Document Actions

IJEGE-11_BS-Volkwein-et-alii

background image
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
1093
DOI: 10.4408/IJEGE.2011-03.B-118
DESIGN OF FLEXIBLE DEBRIS FLOW BARRIERS
a
xel
VOLKWEIN
(*)
, C
oRinna
WENDELER
(**)
& G
uido
GUASTI
(***)
(*)
Swiss Federal Institute for Forest, Snow and Landscape WSL - Switzerland, volkwein@wsl.ch
(**)
Geobrugg AG - Switzerland, corinna.wendeler@geobrugg.com
(***)
Geobrugg Italia - Italia, guido.guasti@geobrugg.com
flow height, density, normal and shear forces.
The developed load model guides design of de-
bris flow retention, and provide impact forces for
correct barrier design.
K
ey
words
: Debris Flows, Barrier, Mitigation, Dimensio-
ning, Design
INTRODUCTION
Debris flows can be mitigated by (a) dewatering
the debris, which removes one necessary component
for a debris flow, and (b) providing a retention sys-
tem. These two principles apply to flexible barriers
that are commonly used for rockfall protection. To
apply flexible barriers to debris flow, several ques-
tions require answers: What are the debris flow
loads? How does the barrier perform during the
filling process? What are the physical limits for the
investigated barriers? Focussing on these questions
a load model is presented that allows the design of
flexible barriers for debris flows.
Experiences from North America, Japan and
Europe (d
uffy
1998 & d
e
n
atale
et alii, 1996)
have proven that flexible protection systems have
an ideal bearing behaviour to stop dynamic loads
such as debris flows due to their large deformation
capacity and their water permeability.
A flexible debris flows barrier (see Fig. 1) is typi-
cally placed in the river channel between the river
banks, with a potential to span up to 15 m (25 m
ABSTRACT
A new type of flexible net barrier system de-
signed to protect against debris flows with volumes
of up to 1000 m
3
has been developed. A detailed
study and testing programme, conducted for the
first time, has demonstrated their highly cost effec-
tive and efficient design in comparison to massive
concrete barriers. A multi-step impact model was de-
veloped describing the filling process and the acting
forces to the barrier simultaneously. During debris
flow events, the total pressure distribution on the net
can be approximated by time-discretizing the contin-
ued filling and by tracking following surges over the
original deposits. In case of a completely filled bar-
rier, overflowing debris material loads the net with a
normal and shear force component. The hydrostatic
pressure and the additional weight of overflowing
material are reduced through compaction and drain-
age over time. The observed overflow of a filled bar-
rier without any damages led to the idea of multilevel
barrier application to gain higher retention volumes.
The theoretical model has been validated and
verified using a full-scale and instrumented field
installation of a net barrier at the Illgraben torrent
in Switzerland. This enabled (a) to investigate its
performance, (b) to measure the impact forces and
(c) to provide information on the expected mainte-
nance. Impact and shear forces were measured at a
shear wall and a force plate which delivered useful
information for the model like pressure profile over
background image
A. VOLkwEIN, C. wENDELER & G. GUASTI
1094
5th International Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction and Assessment Padua, Italy - 14-17 June 2011
compressing the underlying sediment thereby releasing
most of it's pore water over the time (w
endeleR
, 2008).
Should a barrier fill completely following surges
will spill over the barrier adding additional loads by
weight and shear forces. The time to drain fine granu-
lar material depends on grain size composition and
the water content at impact. The static loads transform
from hydrostatic to earth pressures.
Different aspects have to be considered to avoid
barrier failure:
• strong anchorage
• strong support ropes
• energy absorption
• protection of the top support ropes against abrasion
• retention volume commensurate with the design
volume.
The maximum retention capacity of a barrier is
dependent on the channel slope, deposition angle and
the height of the barrier (l
ien
, 2003). Empirical studies
suggest that the deposition angle corresponds to 2/3 of
the original torrent gradient (R
iCkenmann
, 1999, Fig. 4).
with additional posts) with heights ranging 2 to
6 m. A steel net is spanned by support and lateral
ropes. The ropes are anchored in the banks with
anchor lengths depending on the load capacity
of the ground. Plastically deforming and hence
energy absorbing elements in the ropes allow
large plastic deformations in the barrier system
and reduce the peak loads during impact.
The goal for a design model for such barri-
ers is to obtain the forces within the single com-
ponents (ropes, net, posts, anchorage, founda-
tion). Furthermore, the adjustable lengthening
of the energy absorbers allows optimizing the
load distribution within the system.
Like rock fall loads the main force acts dynami-
cally on a protection barrier during debris flow. In
contrast to falling rocks, debris flows produce a dis-
tributed load and debris flows occur in surges.
In the following; a load model is presented that
allows a design of a flexible barrier against debris
flows. This model has been developed and validated
based on laboratory test, full scale field tests and
numerical simulations (w
endeleR
, 2008; w
endeleR
et alii, in prep., Fig. 2).
DESIGN PRINCIPLES
Due to the net’s permeability an impacting granu-
lar debris flow is drained as a result of the retention of
rougher material. The stopped certain length of the de-
bris flow and the amount of debris material gives the so
called relevant length / mass-ratio. The continuous filling
of the barrier can now be modelled step by step: After the
first impact the additional material overrides the first ar-
rested surge (see Fig. 3) providing additional weight and
Fig. 1 - Debris flow barriers 1998 in Aoban-
dani, Japan (left, 750 m
3
retained,
deflection 2 - 3 m, remaining bar-
rier height reduced from original 5
m down to 3.5 m) and 2008 in Villar
Sautoreglia, Italy (right, 1500 m
3
re-
tained). Post-event material (water
and sediments) goes over the barriers
Fig. 2 - Development of a load model for flex-
ible debris flow barriers: Field tests in
the Illgraben (left), physical modelling
with small scaled tests in the labora-
tory (centre) and numerical modelling
with the finite element software (right)
(w
eNDeler
, 2008; v
olKweiN
, 2004)
Fig. 3 - Modelled second filling wave of a debris flow with
flow height h
0
and its loading components of dy-
namic pressure (ΔP) and hydrostatic pressure (P
hyd
)
(w
eNDeler
et alii, in prep.)
background image
DESIGN OF FLEXIBLE DEBRIS FLOW BARRIERS
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
1095
STEP BY STEP APPROACH
Figure 5 shows a step by step approach to design
net barriers. The first step is to estimate a possible
debris flow volume VDF. Numerous different for-
mulas are proposed in the literature, although each
has limited reliability. Therefore, observations and
experiences at the location of the project should be
used in conjunction with the respective formulas. A
further method is to execute a geomorphologic as-
sessment of the sediment potential (R
iCkenmann
,
1999). It is therefore recommended that debris flow
volumes must be determined in detailed site-specific
studies and that a frequency-magnitude relationship
needs to be established using extreme value statis-
tics in order to obtain a reliable design basis. Exam-
ples on relationships between debris flow Volume
V
DF
, catchment area A
c
and mean slope inclination
I
s
can be found in b
eRGmeisteR
et alii (2008), R
iCk
-
enmann
& z
immeRnann
(1993), H
amPel
(1980) or
d’a
Gostino
et alii (1996). The volume capacity for
one flexible net barrier system lies in a range of
V
R
= 100 m
3
- 2,000 m
3
depending on channel topography.
Several studies have proven that the peak discharge
of a debris flow is correlated to its volume. There are
Kinetic energy is mainly dissipated in the en-
ergy absorbing brake elements. To activate the
brake elements, the net must transfer the load to the
support ropes in which they are installed. During
a debris flow both single impact loads from indi-
vidual boulders and fully distributed loads of the
flow’s front occur. The links between the single net
meshes have to be strong enough to withstand the
high forces that must be transmitted to the margins
and supporting structure.
SAFETY CONCEPT
Ideally, intensity and return period lead to a prob-
abilistic density function to describe the debris flow
pressure. However, this safety concept was not based
on a probabilistic analysis because of limited field in-
vestigation data. But the given safety parameter were
deduced from existing Swiss guidelines dealing with
natural hazard impacts on buildings (e
Gli
, 2005) and
snow fences (b
uwal
, 2007).
Resistance: The resistance safety factor can be set
to γ
R
= 1.35 according to b
uwal
(2007).
Load: The safety factor on the loading is first
influenced by the risk potential (Tab. 1). Three risk
classes were defined in Table 2 summarizing pro-
posed safety load factors. A preliminary guideline for
a safety concept of debris flow protection measures is
in review in Austria (ONR 24802, 2009).
DIMENSIONING
DEBRIS FLOw CHARACTERISATION
From a mechanical point of view debris flows can
be divided in two main types:
• Mud flows that consist of water and fine mate-
rial; and.
• Granular debris flows that consist of water and a
coarser grain size distribution, typically lacking
the clay fraction.
Fig. 4 - Deposit in the flow direction behind the bar-
rier (w
eNDeler
, 2008)
Tab. 1 - Classes according to risk potential
Tab. 2 - Safety factor γ
F
on the loading site for different
time periods and risk classes
background image
A. VOLkwEIN, C. wENDELER & G. GUASTI
1096
5th International Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction and Assessment Padua, Italy - 14-17 June 2011
I
s
refers to the gradient of the torrent (tangent of the
slope inclination in degrees). Typical values for Is
are I
s
= 0.18 (10°), I
s
= 0.36 (20°) or I
s
= 0.58 (30°).
v = 2.1 · Q
P
0.34
· I
S
0.2
(v
d
= 2 m/s - 8 m/s).
Japanese guidelines (PWRI, 1988) suggest a Man-
ning-Strickler equation to determine the average flow
velocity (see also G
ReGoRetti
, 2000). Here, n
d
refers to
a pseudo-manning value which typically lies between
0.05 s/m
1/3
and 0.18 s/m
1/3
, while the values for granular
debris flows lay between 0.1 s/m
1/3
and 0.18 s/m
1/3
.
different relations for granular debris flows and mud
flows. m
izuyama
et alii (1992) propose for a granular
debris flow (debris avalanche) the empirical relation-
ship between peak discharge and debris flow volume:
Q
P
= 0.135 · V
DF
0.78
(Q
P,d
= 5 m
3
/s - 30 m
3
/s)
Using the peak discharge Q
Pd
, allows estimat-
ing the average flow velocity v at the front of the
flow. R
iCkenmann
(1999) proposes a regime con-
dition for the relation between velocity, peak dis-
charge and slope inclination (friction considered).
(1)
(2)
Fig. 5 - Diagram for step wise dimensioning procedure for flexible debris flow barriers. A symbol list can be found at
the end of the document
background image
DESIGN OF FLEXIBLE DEBRIS FLOW BARRIERS
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
1097
However, it is recommended to use super eleva-
tion and to back calculate the velocities in combina-
tion with a sensitivity analysis instead of using eqn. 3.
It is recommended to use both equations and com-
pare the results.
The flow depth h is calculated as a function of the
cross section and the peak discharge.
However, the typical flow depth is better meas-
ured in the field based on levees or scour marks.
The density of the material is about ρ ≈ 1,600-
2,000 kg/m
3
for a mud flow and about ρ ≈ 1,900-2,300
kg/m
3
for a granular debris flow R
iCkenmann
(1999).
The post-event barrier height is about 3/4 of its
pre-event height. Thus the minimum barrier height is
determined as follows:
with V
R
retention volume, ε barrier inclination and θ
and θ' the gradient of the material before and after a
debris flow event.
MULTI-LEVEL CONSIDERATIONS
The distance between two barriers is important for
the construction of a multi-level barrier in series. The
distance should be long enough that a hydraulic jump
and a backwater curve next to the check dam can achieve
the greatest loss of energy (see Figure 6, the blue-line is
the water level). The inclination of the river bed behind a
filled barrier I
s
’ should achieve a subcritical flow regime
in order to have more stable river bed conditions.
The distance between the barriers should not be
smaller than the influenced backpressure length.
The behavior of multi-level net barriers was stud-
ied in the Merdenson torrent in the Canton of Valais,
Switzerland. Three barriers were installed in series in
2006. During the following winter in January 2007, de-
bris flows filled the barrier systems. The total retained
volume was 800 m
3
as determined by 3D topographic
measurements before and after the filling event (Fig. 7).
The field tests showed the potential for several bar-
riers in series to increase the retention volume and their
ability to stabilize river bed method. Finally, the long-
time behavior of steel barriers (abrasion, corrosion) was
studied at the test site over a three year period.
RANGE OF APPLICATION OF FLEXIBLE
BARRIERS AGAINST DEBRIS FLOWS
Barrier systems should be located in a relatively
straight torrent section. The torrent’s gradient should be
as low as possible to reduce the impact velocity and to
maximise the retention capacity. The location should be
easily accessible to ensure inspection and debris removal
upstream of the barrier. The bed at the barrier location
should be stable enough to withstand the expected ero-
sion; otherwise the channel bed and the barrier will
require stabilization measures. The banks on both side
of the torrent need to support the anchor loads. After a
debris flow, plastically deformed components must be
replaced; most commonly these are the brake elements.
(3)
(4)
(5)
Fig. 6 - Filled barrier as a check dam and its flow regime
with the change from subcritical flow by a hydraulic
jump to supercritical flow regime (w
eNDeler
, 2008)
Fig. 7 - 3-D Model of the Merdenson torrent with empty
barriers (left) and filled barriers (right)
Fig. 8 - Cross sectional line of the Merdenson torrent
without (blue) and with filled barriers (red)
background image
A. VOLkwEIN, C. wENDELER & G. GUASTI
1098
5th International Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction and Assessment Padua, Italy - 14-17 June 2011
DEVIATION STRUCTURE, REPAIR OF EXI-
STING STRUCTURES
At check dam 25 in the Illgraben in Switzerland
(Fig. 14), the debris flows increasingly eroded the slopes
on the right side of the dam. Two barriers installed in
A gap between the net and the riverbed avoids unwanted
filling through normal bedload or sediment transport.
PROTECTION wITH A SINGLE BARRIER
Figure 9 shows a single barrier system installed in
the Engler torrent of Meiringen, Switzerland. A hospi-
tal is situated beneath a small road. An active landslide
supplies material for small debris flows every period
of heavy rainfall. The function of the barrier is to catch
the material and to slow down and stop the debris
flow. The barrier must be cleaned out after an event.
INCREASED RETENTION USING A MULTILE-
VEL SYSTEM
With several barriers in a row, the retention capacity
can be increased. Figure 10 shows a multilevel system in
the Hasliberg region in the Bernese Alps at the Milibach
torrent. Thirteen barriers were installed in a row and have
a collected retention capacity of approximately 10,000
m
3
of debris. The multilevel system works by successive-
ly filling each barrier in the torrent, should the first barrier
fill to maximum capacity any further material overflows
into the following barrier until the entire system is filled.
The barriers must be cleaned out after an event.
ENHANCEMENT OF A RETENTION BASIN
Figure 11 shows the application of a barrier as
a supplementary structure of a retention basin to in-
crease the retention volume. The barrier is situated at
the Schlucher Ruefe torrent in Liechtenstein.
RIVER BED STABILIZATION
The barriers in the Merdenson torrent, Canton of
Valais, Switzerland (Fig. 12), are intended to stabilize
the river bed. Remaining filled, the step-wise arrange-
ment of the filled barriers leads to an energy loss of the
debris flow regime. The barriers remain filled after an
event. Static loads and corrosion have to be considered.
Fig. 9 - Debris flow barrier in the Engler torrent, Berner
Oberland, Switzerland
Fig. 10 - Multi-level debris flow barriers in Hasliberg re-
gion, Switzerland
Fig. 11 - Debris flow barrier at the Schlucher Ruefe torrent,
Liechtenstein
Fig. 12 - Filled multi-level barriers in the Merdenson tor-
rent, Switzerland
Fig. 13 - Net barriers as a repair and deviation construc-
tion in the Illgraben, Switzerland
background image
DESIGN OF FLEXIBLE DEBRIS FLOW BARRIERS
Italian Journal of Engineering Geology and Environment - Book www.ijege.uniroma1.it © 2011 Casa Editrice Università La Sapienza
1099
far we propose the presented design concept as a basis
for a design guide. We welcome further discussion and
suggestions for its optimization. The proposed design
concept is of course limited to debris flow that can be
covered by the proposed barriers. Huge events in very
wide channels or with enormous flow heights are not
considered yet and a design concept has first to be vali-
dated for such dimensions. But we hope that the future
international developments in this area will improve
the knowledge and the capabilities of such barriers.
LIST OF SYMBOLS
two stages remain filled with debris. The retention of
material redirects subsequent debris flows back over the
repaired concrete check dam. Both barriers were filled
with a natural debris flow and will remain filled.
PROTECTION AGAINST SCOUR
Figure 14 illustrates the consequences of debris
flow scour to the base of a check dam in the Merden-
son torrent. A net barrier was directly installed in front
of the dam. It now acts as protection for the dam toe.
CULVERT BLOCkAGE
In front of culverts where debris and drift wood
is expected, a net can be installed to protect culverts
from blockage (Fig. 15).
DRIFTwOOD RETENTION
The barriers can be applied in torrents to catch drift-
wood. The load distribution is similar to debris flow load-
ing. For the driftwood load case, a different dimension-
ing concept has to be applied (Fig. 16; R
imböCk
, 2003).
SUMMARY AND CONCLUDING REMARKS
A design concept has been proposed that allows
the dimensioning of flexible barriers to mitigate debris
flows. The design load is a debris flow with a certain
flow depth and represents a worst case scenario in-
cluding smaller loads from sediment filling (if there is
no gap between the net and riverbed) or flood events.
Since no useable design concepts are available so
Fig. 14 - Scoured check dam base (left) and
protected check dam base by a
naturally filled net barrier
Fig. 15 - Net barrier against culvert blockage installed di-
rectly in front of a culvert, The Narrows, CA
Fig. 16 - Net barrier against driftwood
background image
A. VOLkwEIN, C. wENDELER & G. GUASTI
1100
5th International Conference on Debris-Flow Hazards Mitigation: Mechanics, Prediction and Assessment Padua, Italy - 14-17 June 2011
ACKNOWLEDGEMENTS
This research was financially supported by the
Swiss Federal commission for Technology and Inno-
REFERENCES
b
eRGmeisteR
k., s
uda
J., H
übl
J. & R
udolf
-m
iklau
F. (2008) - Schutzbauwerke gegen wildbachgefahren. Technical report,
Betonkalender.
b
uwal
(2007) - Lawinenverbau im Anbruchgebiet. Technical guideline, Berne.
d'a
Gostino
v., C
eRato
m. & C
oali
R. (1996) - Il transporto solido di eventi estremi nei torrenti del Trentino Orientale, In:
Internationales Symposium; Interpraevent 1996, Garmisch-Partenkirchen, Band 1: 377-386
d
e
n
atale
J.s.
et
Alii
(1996) - Response of the Geobrugg cable net system to debris flow loading. Report. California Polytechnic
State University.
d
uffy
J.d. (1998) - Case studies on debris and mudslide barrier systems in California. Proc. One Day Seminar on Planning
Design aod Implementation of Debris Flow and Rockfall Hazards Mitigation Metbods: 77-90, Hong Kong.
G
ReGoRetti
C. (2000) - Estimation of the maximum velocity of a surge of debris flow propagating along an open channel.
Interpraevent20000, Villach
H
amPel
R. (1980) - Geschieberechnung für Gefahrenzonenpläne in wildbachgebieten. Internationales Symposium; Interpraevent
1980, Bad Ischl, Band 3: 83-91
l
ien
H. (2003) - Design of slit dams for controlling stony debris flows. Journal of sediment research, 18(1): 74-87.
m
itzuyama
et
Alii
(1992) - Prediction of debris flow peak discharge. Proc. Interpraevent, 4: 99-108, Berne.
m
uRaisHi
H. & s
ano
s. (1999) - Full scale rockfall test of ring net barrier and components. Seminar of Rockfall Tests and
Standardization, Davos.
e
Gli
t
H
. (2005) - Objektschutz gegen gravitative Naturgefahren, Vereinigung Kantonaler Feuerversicherungen VKV.
ö
steRReiCHisCHes
n
oRmunGsinstitut
(2009) - Schutzbauwerke der wildbacherbauung - Bemessung und konstruktive,
Durchbildung. ONR 24802.
PWRI (1988) - Technical Standard for measures against debris flows (draft). Ministry of Construction, Japan.
R
iCkenmann
d. (1999) - Empirical relationships for debris flows. Natural Hazards, 19(1): 47-77.
R
iCkenmann
d. (2001) - Estimation of debris flow impact on flexible wire rope barriers. Internal report WSL, Birmensdorf.
R
iCkenmann
d. & z
immeRmann
m. (1993) - The 1987 debris flows in Switzerland: Documentation and analysis. Geomorphology,
8: 175 - 189.
R
imböCk
a. (2003) - Schemmholzrückhalt in wildbächen. Diss. Techn. Univ. Munich.
v
olkwein
a. (2004) - Numerische Simulation von flexiblen Steinschlagschutzsystemen. Diss, ETH Zurich.
w
endeleR
C. (2008) - Murgangrückhalt in wildbächen - Grundlagen zu Planung und Berechnung von flexiblen Barrieren. Diss.
ETH Zurich.
w
endeleR
C., v
olkwein
a. & b
aRtelt
P. a. (in prep.) - Structural design of flexible steel barriers for torrent debris flow mitigation.
vation (CTI). We further thank the reviewers who did
an excellent job.
Statistics