Journal of Japan Society of Civil Engineers, Ser.

B1of(Hydraulic
Annual Journal HydraulicEngineering), Vol. 68,
Engineering, JSCE, No. 4,2012,
Vol.56, I_1-I_6, 2012.
February

NUMERICAL ANALYSIS OF DEBRIS FLOW

DEPOSITION ON BREAKER STRUCTURE

Yeonjoong Kim1, Hajime NAKAGAWA2, Kenji KAWAIKE3,

and Hao ZHANG4
1Student Member of JSCE, Doctoral Student, Department of Civil and Earth Resources Engineering, Kyoto
University (Katsura Campus, Nishikyo-ku, Kyoto 615-8540, Japan)
2Member of JSCE, Dr. of Eng., Prof., DPRI, Kyoto University (Yoko-oji, Fushimi, Kyoto 612-8235, Japan)
3Member of JSCE, Dr. of Eng., Assoc. Prof., DPRI, Kyoto University (Yoko-oji, Fushimi, Kyoto 612-8235, Japan)
4Member of JSCE, Dr. of Eng., Assist. Prof., DPRI, Kyoto University (Yoko-oji, Fushimi, Kyoto 612-8235, Japan)

Debris-flow breakers have advantage not only to reduce the energy but also to create suitable narrow
area, cost-efficient, simply designed, easily repaired and maintained. It is known that two phenomena
occur when a debris flow crosses the debris-flow breaker. First, infiltration happens as the process of
deposition in the debris flow due to water drains through the permeable deck. Second, the pore water
pressure changes on the debris-flow breaker. In this paper, fundamental experiments and numerical
simulation are conducted to investigate debris-flow breakers. In addition, a methodology is proposed to
assess the suitability of a variable deck shape and change of pressure on the deck according to separation
of water. Numerical model to calculate travel length and deposit thickness of different sediments depends
on opening size and blocking size. As a result, the optimum opening sizes for sediment A, B and C are
0.3cm, 0.3cm and 0.4cm for blocking size 1cm. Furthermore, the simulated results of the travel length
and deposit thickness on the deck are also compared with experimental results.
Key Words : debris flow, infiltration, optimum opening size, pore water pressure, debris-flow breaker

1. INTRODUCTION Debris-flow breakers are designed to reduce debris

Usually, sediment control structures temporarily
stores the excess sediment in the upstream pocket of
sabo dam and reduce discharge safely. The capacity
of sabo dam to control sediment is determined by
sediment storage capacity between the stable slope
and the temporary slope of accumulated sediments.
Therefore, sabo dam should control the increasing
amount of sediment discharge due to gradually
accelerating of debris flow. Using the debris- flow
breaker at upstream of a sabo dam, could be more
effective to control sediment discharge than without
debris-flow breaker. The peak discharge of the flow
must have been effectively reduced and the flow
converted to a less-harmful level because of the
reduced size of the boulder dam1). Debris-flow
breakers have advantage not only to reduce the
energy but also to create suitable narrow area, cost-
efficient, simply designed, easily repaired and
maintained if their size and location are well
planned before construction2)(Fig. 1).
flow energy3) 4). By slowing and depositing the surge
front of the debris flow, downstream reaches the
stream channels and settlement areas which are
exposed to considerable dynamic impact. In an array
of different debris-flow mitigation structures,
debris-flow breakers are always in the most
upstream position. A debris-flow breaker should
retain at least the volume of surge wave. Numerous
functional structures with modern sediment
management systems could be installed downstream
of the breakers.
It is thought that two phenomena occur when a
debris flow crosses the debris-flow breaker: the pore
(mud) water drains through the deck of the debris-
(a) Debris-flow breaker (b) Suitable narrow area
Fig.1 Debris flow breaker

I_1
 Fig.2 Experimental flume setup

flow breaker and the pore water pressure near the Table 1 Properties of sediment material
deck changes. Drainage of the pore water through Sediment (mm) (mm)
the deck increases the sediment concentration of the Sediment A 1.783 10.871 0.341
debris flow increasing the bottom shear stress of the
Sediment B 2.304 11.142 0.353
debris flow. Because the deck of the debris-flow
breaker is open to the air, the pore water pressure of Sediment C 3.054 11.163 0.363
the debris flow near the deck decreases Table 2 Experimental condition
instantaneously5). Experiment No. Opening size (cm) Blocking size(cm)
Watanabe, et al.6) has shown that the spacing of Case0-A,B,C 0.0 0.0
the posts has effects on the trapping capacity of a Case1-A,B,C 0.2 1.0
slit dam. When the relative spacing / 2.0,
Case2-A,B,C 0.4 1.0
where is the spacing of the posts and is the
Case3-A,B,C 0.6 1.0
maximum diameter of the debris flow, the volume
Case4-A,B,C 0.4 3.0
of the debris flow could be reduced by 50% during
peak time. The above studies validated the Case5-A,B,C 0.4 6.0
effectiveness of open-type dams in the prevention of the simulated results of the travel length and deposit
debris flow. They all only considered the relative thickness on the deck are also compared with
spacing factor in designing the spacing of open-type experimental results.
dams.
In this paper, fundamental experiments and 2. LABORATORY EXPERIMENTS
numerical simulation are conducted to investigate
debris-flow breakers. A methodology proposed to A rectangular flume of 5m long, 10cm wide and
assess the suitability of a variable deck shape and 13cm deep is used for the experiments. The slopes
change of pressure on the deck according to of flume are set at upstream with 18° and
separation of water. A numerical simulation is downstream with 7° . The details of experiment
applied to not only new bed kinetic boundary setup are shown in Fig. 2. Silica sand (S1, S2, S3,
condition but also new relation of factor. As a S4, S5, S6) and gravel (G1) are mixed in equal
result, it is able to verify the impact of different deck proportion by weight to prepare the bed sediment-A.
shapes according to variation of opening size. In Silica sand (S1, S2, S3, S4, S5, S6) in proportion
addition, it is possible to decide the optimum (1.6, 1.5, 1, 1, 1, 0.7) and gravel (G1) in (1.7) by
opening size by numerical simulation. Furthermore, weight are mixed to prepare the bed sediment-B.
Silica sand (S1, S2, S3, S4, S5, S6) in proportion
(2.6, 1.7, 0.7, 0.7, 0.7, 0.6) and gravel (G1) in (2.0)
by weight are mixed to prepare the bed sediment-C.
Fig.3 shows particle size distribution of the prepared

Fig.3 Grain size distribution curve of sediment materials Fig.4 Observation points on the debris-flow breaker

2
I_2
 Fig.5 Pore water pressure distribution on the debris flow

material for bed sediment-A, bed sediment-B and Fig.6 Variation of travel length depend on opening size
bed sediment-C. The bed sediment with 1.9m long
Table 3 Permeability of deck
and 7cm deep is positioned 2.8m upstream from the
opening area Blocking size
outlet of the flume by installing a partition of 7cm in Ks 
total area 1cm 3cm 6cm
height to retain the sediment. This sediment bed is
saturated by water. Properties of sediment material 0.2 0.167 0.063 0.032
Opening
and experimental condition are shown in Table 1 size 0.4 0.286 0.118 0.063
and Table 2. Maximum sediment concentration at (cm) 0.6 0.375 0.167 0.091
bed ∗ 0.65 , angle of repose tan 0.7 and
sediment density 2.65g/cm are used. Debris learn that we have to consider the impact of shape of
flow is produced by supplying a constant water deck using experimental results. Table 3 show that
discharge of 300cm /sec for 10sec from upstream permeability of deck depends on blocking size and
end of the flume. Debris flow produced in the opening size. The result of the experiment shows
experiments is the fully stony type debris flow and reduction of the travel length (Fig. 6) by different
the largest particles are accumulated in the forefront. opening size with the fixed blocking size (1cm).
To measure the thickness of deposition (i.e. the flow Moreover, reduction rate of travel length (Fig. 7)
depth plus the deposition thickness in the final shows the similar reduction rate due to the
stage) accurately, a scale bar is used in each point difference between blocking size 1 and 3cm. By
(Fig. 4). introducing consider to variation of deck shape is
described as follows.
3. NUMERICAL MODEL 1 (on the bed )
 
    K s  lo  lb  (1)
1      (on the deck )
W 
(1) Change pore water pressure
The debris-flow breaker is a simple engineering   Kd W

structure which filters fine sediment with water and
traps the course debris on a horizontal screen. It was
designed to separate coarse clastic debris from water
with a fine debris matrix so that the water passes
through the breaker board while the coarse debris
flow is trapped1). When the debris flow reaches the
deck, the infiltration occurs rapidly which can
changes the pore water pressure (Fig. 5). But until
now, the mechanism of the debris-flow breaker
structure has not been well explained. Previous
study5) suggested that change in pore water pressure
is due to the factor using 1D dynamic model.
Previous study considered only permeability of
deck and porosity of sediment. But it was able to
T .L(each cases )
Reduction rate  (2)
T .L( without structure)
where is the permeability of deck, is the
porosity in debris flow, are the opening
size and blocking size in the deck, is the width in
the deck and is the constant coefficient (0.09), T.L
is the travel length(i.e. the maximum length at final
stage) in the deck. Fig. 8 shows that relation factor
depends on variation of blocking size and opening
size using Eq.(1).
(2) Governing Equations
The basic equations used to compute the behavior
of flow motion of debris flow are the two-
dimensional momentum equations, continuity
equation of flow, continuity equation of sediment

Fig.7 Reduction rate of travel length for sediments A, B and C:

fixed opening size(0.4cm) with variable blocking size case Fig.8 Variation of depend on blocking size (Sediment B)

3
I_3
and river bed surface equation7). The pore water (3) Bottom shear stress
pressure will be changed on the debris-flow breaker In the upstream region of a debris-flow breaker,
due to factor relationship. Momentum equations sediment concentration is higher than that of
of pressure term could be integrated assuming the equilibrium state and becomes maximum
kinetic boundary conditions at bed( and concentration due to existence of the deck, and the
water surface 0 . But, bed kinetic boundary yield stress exceeds the driving force, then debris
condition is not on the deck. To consider flow stops and deposition occurs, before filling up
change of the pore water pressure the bed kinetic upstream of the deck. This mechanism of deposition
boundary condition could be changed from to is incorporated in momentum equation of the flow
at the deck. By introducing these pore water mixture as considering yield stress in bottom shear
pressure at the debris-flow breaker, the depth-wise stress. For a fully developed stony debris flow
averaged two-dimensional momentum equations of ( CL  0.4C* );
debris flow for the x–wise (down valley) and y-wise u
(lateral) directions are described as follows.  bx   yx  f bu u 2  v 2 (9)
M  (uM )  ( vM )  ( zb  h) u v2 2
   gh sin  bx 0  gh cos  bx 0
t x y x v
 by   yy  f b v u 2  v 2 (10)
zb  bx u 2  v2
 ( gh  gh)(sin bx0  cos bx0 ) (3)
x T
in which  yx and  yy are the yield stresses in x
N  (uN )  ( vN )  ( zb  h)
   gh sin  by 0  gh cos  by 0
t x y y and y directions, which can be expressed by using
zb  by constitutive equations of Takahashi et al. (1997) 8) as
 ( gh  gh)(sin yx0  cos by0 ) (4) follows:
y T
The continuity equation of the total volume is  yx  f (CL )     C L gh cos  x tan  (11)
h M N  yy  f (CL )     C L gh cos  y tan  (12)
   ib (5)
t x y  CL  C3
 ; CL  C3
The continuity equation of the coarse particle f (CL )   C*  C3 (13)
fraction that is sustained in the flow by the action of 0 ; CL  C3

particle encounters is
 C L h   C L M   C L N   ib C* L (ib  0) where x and  y are the x and y components of
   (6)
t x y i b C* DL (ib  0) slope of the bed surface. is the limitative
The continuity equation for fine particle fraction that sediment concentration(0.48). The coefficient of
is suspended in the interstitial fluid by the action of resistance, fb , is described as
turbulence is 2
1 ( /  )  dm 
1  CL  CF h  1  CL  CF M   1  CL  CF N  fb    (14)
t

x

y 
8  C C 1 3  1   h 
 * L

2



(7)
 i 1  C* L  C* F ; (ib  0) For an immature debris flow ( 0.02  CL  0.4C* );
 b
ib 1  C*DL  CF ; (ib  0) 2
T  d m 
where and are flow flux in ,  bx    u u 2  v2 (15)
0.49  h 
directions, are the mean velocity, is flow 2
depth, is erosion 0 0 or deposition T  d m 
 by    v u 2  v2 (16)
0 velocity, is the sediment concentration in the 0.49  h 
flow, ∗ is maximum sediment concentration in the For a turbulent flow ( CL  0.02 );
bed, is momentum correction factor equal to 1.25
for stony debris flow, is the acceleration due to  gn2 u u 2  v 2
 bx  (17)
gravity, is bed slope, and are bottom shear h1 / 3
stress, is mixture density 1 ,  gn 2 v u 2  v 2
is density of the sediment particle, is density of  by  (18)
h1 / 3
the water.
z b
 ib  0 (8) (4) Erosion and deposition velocity equations
t The erosion and deposition velocity that have been
where is erosion or deposition thickness of the given by Takahashi et al.9) are used as follows.
bed measured from the original bed surface Erosion velocity, if ;
elevation.

4
I_4
 (a) Flow and sediment discharge (Case0-B) (b) Sediment concentration (Case0-B)

(c) Case2-B (d) Case4-B (e) Case5-B

Fig.9 Numerical and experimental results

C  C L u2  v2 h laboratory experiments. To measure the thickness of

ib   e (19) accurately, a scale bar is used in each point.
C*  C  dm
To verify the model, the simulated results of
Deposition velocity, if ; outflow discharge and sediment discharge at the
 u 2  v2 C C downstream end of flume without debris-flow
ib   d 1    L
u 2  v2 (20)
 pU e  C* DL breaker are used. Comparison of deposition
  thickness and travel length are made using
where 2/3 is numerical constant and is observation points by each case. In Fig. 9 (a) and (b)
the equilibrium velocity at which neither erosion nor which are the temporal variations of flow, sediment
deposition takes place as follows: discharge and sediment concentration in the case of
1/ 2  1/ 3 
2  g sin  e   m   C* DL   bed sediment-B, (c), (d) and (e) are the comparison
Ue   C L  (1  C L )     1h 3 2
5d m  a i sin  i     of deposition thickness on the observation points of
 C L  
deck. In Fig.9 (e), simulation results are larger than
(21)
experiment results. The reasons for this study did
where channel slope in which coarse sediment
not consider separated fine sediment through the
concentration is in equilibrium, which can be
opening size. Fig. 10 shows the correlation
obtained as follows.
coefficient of deposition thickness for all cases. The
C    m  tan 
tan  e  L (22) observed results of the deposition thickness were
C L    m    m almost equal to the numerical ones. But, simulation
where is internal friction angle of sediment. was slightly larger than experiment. The simulation
shows a good agreement between observed
4. RESULTS AND DISSCUSION deposition thickness and travel length (Table 4).
The results of experiments have a reduction of
Fundamental experiment and numerical range from minimum 34%(Case5-B) to maximum
simulation are conducted to investigate debris-flow 65%(Case3-A) by compare Case0-A,B,C with each
breakers. The parameters of the simulation are as another case. An analysis of Table 4 shows that
follows; Δ 2cm, Δ 1cm, Δ 0.001sec. changes in the reduction of travel length depend on
The effectiveness of deck shape in a debris flow opening size and blocking size. The above result
fan was investigated through numerical model and shows that we have to consider shape of deck.

(a) Fixed blocking size with variable opening size case (b) Fixed opening size with variable blocking size case
Fig.10 Correlation coefficient of deposit thickness between simulation and experimental results

5
I_5
Table 4 Results of travel length (unit : cm)
Sediment A Sediment B Sediment C
No.
Exp Sim Exp Sim Exp Sim
Case0 80 70 70 61 65 57
Case1 42 37 45 35 41 34
Case2 42 33 34 30 26 28
Case3 29 32 30 29 28 26
Case4 42 33 38 30 32 28
Case5 49 43 46 41 40 39

Because reduction rate of travel length was almost

equal to the results comparing the experimental and
simulation results from Case2(blocking size, 1cm)
to Case4(blocking size, 3cm). But, Case5(blocking
size, 6cm) was confirmed that not cut the mustard as
effects of deck(Fig. 11). From this results, in the
case of blocking size 1cm: it is verified the effect of
suitable opening size by reduction rate of travel
length. As a rough result, the optimum opening sizes
for sediment A, B and C are 0.3cm, 0.3cm and
0.4cm for blocking size 1cm (i.e. a select positive
integer) in Fig.12. Using reduction of rate analysis,
the following equation was deduced from the
experimental results:
Lo
 1 .3 (23)
Dmean
where are the opening sizes and
mean diameter of sediment. In which the estimated
values of the parameters was applied for reduction
rate of travel length with the range around 0.5. In
this equation have no consideration of variation of
blocking size. It is necessary to consider that
variation of deck shape in the next study.
In study was evaluated the effectiveness of debris-
flow breaker using relation of factor. Fixed value
of was used in this model. But this factor might
be changing the porosity of sediment and mean
diameter in debris flow. By considering the
depositing processes on the fan area, we could more
clearly calculate on debris-flow breaker.
5. CONCLUSIONS
This study has shown that a permeable debris-flow
breaker effectively reduces the travel length and
deposit thickness. It is possible to quantitatively
Fig.12 Reduction rate of travel length for sediments A, B and C:
fixed blocking size(1.0cm) with variable opening size case
Fig.11 Reduction rate of travel length
estimate the change in the pore water pressure using
factor relation and determined the optimum
opening size. In the next step of our work is to
clarify the relationship of the value with the
sediment concentration by volume. It is also
necessary to modify mathematical model for the
particle size distributions and consider the
infiltration, separate water and variation of deck
shape.
ACKNOWLEDGMENT: This research is
supported by JSPS AA Science Platform Program
(Coordinator: H. Nakagawa).
REFERENCES
1) Suwa, H., Okano, K, and Kanno, T.: Behavior of debris flows
monitored on test slopes of Kamikamihorizawa Creek,
Mount Yakedakem, Japan, International Journal of Erosion
Control Engineering, Vol.2, No.2, 2009.
2) ICHARM: Debris-flow dewatering brakes: a promising tool
for disaster management in developing countries,
International Center for Water Hazard and Risk
Management Newsletter, Vol.3, No3, pp.10, 2008.
3) Kettl, W.: Vom Verbauungsziel zur Bautypenentwicklung -
Wildbachverbauung im Umbruch. Wildbachund
Lawinenverbau, 48, Jahrgang, Sonderheft, pp.61-98, 1984.
(in German)
4) Fiebiger, G.: Structures of debris flow countermeasures,
American Society of Civil Engineers, pp.596-605, 1997.
5) Gonda,Y.: Function of a debris-flow brake, International
Journal of Erosion Control Engineering, Vol.2, No.1, 2008.
6) Watanabe M., Mizuyama, T, and Uehara, S.: Review of
debris flow countermeasure facilities, J.of the Japan Erosion
Control Engineering Society, Vol.115, pp.40-48, 1980.
7) Shresta, B.B, Nakagawa, H., Kawaike, K., Baba, Y. and
Zhang, H.: Numerical simulation on debris-flow with
driftwood and its capturing due to jamming of driftwood on
a grid dam, Annual Journal of Hydraulic Engineering,
JSCE, Vol.52, pp.139-144, 2008.
8) Takahashi, T., Nakagawa, H., Harada, T., and Yamashiki, Y.:
Routing debris flows with particle segregation, J. of Hyd.
Eng., ASCE, Vol.118, No.11, pp.1490-1507, 1992.
9) Takahashi, T., Satofuka, Y., and Chishiro, K.: Dynamics of
debris flows in the inertial regime, Proc., 1st Conf. on
Debris-Flow Hazards Mit.: Mech., Pred., and Assessment,
pp.239-248, 1997. (Received September 30, 2011)

6
I_6