstract
Rockfall nets are often tested in unrealistic, academic conditions that ignore common load cases. Two,
naturally occurring important load cases are missing in certification norms: Firstly, fast rotating rocks
bear the potential, depending on their rock size, impact location and mesh geometry, to over climb the
nets. Secondly, non-ideal, acentric net impacts including direct column hits are inexistent as well.

In order to approach real-world conditions, we present a novel full scale rockfall net testing site in natural
terrain. The shift towards realistic impact conditions imposes the use of sophisticated measurement
devices in order have a fully determined impacting system. We exploit recent advances in sensor
technology and progress made in rockfall reconstruction methodologies. In-situ accelerometer and
gyroscope data are combined with videogrammetric and unmanned aerial systems mapping techniques
to understand the role of rock rotations, ground penetration and translational scarring in rockfall motion.
The exhaustive target tracking and succeeding trajectory reconstruction provides information over the
complete flight path such as translational velocity vectors, angular velocities, impact duration and forces,
ballistic jump heights and lengths during ist downslope travel as well as upon impacting the mitigation
structure.

The experimental data provides insight into the basic physical processes detailing how rotating rocks of
general shape penetrate, rebound and scar ground terrain and consequently rockfall nets. The adequate
simulation of change of kinematics during the impact becomes of key importance in order to match
realistic mitigation structure behaviour.

The presented methodologies pave the way to a comprehensive understanding of rock-ground and rock-
net interaction, a key requirement to develop numerical models and loading procedures that account for
the role of rock shape, spin and eccentric impacts in rockfall protection.

Keywords
Rockfall net, barrier, full scale testing
 Eurock 2020 – Hard Rock Engineering

1 Introduction
Technical mitigation measures such as nets and dams are widely used to protect roads, railways or
infrastructure from rockfall. Accurate dimensioning of such protection infrastructure still poses a
challenging task to hazard engineers. A state-of-the-art approach is the use of computer assisted tools
that not only can objectively forecast kinetic energies, run out distances, jump heights and lateral
dispersion of rockfalls but preferably the complex interaction of the impacting rock upon a mitigation
structure. Rockfall nets are often tested in well controllable, but unrealistic and hence academic
conditions. Practical and important load cases pose a greater experimental challenge. Important load
cases that occur in natural environments are presently not considered in the certification norms: Off-
centre field impacts, rock spinning effects upon impact and border field hits. While the first present
asymmetric load cases to be considered for the energy level classification process of the individual
construction parts, the second case can cause rocks to climb over nets depending on rock size, impact
location and mesh geometry. The latter case aims at the uncertainty of the certification level for the
border fields of any flexible rockfall barrier. While certification procedures only consider constructions
of center field, the performance decrease for border fields is not well established yet. This uncertainty
even yields to different risk management and installation guidelines across different countries. While
off-centre field impacts can also be addressed in more technical setup, the generation of fast rotating
rocks is merely possible via a sufficiently large acceleration zone in natural terrain.

2 Experimental test site and methods

2.1 Experimental Site
The experimental site named Chant Sura (46.74625◦ N, 9.96720◦ E) is located on the Flüelapass, 12 km
south-east of Davos, Switzerland (see Fig. 1). The release point is located at an altitude of 2380 m a.s.l.,
yielding a projected travelling distance of roughly 145 m until reaching the installed rockfall barrier.
The slope has been used during the last three years for rockfall tests in unobstructed conditions (Caviezel
et al. 2019), whose deposition points are indicated as red dots in Fig. 1. The soil characteristics feature
typical alpine meadow interspersed with rocks featuring slope angles between 40 and 80◦ in the
transition zone leading towards the rockfall barrier. The slope serves as an ideal representative of an
alpine environment subjected to rockfall hazards. The easy accessibility provided via the pass road and
the lack of infrastructure or hiking paths facilitates repetitive full-scale experiments.

Fig. 1 Test site overview: indicated is the release

point (black cross) and the the deposition points
obtained during the unobstructed experimental
campaigns (red dots). The installed six-field 2000
kJ flexible rockfall barrier is indicated as blue line.
Data acquisition location for barrier sensor data is
indicated as white astersik Inset: geographical
location of the Chant Sura test site within
Switzerland.

2
 Eurock 2020 – Hard Rock Engineering

2.2 Experimental methods

The deployed external and in-situ measurement devices are equivalent to the previous experimental
campaigns in the same, but unobstructed slope carried out in preceding years (Caviezel et al. 2019).
They comprise a pre- and post-experimental UAS survey in order to generate latest high-resolution
digital terrain surface models with a grid resolution of ~5 cm and the possibility of a-posteriori scar
mapping via the difference map. Static 6K videogrammetry serves as overview footage. In-situ sensor
StoneNode v1.2 mounted in the rock’s centre of mass is recording gyroscopic motions up to 4000 °/s
and in-situ accelerations up to 400 g with a data acquisition rate of 1 kHz. Recording times of several
hours allows the tracking of all runs with the use of a single sensor (Niklaus et al. 2017, Caviezel et al.,
2018). Test rock are the equant EOTA111 and platy EOTA221 variants of the perfectly symmetric norm
rock of the European Organization for Technical Assessment used in standardized rock fence testing
procedures in official European Technical Approval Guidelines (ETAG027 , 2008). Test blocks are
made from reinforced concrete with weights of 800 kg and 2600 kg. Artificially constructed rocks allow
for full control over rock shape and mass and reproducibility of test specimens. Additionally, reinforced
concrete approaches natural rock densities well and excels in durability allowing for more experimental
runs with the same test block. The rock is released either manually via a hydraulic platform or
disengaged directly from the helicopter’s longline via electrical hook. Small rocks can are transported
back to the release point via an Airbus H125 helicopter while the slinging work for the large rocks
require an Airbus H225.

The flexible rockfall barrier is a six-field ROCCO-2000 EOTA class 5 barrier (see Fig. 2a). The field
and post labelling is ascending from westerly to easterly positions. Accordingly, Figure 2a shows an
off-center impact in field four, close to post five. Impacts into field one and six are considered as border
field impacts Modification to normal installation includes a slightly larger concrete foundation for the
posts in order to accommodate the modified base plate hosting load cells at post 3-6 (see Fig. 2b). Lateral
support ropes (Fig 2c) as well as slope facing anchor ropes (Fig 2d) are equipped with rope load cells
measuring occurring strain forces upon rock impact. The data are collected into two amplifiers and
collected on-site in realtime at the data acquisition point indicated in Fig. 1. Miguel: Can you input
exact measurement dvices and their measurement ranges: how many kN, etc., which post is equipped,
etc…

Fig. 2: a) Installed 6-field 2000kJ flexible rockfallbarrier at Chant Sura, Flüelapass, Switzerland. The field and post labelling
is ascending form westerly to easterly position. b) Custom baseplate conection for incorporation of load cells. c) and d) rope
load cell installation for top support rope and back ropes.

3
 Eurock 2020 – Hard Rock Engineering

3 Data and First Insights

Here, we report first ever collected data at the newly installed flexible barrier at the Chant Sura test site.
The presented data originates from the first impact onto the barrier during the experimental campaing
800 𝑘𝑔
of fall 2019, using a 𝐸𝑂𝑇𝐴111 block. Figure 3 shows movie stills extracted from the 6K video footage.
While hardly discernible in the stills, the motion pictures clearly show a hit to the back support rope in
Fig 3a at t = 16.43 and induced motion at the border field posts 6 and 7. Figure 3b and 3c visualize the
transition into the net, covering a projected distance of roughly 9 m from anchor to the net within
approximately 0.67 s resulting in an impacting velocity of 14.2 m/s. Figure 3d-f visualize the net impact
from first contact over full elongation, that is reversal of sense of motion, to rebound of the block from
the net within 1.25 s.

Fig. 3: Cropped video stills extracted from the 6K footage. (a) Impact on barrier support rope at t = 16.43 s. (b) Airborne phase
before impact at t = 16.68 s. (c) Airborne phase before impact at t = 16.93 s. (d) First contact to rockfall barrier at t = 17.01 s.
(e) Reversal, that is maximal net elongation upon impact at t = 17.46 s. (f) Rebound impact of block at t = 18.26 s.

Figure 4 shows the StoneNode gyroscope sensor stream. Plotted are the angular velocities of every single
axis together with its resultant. As the rockfall barrier has been installed at the end of the transition zone,
rotational speeds are close to its maximum when impacting into the barrier. This run features maximal
angular velocities of roughly 1800 °/s just before the impact. The resultant rotational energy amounts to
roughly 30 kJ. The calculated translational speed yields an energy level of 75 kJ. The boulder thus carries
a total energy of 105 kJ, 72% of which being stored translational degrees of freedom and 28% in
rotational motion, respectively. This only signifies a five percent load case to the installed barrier.

Figure 5 shows load cell measurements for the same impact. Due to an amplifier failure on part of the
instrumentation only partial data is available for the first impact. Meaningful data streams have been
extracted for load cell measurements for top support rope (left), bottom support rope (left), lateral
support rope (left) and Post 4 and 7. The first signal, predominantly visible in the Post 7 measurement
is in accordance with the pre-impact loading via the support rope at t = 16.43 s. Main impact occurs at t
= 80 s (load cell internal time axis). As the reported impact was a border field hit into field six, larger
forces to lateral posts are expected. Main loading is visible during the previously determined contact
time of 1.25 s. Subsequent relaxation motion of posts and support ropes are visible in the load cell data.

Maximum forces of 100 kN in Post 7 and 60 kN in the top support rope are measured. Official
certification procedure in Switzerland is done via a 50% pre-test loading followed by a 100% loading.
The official type certificate from the Federal Office for the Environment FOEN (BAFU, 2013) reports
braking times of 0.3 s and maximum tensile forces in a stay camble of 150 kN for the 50%, that is a
1000 kJ, load case. The presented Chant Sura load case features only 20 kN in the stay cable and a
braking time of 0.45 s.

Miguel: Do we have measurement range? Than we know the loading percentage? Do we have any
data on impacts in Walenstadt? What forces are the cable dimensioned?…

4
 Eurock 2020 – Hard Rock Engineering

Fig. 4: In-situ StoneNode measurement of the angular velocity. Plotted are the individual axis as well as the resultant angular
velocity during the drop.. The temporal positions according to the movie stills are indicated with dashed lines: Impact at t =
17.01 s - the drastic decline of rotational speed and its rebound motion is clearly visible within the sensor stream; Reversal at t
= 17.46; Rebound impact of block at t = 18.26 s.

Fig. 5: Load cell measurements for top support rope (left), bottom support rope (left), lateral support rope (left) and post 4 and
7. The pre-impact loading via the support rope hit featuring in Post 7 is nicely visible in the data stream. Main impact occurring
at t = 80 s (internal time axis). As the reported impact was a border field hit into field six, larger forces to lateral posts are
measured. Main loading during the 1.25 s from first contact to rebound impact with subsequent relaxation motion of posts and
support ropes are visible in the load cell data.

5
 Eurock 2020 – Hard Rock Engineering

4 Conclusions
The first experiments have been conducted successfully at the Chant Sura test site with a ROCCO 2000
kJ net installed. This first test shows, that impacting energies are rather low for 800 kg blocks, as
expected. In-situ instrumentation combined with an instrumented net allows for a comprehensive
approach to acting forces and impulses during contact situations. Further analysis of the breaking
motions, above all the energy dissipation mechanisms from the different degrees of freedom into the
mitigation structure will be main focus with respect to data analysis. The small rock impacts represent
a valuable low-load situation generating viable data in order to study scaling effects and evolution of
material strain upon increased loadings. Future tests seek to recreate increased load cases, preferably
close to and beyond the specified barrier capacity. Such a data set then can serve as calibration basis for
a real-world data based rockfall net simulation program.

Merging abundant test data from common vertical drop test with these realistic load cases fosters
enhancements in existing rockfall barrier simulation codes. It provides a multidimensional calibration
basis for the different elements. Prospective benefits are not only more accurate and faster modelling of
barrier loadings but also a seamless integration of mitigation structure simulation in the rockfall
trajectory simulation program RAMMS::ROCKFALL. v

The continuation of the experimental campaigns will generate further viable data for single field impacts
or special load cases to a flexible rockfall barrier under realistic conditions. It provides insight into the
basic physical processes detailing how rotating rocks of general shape impact, penetrate and rebound
from rockfall nets. The adequate simulation of changes in kinematics during the impact becomes of key
importance in order to match realistic mitigation structure behaviour. The presented methodologies
supports the development of numerical models and simulation of realistic loading procedures that
account for the role of rock shape, spin and eccentric impacts in rockfall protection.

References
BAFU (2013), Zertifikat Geobrugg RXE-2000,
https://www.bafu.admin.ch/dam/bafu/de/dokumente/naturgefahren/fachinfo-daten/rxe-
2000.pdf.download.pdf/rxe-2000.pdf, Accessed at 29.10.2019
Caviezel A, Schaffner M, Cavigelli L, Niklaus P, Bühler Y, Bartelt P, Magno M, Benini L (2018), Design and
Evaluation of a Low-Power Sensor Device for Induced Rockfall Experiments IEEE Transactions on
Instrumentation and Measurement, 67, 4: 767-779 , https://doi.org/10.1109/TIM.2017.2770799
Caviezel, A., Demmel, S. E., Ringenbach, A., Bühler, Y., Lu, G., Christen, M., Dinneen, C. E., Eberhard, L. A.,
von Rickenbach, D., and Bartelt, P (2019), Reconstruction of four-dimensional rockfall trajectories using
remote sensing and rock-based accelerometers and gyroscopes, Earth Surf. Dynam., 7, 199-210,
https://doi.org/10.5194/esurf-7-199-2019
Niklaus P, Birchler T, Schaffner M,Cavigelli L, Caviezel A,Magno M, and Benini L (2017) Stone-Node: A Low-
Power Sensor Device for Induced Rockfall Experiments, IEEE Sensors Application Symposium,
Glassboro, New Jersey (USA), https://doi.org/10.1109/SAS.2017.7894081
ETAG027 (2008), Guideline for European technical approval of falling rock protection kits. EOTA. Brussels,
Belgium

https://www.bafu.admin.ch/dam/bafu/de/dokumente/naturgefahren/fachinfo-daten/rxe-
2000.pdf.download.pdf/rxe-2000.pdf