Fonseca R.L. Et Al. (2007) - Protection Systems Against Debris Flows
Fonseca R.L. Et Al. (2007) - Protection Systems Against Debris Flows
Fonseca R.L. Et Al. (2007) - Protection Systems Against Debris Flows
Abstract
Debris flows are natural hazards which look like a combination of flood, land
and rock slide. The phenomenon is also called mud or debris avalanche due to its
similar flow behavior compared with snow avalanches. Debris flows regularly
cause severe damages in mountainous areas like such as the Alps (Brig 1993,
Sachseln 1997, Gondo 2000). Like rock fall, the debris flows loads act mainly
dynamically on a protection barrier. But in contrast to falling rocks debris flows
are not punctual impacts but a distributed load on the protection system. A
further difference between the two hazards is the fact that rockfalls are single
events while debris flows mostly occur in surges. Experiences from North
America, Japan and Europe prove that flexible protection systems like the
Geobrugg VX/UX Systems have an ideal bearing behavior to stop dynamic loads
such as debris flows due to their large deformation capacity
Keywords: debris flow, ring net, ROCCO, dynamic barrier, muds.
1 Introduction
After the successful result of the introduction of the Geobrugg ROCCO® ring net
several years ago, in the Rockfall Protection Systems, and as a result of the
necessity of the avalanche control of materials dragged by the action of the water
(rocks, mud, trees, etc.). Geobrugg was studying the possibilities of placing
properly braced, this type of ring net within the natural channel of these flows,
obtaining with it to stop the heavy and dangerous blocks and trees and being let
pass the water.
In February 1995 a debris flow of 60m3 was held back by a rockfall barrier of
BRUGG Cable Products Inc. (BCPI) along California State Road 41 (County of
San Luis Obispo). The result was that the road was not affected by the natural
WIT Transactions on The Built Environment, Vol 94, © 2007 WIT Press
www.witpress.com, ISSN 1743-3509 (on-line)
doi:10.2495/SAFE070081
74 Safety and Security Engineering II
hazard. As a result of this event detailed studies were carried out in the year
1.996 by the United States Geological Survey (USGS), the California
Polytechnic State University (CALPOLY) and the California Department of
Transportation (CALTRANS). On the test flume of H.J. Andrews Experimental
Forest, Blue River, Oregon the debris flow load on different protection barriers
under different conditions was analyzed. The concrete channel has a total length
of 95m, a width of 2m and the lateral walls are 1,2m high. The top 88m have an
inclination of 31° and the lower 7m one of 3°. The test program contained 6 tests
with debris volumes of approx. 10m3 and impact velocities between 5 and 9m/s.
The composition of the test debris was 24% gravel, 74% sand, 2% silt and clay.
The tests showed that ring nets with laid-on chain-link mesh achieve the best
retention performance. Only 0,05% of the test material passed the barrier. This is
mainly important if there is a road which has to be protected and has to be
drivable all the time. The brake elements engaged, only border rings were
deformed plastically and the ropes and posts were not damaged.
In 1998 the RX-150 system (1.500kJ) installed in Aobandani, Japan stopped a
debris flow with a volume of 750 m3. The maximum deflection of the system
was between 2 and 3m and the remaining system height was about 3,5m
(original height 5m). After cleaning of the barrier only the brake elements had to
be replaced while the ropes and the ring net were not damaged.
In March 2000 a 200m3 debris flow hit a RX-075 barrier (750kJ) in
Seewalchen, Austria. There were also rooted out logs in the debris flow material,
2m of the original height of 3m, remained after the impact and the deflection of
the system was about 2m. The brake elements engaged properly and were
replaced after the cleaning of the system. No further repairs were necessary.
In November of 2001 in Fikushima, Japan, happened a snow sliding of a
volume of 400 m3, stopped by a barrier RX-075. In 2002 in Japan the Tabata’s
project was made, which has stopped events that have been recorded of more of
3.000m3 of blocks and muds. The design of big concrete blocks for anchorage is
not allowed in many countries because of environmental reasons.
After storms of the winter of 2.002 in Santa Cruz de Tenerife in the Canary
Islands, Geobrugg designed and installed a solution of debris control by means
of two lines of barriers, with capacity of 850m3.
In 2.004 Geobrugg installed in the port Gaviota the South of California,
Highway 101, barriers for debris flow control. In June happened an event that
dragged a volume 300m3, it was contained successfully by the installation.
2 Design principles
2.1 Principles
The Geobrugg VX/ UX Protection System against Debris Flow is based on the
approved and from independent institute certified RX Protection System against
Rockfall. Due to the aerial load of debris flows some adaptations are necessary:
stronger support ropes, brake elements with higher capacities and weaker
ROCCO® ring net due to the distributed load; stronger anchorage; protection of
the top support ropes against abrasion.
WIT Transactions on The Built Environment, Vol 94, © 2007 WIT Press
www.witpress.com, ISSN 1743-3509 (on-line)
Safety and Security Engineering II 75
WIT Transactions on The Built Environment, Vol 94, © 2007 WIT Press
www.witpress.com, ISSN 1743-3509 (on-line)
76 Safety and Security Engineering II
1.- steel post 3.- top support rope 5.- flexible anchor
2.- bottom support 4.- middle support rope 6.- lateral support rope
rope 7.- over flow
The remaining barrier height is about 3/4 of the original height. Thus the
minimum barrier height is determined as follows:
If the design height gets to high, a flatter and/or wider barrier location has to be
chosen.
3 Dimensioning parameters
3.1 General characterization
Natural phenomena which are located between landslide/ rockfall and bed load
transportation in water flows are called debris flows. They mostly occur as a
result of heavy rainfall but can also be triggered by other events such as melting
snow or dam failure. The pre-conditions for the appearance of debris flows are
mainly steep slopes, enough material which is easy to mobilize and enough water
to trigger the flow.
Under a mechanical point of view debris flows can be divided in two main
types:
Mud flows, which mainly consist of water and fine material, which is more or
less uniformly distributed.
WIT Transactions on The Built Environment, Vol 94, © 2007 WIT Press
www.witpress.com, ISSN 1743-3509 (on-line)
Safety and Security Engineering II 77
Granular debris flows, which consist of water, fine and rougher material. The
larger components are mostly accumulated at the front of the flow and play an
important role in the overall flow behavior of a granular debris flow.
Observations of debris flows show that they mostly occur in surges. The
observations show further that the velocity and the consistency of the surges may
vary from surge to surge. Therefore it is important to use load parameters always
with a sufficient variation.
3.2 Parameters
Although debris flow load parameters are crucial input data to dimension
protection systems, only few research projects were carried out on this subject so
far. This is a result of the still limited understanding of the mechanics of debris
flows. It is further hard to measure debris flow parameters adequately during real
events. Several mechanical and rheological models were proposed to analyze
and predict debris flows. Due to the lack of field data for comparison,
Rickenmann suggests to use empirical relationships. These empirical
relationships are applied in the following design concept. In the direction profile
the flow depth h, the cross section A and the front flow velocity v of the debris
flow can be defined:
At the location of the barrier the stopped debris flow has to be modelled in the
cross section. In most of the cases there are inclined banks and the experiences
show that the maximum depth of the accumulated material is in the middle of the
torrent. For simplifying it is assumed that the width of the flow corresponds to
the average bed width. The idealised cross section of the stopped flow should
have more or less the same area than the expected flow face.
A0 A1
WIT Transactions on The Built Environment, Vol 94, © 2007 WIT Press
www.witpress.com, ISSN 1743-3509 (on-line)
78 Safety and Security Engineering II
The first step is to estimate a possible debris flow volume VDF. A lot of
different formulas are proposed in the literature, but they are all not very reliable.
Therefore observations and experiences at the location of the project should be
used. A further method is to execute a geomorphologic assessment of the
sediment potential. The volume for flexible protection system lays in a range of
100m3 to 1.000m3.
Several studies proved that the peak discharge of a debris flow is correlated to
its volume. There are different relations for granular debris flows and mud flows.
Mizuyama et al. propose for a granular debris flow (debris avalanche) the
following empirical relationship between peak discharge and debris flow
volume:
QP = 0,135 VDF0,78
Equation following represents the according relationship for mud flows:
QP = 0,0188 VDF0,79
By using the peak discharge it is possible to estimate the average flow velocity v
at the front of the flow. Rickenmann proposes a regime condition for the relation
between velocity, peak discharge and slope inclination (friction considered). S
refers to the gradient of the torrent (tangent of the slope inclination in degrees).
Typical values are S=0,18 (10°), S=0,36 (20°) or S=0,58 (30°).
v = 2,1 QP0,33 S0,33
Japanese guidelines suggest a Manning-Strickler equation to determine the
average flow velocity nd refers to a pseudo-manning value which is typically
between 0,05 s/m1/3 and 0,18 s/m1/3, while the values for granular debris flows
lay between 0,10s/m1/3 y 0,18 s/m1/3.
v =1/ nd H0,67 S0,5
The flow depth h is calculated by using the cross section and the peak discharge.
h = Qp / v b
It is recommended to use both equations and compare the results.
The density of the material is based on empirical values and is about:
γDF = 18 – 23 kN/m3
As a result of the dewatering of the debris flow during the impact on a permeable
barrier not the whole mass of the flow has to be stopped but only a relevant
length or mass. The relevant mass M is determined as follows: It is assumed that
only this part of the flow is acting dynamically which fills up the barrier with
debris between the time of contact and the time of the maximum deflection of
the barrier. In the Oregon tests with volumes of 10m3 Timp was about 1s. Real
debris flows are expected to be much larger and therefore the braking time will
also last longer. Timp is estimated to be between 1s and 4s.
Timp = 1s - 4s (depending on velocity and barrier length)
WIT Transactions on The Built Environment, Vol 94, © 2007 WIT Press
www.witpress.com, ISSN 1743-3509 (on-line)
Safety and Security Engineering II 79
Table 1.
Influence of a debris
Rockfall Debris Flow flow on a flexible barrier
compared to a rockfall
Punctual distributed positive due to smaller
Load
(1 section) (several sections) local loads
Impact positive due to smoother
0,2 – 0,5s 1 – 4s
time deceleration
negative due to static
Type of
single impact in surges loads in the system after
impact
the first impact
Braking negative due to higher
5 – 8m 2 – 3m
distance dynamic forces
Within the scope of a KTI project (Commission for Technology and Innovation)
a computer program was developed to simulate the impact of rocks into flexible
ring net barriers. The project was executed with collaboration of the ETH Zurich
(Federal Institute of technology) and the WSL Birmensdorf (Research Institute
for Forest, Snow and Landscape). The program is called FARO (falling rocks)
and was calibrated by using the data resulting from static pull tests of the single
elements and 1:1 field trials.
With this simulation program it is not only possible to model punctual
impacts but also distributed impact loads. The following figure illustrates an
impact of a debris flow into a UX system. It is assumed that the debris flow only
hits the middle section and that it hits the barrier first in the bottom part and then
fills up the whole system.
WIT Transactions on The Built Environment, Vol 94, © 2007 WIT Press
www.witpress.com, ISSN 1743-3509 (on-line)
80 Safety and Security Engineering II
Figure 6: Impact of the debris flow into the UX barrier (modeled with
FARO).
WIT Transactions on The Built Environment, Vol 94, © 2007 WIT Press
www.witpress.com, ISSN 1743-3509 (on-line)
Safety and Security Engineering II 81
This barrier has been hit by three events during 2005. The barrier was filled
since the first event, and successful supported the loads, generated in the rest of
the events.
Table 2.
With the results of these first measurements in 2.005, have been able to
corroborate several hypotheses, some of general character, that allow
demonstrate the suitability of these flexible systems in front to the traditional
ones. While, on the other hand it has been possible to verify, the form of work of
the different devices integrates in the measurement system, its precise operation,
in the hard conditions of work. With the occurrence of these events, also it has
been possible to control the system mobile communications GSM, indispensable
for the on line control of the operation. The knowledge of the weather conditions
in the place before, during and after to the event, allows to make comparisons
and establish, a certain extent, try to predict their occurrence. Although this is
extremely difficult, since the generation of the torrent, does not follow standard
rules.
5 Conclusions
WIT Transactions on The Built Environment, Vol 94, © 2007 WIT Press
www.witpress.com, ISSN 1743-3509 (on-line)
82 Safety and Security Engineering II
anchorages system and the bearing capacity all the system in front to quasi-static
loads. The steel rings net of the barriers are transparent and fit better in landscape
than the massive steel or concrete structures.
References
[1] Ammann, Walter. Volkwein, Alex (2006). Debris flow test site Illgraben,
Switzerland. Research result from test barrier 2005. WSL, Switzerland.
[2] Luis Fonseca, Roberto, et al. Madrid (2006). Sistemas flexibles de
protección contra flujos de detritus.
[3] Roth Andrea, et al. Valencia (2005). Sistemas de protección contra flujos
de detritus.
[4] Luis Fonseca, Roberto y Laguna Megal, Luis (2003). Dimensionamiento
de la solución para la contención de flujos de detritus. Barranco de
Casalón. Santa Cruz de Tenerife. Canarias.
[5] Roth Andrea, (2003). Design Concept VX / UX Protection System against
Debris Flow.
[6] Duffy J.D., Peila A. (1999), Case studies on debris and mudslide barrier
systems in California
[7] Mitzuyama et al. (1992), Prediction of debris flow peak discharge,
Interpraevent, Bern, Bd. 4, 99-108
[8] Muraishi H., Sano S. (1997), Full scale rockfall test of ring net barrier and
components
[9] PWRI (1988), Technical Standard for measures against debris flows
(draft), Ministry of Construction, Japan.
[10] Rickenmann D. (1999), Empirical relationships for debris flows, Natural
Hazards, 19(1), 47-77
[11] Rickenmann D. (2001), Estimation of debris flow impact on flexible wire
rope barriers, Birmensdorf, interner Bericht, unver. ffentlicht
[12] Salzmann H. (2001), Debris Flow Mitigation by Means of Flexible
Barriers, proceedings 14th Southeast Asia Geotechnical Conference, Hong
Kong
[13] Wartmann St., Salzmann H. (2002), Debris Flow and Floating Tree
Impacts on Flexible Barriers
[14] Grassl, Hans; Volkwein, Axel; Bartelt; Perry and Wartmann,
Stephan.MIT, USA (2003). Experimental and Numerical Modeling of
Highly Flexible Rockfall Protection Barriers.
[15] Rorem, Erick. USA (2004). Performance of flexible Debris flows barriers
in fire burned areas, San Bernardino.
WIT Transactions on The Built Environment, Vol 94, © 2007 WIT Press
www.witpress.com, ISSN 1743-3509 (on-line)