Earth Surf Processes Landf - 2015 - Cui - Experimental Analysis On The Impact Force of Viscous Debris

EARTH SURFACE PROCESSES AND LANDFORMS
Earth Surf. Process. Landforms 40, 1644–1655 (2015)

© 2015 The Authors. Earth Surface Proccesses and Landforms published by John Wiley & Sons, Ltd.
Published online 24 May 2015 in Wiley Online Library
(wileyonlinelibrary.com) DOI: 10.1002/esp.3744
Experimental analysis on the impact force of

viscous debris flow
Peng Cui,1,2* Chao Zeng1,3 and Yu Lei1,4
1
Key Laboratory of Mountain Hazards and Earth Surface Processes/Institute of Mountain Hazards and Environment, CAS, Chengdu,
610041, China
2
CAS Center for Excellence in Tibetan Plateau Earth Sciences
3
Sichuan Geomatics Center/Sichuan Engineering Research Center for Emergency Mapping & Disaster Reduction, Chengdu,
610041, China
4
University of Chinese Academy of Sciences, Beijing 100049, China
Received 26 June 2014; Revised 19 March 2015; Accepted 23 March 2015
*Correspondence to: Peng Cui, Institute of Mountain Hazards and Environment/Key Laboratory of Mountain Hazards and Earth Surface Processes, CAS, Chengdu,
610041, China. E-mail: pengcui@imde.ac.cn
The copyright line for this article was changed on 6 July 2015 after original online publication.
This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in
any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.
ABSTRACT: A miniaturized flume experiment was carried out to measure impact forces of viscous debris flow. The flow depth
(7.2–11.2 cm), velocity (2.4–5.2 m/s) and impact force were recorded during the experiment. The impact process of debris flow can
be divided into three phases by analyzing the variation of impact signals and flow regime. The three phases are the sudden strong impact
of the debris flow head, continuous dynamic pressure of the body and slight static pressure of the tail. The variation of impact process is
consistent with the change in the flow regime. The head has strong–rapid impact pressure, which is shown as a turbulent-type flow; the
body approximates to steady laminar flow. Accordingly, the process of debris flows hitting structures was simplified to a triangle shape,
ignoring the pressure of the tail. In order to study the distribution of the debris flow impact force at different depths and variation of the
impact process over time, the impact signals of slurry and coarse particles were separated from the original signals using wavelet
analysis. The slurry’s dynamic pressure signal appears to be a smooth curve, and the peak pressure is 12–34 kPa when the debris flow
head hits the sensors, which is about 1.54 ± 0.36 times the continuous dynamic pressure of the debris flow body. The limit of application
of the empirical parameter α in the hydraulic formula was also noted. We introduced the power function relationship of α and the Froude
number of debris flows, and proposed a universal model for calculating dynamic pressure. The impact pressure of large particles has the
characteristic of randomness. The mean frequency of large particles impacting the sensor is 210 ± 50–287 ± 29 times per second, and
it is 336 ± 114–490 ± 69 times per second for the debris flow head, which is greater than that in the debris flow body. Peak impact
pressure of particles at different flow depths is 40–160 kPa, which is 3.2 ± 1.5 times the impact pressure of the slurry at the bottom of
the flow, 3.1 ± 0.9 times the flow in the middle, and 3.3 ± 0.9 times the flow at the surface. The differences in impact frequency indicate
that most of the large particles concentrate in the debris flow head, and the number of particles in the debris flow head increases with
height. This research supports the study of solid–liquid two phase flow mechanisms, and helps engineering design and risk assessment
in debris flow prone areas. © 2015 The Authors. Earth Surface Proccesses and Landforms published by John Wiley & Sons, Ltd.
KEYWORDS: viscous debris flow; dynamic pressure of slurry; particles impact process; depth variation; experimental study
Introduction phase vitally influences the motion and impact force of a

debris flow. Many researchers have been concerned about
Debris flows are a common phenomenon in mountain areas these differences, especially the dynamic pressure of slurry and
and are characterized by high capacity for sediment trans- the impact force of grains.
port, catastrophic occurrences, high-concentration sediment, Few field experiments have been carried out to detect the im-
wide range of grain size, high velocity and short period of pact force of a debris flow since the 1970s. An observation sys-
movement (Cui, 1999), which is the typical solid–liquid tem was devised at Mt Yakedake in Japan in 1972 that recorded
two-phase flow (Fei and Shu, 2004). The fluid phase of a the dynamic pressure of volcanic muddy flow at about 130
debris flow is composed of slurry or water mixed with clay, kPa, and the big boulder impact pressure was estimated to be
silt and sand, which surges down as a continuous fluid. about 20 times the dynamic pressure (Suwa et al., 1973). Field
The solid phase is coarse-grained material, which moves as observations have been conducted three times in Jiangjia
suspended load or bed load in the fluid phase (Qian and Ravine, China from 1973 to 2004 (Hu et al., 2006). In 1973,
Wang, 1984). The difference between the solid and fluid the dynamic pressure was record at about 190 kPa and large
10969837, 2015, 12, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/esp.3744 by National Institutes Of Health Malaysia, Wiley Online Library on [20/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
IMPACT FORCE OF DEBRIS FLOW 1645
particle impact pressure was 920 kPa (Zhang and Yuan, 1985). have realized that the variance of the impact force is due to the
In 2004, the dynamic pressure of a debris flow in Jiangjia Ravine distribution of grains in the flow (Hübl et al., 2009). It may cause
was 10–50 kPa, and the peak impact loading of grains exceeded a possible risk for prevention systems if the countermeasures are
2600 kPa (Hu et al., 2011). In addition, the impact force can be designed according to the impact value calculated by the models
obtained from the impact trace of debris flow such as from back- (Wendeler et al., 2007; Suda et al., 2009). Therefore, separate
analysis of the indentation and craters in dams arising from the calculation of the slurry impact pressure and the particle load
impact of big boulders (Hungr et al., 1984; Chen and Tang, force of debris flows is useful for designing countermeasures
2006) and calculation of the critical condition of bridges and build- against debris flows, and need to be further investigated.
ings damaged by debris flows (Zanchetta et al., 2004). Although In this work, the signals of the impact force of viscous debris
field measurement is an efficient way to obtain the impact force flows were recorded in a miniaturized flume experiment. The
of debris flow, there are still many limits to monitoring and dynamic pressure of slurry and the impact pressure of particles
modeling the behaviors of debris flow impact. For instance, the were separated from the original debris flow impact signals by
occurrence frequency of debris flow events is hard to predict, wavelet analysis. The distribution and processes of debris flow
and it is very difficult to install and to operate measurement devices impact force at different depths were also analyzed. The key
during field observations. And for large scale experiments, the ex- elements obtained in this study will help engineering design
pected volume of the event makes the design cost-intensive and and risk assessment.
flow material parameters are often lacking (Bugnion et al., 2012).
Therefore, miniaturized flume experiments became popular
in studying the impact force of debris flow (Armanini, 1997; Experimental Design
Arattano and Franzi, 2003; Armanini et al., 2004, 2011;
Zanuttigh and Lamberti, 2006; Huang et al., 2007). Much progress Experimental equipment
has been achieved by many researchers. Zanuttigh and
Lamberti (2006) analyzed the impact process of debris flows The experimental equipment consisted of three parts. The first part
on structures of different shapes that reproduce check-dams, was the flume which was 0.2 m wide, 0.4 m deep and 3 m long
filter-dams, buildings and flow-breakers, and compared the im- (a schematic diagram of the flume is shown in Figure 1). The
pact forces obtained in the experiments with available analyti- flume’s length was sufficient to allow the front and body of the
cal formulas. By integrating hydrostatic and dynamic pressures debris flow to reach an asymptotic configuration (constant shape
in their formula, the maximum impact force of debris flows was of the front and of the body) that moves without deformation, and
calculated for a given upstream flow depth and velocity. Some to make sure the flow regime of debris flows is similar to those of
experiments were also carried out to detect the impact force of the Jiangjia Ravine. The bottom of the flume was made from steel
debris flow and granular flow (Huang et al., 2007; Ishikawa plate and the sides were constructed of glass walls to allow optical
et al., 2008), but the impact signals of particle and slurry were measurements and video observations. The sensor support was
not separated, which limits its utilization. Numerical methods made of steel plate 30 mm wide and 180 mm long. In order to
can be adopted to simulate the behavior and impact process of avoid detection error caused by accumulation in front of the
debris flows (Zanuttigh and Lamberti, 2004), but assumptions sensors, the sensor support was fixed at the end of the flume to en-
about the rheological parameters and material concentration able the debris flow to pass the sensor as smoothly as possible. The
may cause uncertainty. Therefore, an accurate analytic method second part was the square tank of 1 m side length, which was
is an effective way to obtain debris flow impact force. For most fixed at the upstream end of the flume. The gate was a steel plate
analytic formulas, the empirical coefficients, the relationship be- which was installed at the bottom of the tank and could be pulled
tween the dynamic pressure and ρv2, need to be estimated with away from the tank suddenly, so that all the material was flushed
laboratory data. However, because most analytical and numeri- down rapidly. There was a chute connecting the gate and flume
cal models have worked on the assumption that particle concen- to let the material flow smoothly. The third part was a pool placed
tration and velocity is uniform at different positions along the flow at the end of the flume to collect the material for reutilization.
depth (Wei, 1996; Zanuttigh and Lamberti, 2006), the impact
force of large particles was accordingly ignored. Therefore, the
models mentioned above are more suitable for a pseudo- Experimental material
homogenous fluid, in which the motion state of particles is similar
to that of fluid (Armanini et al., 2011; Proske et al., 2011), rather In order to simulate the behavior of natural debris flows in the
than the natural debris flow with fluid–solid phases. Some experts field, material from Jiangjia Ravine was used in our
Figure 1. The experimental equipment.
© 2015 The Authors. Earth Surface Proccesses and Landforms published by John Wiley & Sons, Ltd. Earth Surf. Process. Landforms, Vol. 40, 1644–1655 (2015)
1646 P. CUI ET AL.
Figure 2. Cumulative grain size distribution of the debris flow.
experiments. To make sure all particles flow smoothly and pass capture the movement of the debris flow. It had a sampling fre-
freely through the section where the sensor supporting bar is quency of 50 pictures per second. The mean surface velocity
installed, particles with a diameter of more than 20 mm were and the flow regime were obtained by image back-analysis.
removed by screen separation. The composition of the solid For bulk density calculations, samples were collected as the
phase of the material remained constant during the experi- flow front reached the sampling section (the end of flume) dur-
ment, but the liquid concentration was changed in different ing the test, using a cylinder sampler with 15 cm diameter and
tests. The volumetric concentration was varied from 0.34– 12 cm in depth.
0.76.The cumulative grain size distribution of the experimental
sample (solid phase) is shown in Figure 2. The clay, silt, sand,
and gravel content of the material was 2.9%, 5.1%, 37%, and Tests and Data Analysis
55%, respectively. The diameter d50 was estimated at 2.7 mm.
The sorted index (δx = (d84 + d16)/2) describing the range of Tests and the fluid properties
grain size was about 4.7 mm.
A total of 27 tests were carried out with different densities and
slopes (Table I). In order to examine whether the sensors
Measurements system worked well and to guarantee the accuracy of the acquired
data, three tests with clean water were carried out first. An ex-
Debris flow velocity, depth and impact forces were recorded in ample of the impact pressure of water is shown in Figure 4. A
real time during the experiment. The impact detection system further 24 tests for debris flow were developed with flume
composed of five sensors (load cells), a data collection device slopes of 10°, 13° and 15° (eight tests for each slope). The de-
and a computer (Figure 3). The diameter of the sensor was 20 bris flow density was about 1600–2300 kg/m3.
mm with a sampling frequency of 2000 Hz; the high frequency In order to analyze the dynamic behavior of debris flow be-
was to allow detection of the transient impulse of the boulder fore impacting, the dimensionless numbers of Re (Reynolds
impact force and fluid dynamic pressure. In addition, the laser number) and Fr (Froude number) which implicate fluid dynamic
distance measuring instrument (Code: LDS-30-VCR-S12) with a properties of debris flow were obtained, and the dynamic vis-
sampling frequency of 10 Hz was installed upstream of the sen- cosity and yield stress of the debris flow were also calculated.
sors and was used to record the flow depth of debris flow. A It was assumed that the debris flow in the test is a Bingham
camera with high definition was installed above the flume to fluid, the (NRey) number can be estimated with Equation (1)
Figure 3. Impact pressure measurement system. This figure is available in colour online at wileyonlinelibrary.com/journal/espl
Table I. Tests and fluid properties
Density Water Volumetric Total Flume Dynamic Yield Froude

3 3
Test no. (kg/m ) concentration concentration volume(m ) slope(°) viscosity (Pa s) stress (Pa) number
1 1000 1 0.00 0.23 13 0.001 — 5.9

2 1000 1 0.00 0.24 13 0.001 — 5.6
3 1000 1 0.00 0.26 13 0.001 — 5.4
4 2089 0.34 0.64 0.24 10 0.112 34.10 4.5
5 2296 0.21 0.76 0.27 10 14.106 172.05 4.5
6 2148 0.30 0.68 0.26 10 0.223 53.92 4.3
7 2123 0.32 0.66 0.24 10 0.162 44.44 4.7
8 2111 0.33 0.65 0.24 10 0.141 40.50 4.8
9 2093 0.34 0.64 0.23 10 0.116 35.10 4.3
10 2005 0.39 0.59 0.24 10 0.054 17.59 3.2
11 1586 0.64 0.34 0.28 10 0.006 0.38 5.8
12 1829 0.50 0.49 0.22 13 0.02 4.44 4.5
13 2116 0.32 0.66 0.24 13 0.149 42.02 2.5
14 1931 0.44 0.55 0.22 13 0.033 9.85 4.1
15 2113 0.33 0.65 0.21 13 0.144 40.96 2.8
16 2144 0.31 0.67 0.24 13 0.213 52.53 2.6
17 1590 0.64 0.35 0.26 13 0.007 0.59 5.6
18 2240 0.25 0.73 0.24 13 1.297 110.84 4.4
19 2095 0.34 0.64 0.24 13 0.119 35.74 5.2
20 1706 0.57 0.42 0.22 15 0.012 1.70 4.8
21 2051 0.36 0.62 0.22 15 0.078 25.25 3.7
22 1980 0.41 0.58 0.21 15 0.045 14.47 4.1
23 2074 0.35 0.63 0.23 15 0.096 30.32 4.0
24 1961 0.42 0.57 0.23 15 0.04 12.46 4.1
25 1877 0.47 0.52 0.22 15 0.025 6.48 4.4
26 1810 0.51 0.48 0.22 15 0.018 3.84 4.5
27 1849 0.49 0.50 0.22 15 0.022 5.21 4.4
20 which is 0.4 m from the sensor support. The distance of 0.4

Detected signal m between the sensor and detecting section is considered to re-
18
Denoised signal duce the influence of the reflecting wave that propagated
16 Filtered signal upstream.
14 The dynamic viscosity and yield stress of debris flow were
Impact pressure (kPa)
12 estimated with the models proposed by Fei and Shu (2004)

(Equations (3) and (4)). These formulae, based on observed
10
data of debris flows in Jiangjia ravine, were recommended
8 because of their good application to debris flow slurry mixed
6 with particles.
4
2:5 !2:5
S vf S vc
2 μ ¼ μ0 μr ¼ μ0 1 k 1 ′ (3)
S vm S vm
0
-2
0 1 2 3 4 5 6 7 8
Time (s) S v S v0
τ B ¼ 0:098 exp 8:54 þ 1:5 (4)
S vm
Figure 4. Time series of denoised and filtered impact pressure of
clean water in the first test. This figure is available in colour online at where μ is the dynamic viscosity of debris flow (Pa s), and τ B is
wileyonlinelibrary.com/journal/espl
the yield stress (Pa). μ0 is the dynamic viscosity of clean water
(the value is 0.001 Pas for water with a temperature of plus
20°C), μr is relative dynamic viscosity, k is the viscosity modified
(Kang et al., 2004). The Froude number is calculated with coefficient, Svf is the volume concentration of slurry without par-
Equation (2). ticles, the maximum grain size of slurry is 2 mm (Wang et al.,
2003; Kang et al., 2004). Svc is the volume concentration of par-
N Rey ¼ 4ρm vh=μ (1) ticles, Svm is the limit concentration of slurry, which can be cal-
culated with Equation (5), and S′vm is the limit concentration of
pffiffiffiffiffiffi particles in debris flow (0.8). Sv is the concentration of the debris
Fr ¼ v= gh (2) flow, Svo is the critical concentration with the Newtonian flow
transform to Bingham flow, Sv0=1.26Svm3.2.
where NRey is the Bingham Reynolds number, ρm is the density
of debris flow (kg/m3), v (m/s) and h (m) are the velocity and S vm ¼ 0:92 0:2 lg∑pi =d i (5)
depth, μ is the dynamic viscosity (Pas), g is the acceleration of
gravity (9.8 m/s2). The velocity and depth of the surge front where di (mm) and pi are the average diameter and weight ratio
were measured in the same section and at the same time, of particles in the grain size distribution curve.
1648 P. CUI ET AL.
The dynamic viscosity of the debris flow was estimated at represents the threshold of the detailed coefficient to eliminate
0.006–14.106 Pas (Table I), the yield stress at 0.384–172.045 noise; the coefficient of each level was set to zero if it was
Pa, and the Bingham Reynolds number in range 314–448818. greater than the threshold T.
The dynamic viscosity and yield stress grow gradually with
the increase of density when the debris flow density is lower 3. Reconstruct
than 2200 kg/m3. If the density is greater than 2200 kg/m3, Wavelet reconstruction was computed using the original
the two parameters will increase dramatically, and the Bingham approximation coefficients for level 8 and the modified de-
Reynolds number will drop below 1000. The Froude number tail coefficients of levels 1to 8 (Equation (9)):
was calculated at about 2.5–5.9, encompassing the value for
most debris flow types. ^ 8f þ D
A8 f þ D ^ 7 f þ ⋯ þD
^ 1 f ¼ f ðk Þ (9)
Results and Analysis

Denoising and filtering of impact signals
The impact force recorded by load cells was converted into the
Separation of the grain and slurry impact signal
impact pressure of the cross-section facing the debris flow. In
The wavelet method was confirmed for denoising the impact
order to eliminate the random noise resulting from the measur-
pressure signal of the debris flow. The impact pressure recorded
ing system and external disturbances, the impact pressure sig-
before and after the impact process was converted to zeros, and
nals were denoised to obtain the real impact signals of debris
the 99% original signal energy was preserved after the noise re-
flow. The denoised signals contained the slurry pressure and
duction. Examples are provided in Figures 4 and 5. The impact
the grain impact loading. Hu et al. (2011) suggested that the
signals were reconstructed using the original approximation
rectangular and smooth impulse results from a continuous uni-
coefficients of level 8 (A8f) recommended as the dynamic pres-
form fluid pressure while the peak impulse is from individual
sure of debris flow, which means all detailed signals from levels
boulders, and some researchers have testified that the smooth ^ 8f D^ 1 f ) were disposed (filtered).
1–8 (D
impact signals can be extracted using wavelet analysis (Zhang,
By considering the similarity of the water and slurry impact
2009; Yang et al., 2011; Tang et al., 2013). Therefore, the wave-
characteristics, the accuracy of the filtering results for clean
let method was utilized to denoise and filter the debris flow sig-
water impact pressure was confirmed. As Figure 4 shows, the
nals, so as to separate the fluid pressure and grain impulse.
dynamic pressure of water was smooth and the correlative
The original impact signal of a debris flow can be expressed
coefficients of the original and filtered signals reached
as the superposition of the real impact signal and the noise
0.97–0.99, which indicates that the filtered signals closely
(Equation (6)):
represent the real impact signals. The debris flow impact signals
were also filtered in the same way. The impact pressure of the
f ðt Þ ¼ ^f ðt Þ þ ε e ðt Þ; t ¼ 0; 1; ⋯ n 1 (6) grains was obtained by subtracting the dynamic pressure of
the slurry from the original pressure (Figure (6)).
where f(t )is the detected impact signal, ^f ðt Þ is the real impact The velocity was estimated at 4.6–4.9 m/s for clean water
signal, the low frequency part and e(t) is the noise, the high fre- when the flume slope was 13° (Table II), the Froude number
quency component. was 5.4–5.9, and the detected peak dynamic pressure of the
Three steps are necessary to denoise and filter signals (Misiti water was 10–13 kPa. The peak dynamic pressure of the slurry
et al., 1996): was 12–34 kPa, which is greater than that of water. The range
of velocities of the debris flow was 2.4–5.2 m/s and the maxi-
1. Decompose mum depth in front of the sensors was 7.2–11.2 cm. The impact
In order to obtain a smooth and continuous dynamic pressure pressure of most big grains is much larger than that of slurry,
for the fluid phase, the Daubechie (db4) function was chosen and is about 40–160 kPa.
as the wavelet basis function because of orthonormality.
Then, the original signals were decomposed into eight levels
with the basic function (Equation (7)):
55
f ðt Þ ¼ A8 f þ D8 f þ D 7 f þ ⋯ þD1 f (7) 50
Detected signal
Denoised signal
45 Filtered signal
where A8f represents the smooth signal which implies a low
frequency component, A8 is the approximation coefficient, 40
D8f–D1f represents the grain impulse signal, which implies a 35

high frequency component mixed in the noise, and D8–D1 30
are the detailed coefficients.
25
2. Threshold detail coefficients 20

For each level from 1 to 8, the threshold detail coefficients were 15
estimated using Equation (8). The threshold T strongly affects 10
the accuracy of denoised results. In our study, the thresholds
5
for each level were estimated using the principle of the least
standard deviation between the original and denoised signal. 0
0 1 2 3 4 5 6 7 8
(
Dj Dj > T Time (s)
^j ¼
D (8)
0 D j ≤ T
Figure 5. Time series of denoised and filtered impact pressure of de-
bris flow in the fifth test recorded at 1.5 cm above the bottom. This figure
where Dj is the detailed coefficients wavelet, j = 1,2…8 and T is available in colour online at wileyonlinelibrary.com/journal/espl
50
force. Therefore, the change of flow regime during the impact
45 processes is discussed first. The Bingham Reynolds number
(Re) was used to predict the flow regime in our study.
40
Table I shows that the Bingham Re in our experiment was
35 0.003 × 105–4.49 × 105. According to the field observation
30
and calculation of the flow regime in the Jiangjia Ravine,
viscous debris flow will become turbulent when the velocity
25 reaches 5–7 m/s and the Bingham Reynolds number is larger
20
than 1.4 × 105–2.06 × 105 (Kang et al., 2004), and it will
become intensely turbulent if the velocity is 7–9 m/s and the
15 Bingham Reynolds number is larger than 2.4 × 105– 8.7×105.
10 By contrasting the Re number estimated in our experiment with
that observed in the field at Jiangjia Ravine, the flow regime and
5 impact processes of debris flow can be divided into three stages:
0
0 1 2 3 4 5 6 7 8
1. The rapid and powerful impact of the debris flow head: the flow
Time (s)
reached two sensors at heights 4.5–6.0 cm. For the debris flow
(a) Time series of impact load of grains densities were greater than 2000 kg/m3, the Bingham Reynolds
recorded at 1.5 cm above the bottom
number of the head was smaller than 2.06 × 105, and
50
the debris flow appeared stable or slightly turbulent. If
45 the densities are lower than 2000 kg/m3, the flow
becomes turbulent or even intensely turbulent. After the
40
short powerful impact, the debris flow rose rapidly, with
the fifth sensor at 13.5 cm above the bottom surrounded
35
by the flow (Figure 7(a)).
30
2. The continuous and steady impact of the debris flow body:
25 the flow buried three sensors to depths of about 7.5 cm
20
(Figure 7(b)). At this stage the turbulence became very
weak, and the debris flow approximated a steady laminar
15 flow; this phenomenon in the experiment is similar to that
10
observed in Jiangjia Ravine by Takahashi (2007).
3. The slip flow of the debris flow tail: the velocity became
5 much lower, and the sensors were subjected to static pres-
0 sure (Figure 7(c)).
0 1 2 3 4 5 6 7 8
Time (s) Because the static pressure of the debris flow tail is much
(b) Time series of impact loads of grains smaller than the rapid impact pressure, the first two steps were
taken as the whole impact process of debris flow and the im-
50 pact signals recorded by the sensors at 10.5 cm and 13.5 cm
45 high were disposed of owing to the run up effect. The impact
signals recorded at 1.5, 4.5 and 7.5 cm were selected to ana-
40
lyze the impact characteristics of the debris flow.
35
30
The impact processes of slurry and the empirical
25 model
20
The typical processes of dynamic pressure
15 The impact time became shorter with the increase of sensor
10
height. The impact processes at the bottom sensor (1.5 cm high)
were sustained for about 5 s, and the impact was sustained for
5 2–3 s for the upper sensor at 7.5 cm high (Figure 8). Based on
0 the record of maximum flow depth, the processes for the
0 1 2 3 4 5 6 7 8 dynamic pressure of slurry can be separated into two parts,
Time (s) the flow before maximum flow depth, process 1, and that after,
(c) Time series of impact loads of grains process 2.
Figure 6. Time series of impact load of grains in the fifth test; left of
1. Process 1 features as the sudden and powerful impact by
the dashed line are impact loads of grains in the debris flow front and the debris flow head, as shown in Figure 8. The impact time
those to the right side are in the debris flow body. This figure is avail- of the process 1 (flow head) was estimated at about 0.8–1s,
able in colour online at wileyonlinelibrary.com/journal/espl from the time when a debris flow touches the load sensor till
the time the flow depth reaches a maximum (Figure 9) (to
the left part of the dashed line). When the debris flow head
The impact processes of debris flow hits the bottom (1.5 cm) and middle sensors (4.5 cm), the
impact force at the two heights increased rapidly, and the
It was found that the flow regime influences the velocity and mean peak impact pressure at the middle sensor was larger
the geometry of the flow, which in turn determines the impact than that at the bottom by 2–3 kPa. The mean peak impact
1650 P. CUI ET AL.
Table II. Impact pressure of debris flow slurry and big grains
Pressure of slurry (kPa) Pressure of grain (kPa)

Density Velocity Depth
3
Test no. (kg/m ) (m/s) (cm) 1# 2# 3# 1# 2# 3#
1 1000 4.86 7.00 11.15 12.54 10.82 — — —

2 1000 4.74 7.20 10.89 13.04 10.61 — — —
3 1000 4.62 7.50 12.50 12.56 10.92 — — —
4 2089 4.59 10.40 21.08 21.03 17.66 75.36 55.19 64.59
5 2296 4.59 10.50 19.92 24.03 17.02 47.41 70.24 60.77
6 2148 4.11 9.40 20.77 21.23 14.48 47.61 93.48 93.71
7 2123 4.59 9.70 18.15 21.94 16.54 56.88 128.09 42.67
8 2111 4.59 9.20 19.20 15.21 15.91 44.55 56.55 49.04
9 2093 4.33 10.20 28.30 27.43 17.33 58.38 48.93 48.93
10 2005 3.00 9.20 13.79 19.20 11.31 34.56 61.28 31.77
11 1586 5.20 8.10 25.23 13.78 11.63 117.44 38.80 34.03
12 1829 3.71 7.00 20.88 19.74 10.88 72.86 54.21 48.88
13 2116 2.36 8.80 18.75 22.04 13.64 49.98 41.01 44.44
14 1931 3.71 8.30 24.81 30.57 18.22 45.82 71.30 49.70
15 2113 2.79 10.30 24.43 32.62 15.38 44.32 51.23 29.01
16 2144 2.69 10.80 27.74 30.10 21.09 35.38 51.36 29.14
17 1590 4.33 6.20 20.50 15.76 16.67 39.80 56.62 37.85
18 2240 4.33 9.80 20.14 25.83 19.65 142.13 91.10 68.57
19 2095 4.88 9.00 31.45 22.84 17.97 87.81 82.49 63.81
20 1706 4.59 9.40 30.87 23.82 15.14 79.85 57.94 59.70
21 2051 3.90 11.20 22.98 28.40 25.46 49.09 88.40 83.14
22 1980 4.11 10.10 26.38 29.30 15.60 168.56 107.72 61.67
23 2074 4.11 10.60 33.13 29.49 26.62 66.05 90.52 69.50
24 1961 4.11 10.20 28.94 25.86 20.58 108.39 74.25 88.04
25 1877 4.33 9.90 24.73 28.40 14.94 160.07 91.05 51.78
26 1810 4.33 9.50 30.95 33.97 19.26 78.31 118.49 80.06
27 1849 3.90 8.20 27.19 24.61 20.69 125.37 102.80 60.19
#1,#2,#3 indicate the three sensors located at different vertical positions, 1.5 cm, 4.5 cm and 7.5 cm from the bottom of the flume, respectively.
(a) The rapid and powerful impact of the debris flow head
(b) The continuous and steady impact of the debris flow body
(c) The static pressure of the debris flow tail on sensors
Figure 7. Impact processes of debris flow on sensors.
40
Dynamic imapct pressure at 1.5cm
35 Dynamic imapct pressure at 4.5cm
Dynamic imapct pressure at 7.5cm
30
25
20
15
10
0
0 1 2 3 4 5 6 7 8
Time (s) Figure 10. Simplification of time series of debris flow dynamic pres-
sure: ① represents the stage of debris flow head impact ② represents
Figure 8. Time series of debris flow dynamic impact pressure at three the stage of debris flow body impact.
depths (1.5, 4.5 and 7.5 cm) during the second test. This figure is avail-
able in colour online at wileyonlinelibrary.com/journal/espl
time is about 4 s and can be represented by a triangle shape. The
top of the structures will be damaged where subjected to higher
24 12
shearing and bending at this stage. Therefore, the impact pro-
22 Dynamic pressure 11
cesses of debris flow on structures can be simplified to the com-
Debris flow depth
20 10 bination of the impulsive forms of two triangles (Figure 10).
18 9
Debris flow depth(cm)
Impact pressure(kPa)
16 8 Empirical model to calculate dynamic pressure

14 7 The hydrodynamic model (Equation (10)) is the most popular
12 6
one for calculating the dynamic pressure of the slurry, because
the debris flow velocity and coefficient α are more easily ac-
10 5
quired. It is mainly derived from fluid momentum balance
8 4
and the Bernoulli equation, which can be applied to calculate
6 3 impact pressure of most fluids, even in the case of pure water
4 2 and saturated mixtures. The key issue for the model is the deter-
2 1 mination of the empirical coefficient.
0 0
0 1 2 3 4 5 6 7 8 9 10
Time(s) P ¼ αρm v 2 (10)
Figure 9. Comparison of debris flow impact pressure (measured by

bottom sensor, 1.5 cm from floor of flume) and debris flow depth vari- where P is the dynamic pressure of debris flow in N/m2, ρm is
ation in the second test. This figure is available in colour online at the density of debris flow (kg/m3), v is the velocity of debris
wileyonlinelibrary.com/journal/espl flow (m/s). The empirical coefficient α is used to represent the
difference of flow regime and proportions of granular composi-
tion. The values of the coefficient have been estimated based
pressure for the three slopes of 10°, 13° and 15° were on laboratory tests and field observations (Table III). However,
28 kPa, 32 kPa and 37 kPa, respectively. a model is still lacking that can be applied to all types of flow
2. In process 2, the depth of debris flow was between 7.2 and or some sort of debris flow.
11.2 cm when the debris flow body reached the sensors.
During this stage, the impact pressure at the bottom Table III. Empirical coefficients in hydrodynamic models
decreased smoothly and the impact pressure at the top
(7.5 cm) reached a peak. The mean impact pressures for Empirical
each of the three slopes were 14 kPa, 19 kPa and 22 kPa. Resource coefficient α Description
In considering the change of flow regime, the impact pro- Zhang (1993) 3–5 Estimated with the data measured in
cesses, and the most vulnerable condition for structures hit by de- Jiangjia Ravine, China, characterized
by viscous debris flow.
bris flows, it is reasonable to speculate that the processes of
Watanabe and 2.0 Estimated with the data measured in
structure damage by debris flow can be simplified into two parts. Ikeya (1981) Nojiri Ravine, Japan, characterized
The first is the debris flow front impact. During the impact process by volcanic debris flow (mudflow).
of the debris flow front, the impact pressures detected at the bot- Hungr et al. 1.5 Back analysis of the stony type debris
tom and middle sensors are maximum. Therefore, the mean of (1984) flow in British Columbia, Canada.
maximum impact pressures at these two sensors is regarded as Mizuyama 1.0 Derived from jet impact theory,
the peak impact pressure Pmax of debris flow front. The stress (1979) applied in Yakedake, Nigorisawa
point of Pmax is relative low and the impact location is much Urakawa debris flow in Japan.
lower than the height of the structure. The bottom of the structure Armanini 0.45–2.2 Estimated with the data measured in
(1997) laboratory experiments, the material
would be damaged first where it is subjected to higher shearing
is a mixture of PVC and water with
and bending. The second part is the continuous pressure of the 3
densities 1080–1300 kg/m .
debris flow body, which is equal to 0.68 ± 0.14Pmax. The impact
1652 P. CUI ET AL.
The Froude number indicates the kinetic energy component ra- estimated at 45 kPa, 23 kPa and 14 kPa, respectively. This indi-
tio between horizontal and vertical directions. The impact force cates that particles may scatter at different depths and big grains
of debris flow is mainly determined by the kinetic energy compo- tend to concentrate in the upper layer.
nent in the horizontal direction. Therefore, some experts have The peak impact pressure of particles at different flow depths
tried to use the Froude number to determine the dynamic pres- was 40–160 kPa. The ratios of the peak grain impact loading
sure (Armanini 1997; Scotten and Trivellato, 1995; Hübl and to the fluid pressure were different at the three depths for the
Holzinger, 2003; Tiberghien et al., 2007). Zanuttigh and Lamberti 24 measurements of the surge: 3.2 ± 1.5 times for the bottom
(2006) proposed a simple expression derived from momentum sensor, 3.1 ± 0.9 times for the middle sensor and 3.3 ± 0.9
balance to calculate the impact force on the structure by intro- times for the surface sensor.
ducing Froude number (Fr). Hübl et al. (2009) also recommended The responses to grain impact pressure at the three depths
that the relationship of the Fr and empirical coefficient (α) was de- were not synchronous. The bottom sensor recorded more grain
termined by a power function (Equation (11)): impacts, but the highest impact frequency occurred on the mid-
dle sensor. The total number of particle impacts on the bottom
α ¼ P max =ρm v 2 ¼ a Fr b (11) sensor was 1048 ± 250 times and as the impact was sustained
for 5 s the frequency was estimated at about 210 ± 50 times per
second; the number of impacts at the middle sensor was 861 ±
where Fr is the Froude number of the debris flow, Pmax is the de-
144, the impact was sustained for 3 s and the frequency was
tected peak impact pressure of the debris flow (N/m2), which is
about 287 ± 29 times per second; for the top sensor these pa-
equal to the mean impact pressure at 1.5 cm and 4.5 cm height
rameters were 452 ± 176 times, 2 s and 226 ± 35 times per
in this paper, ρm is the density of the debris flow (kg/m3), and a
second.
and b are the coefficients.
The grain impact frequency of the debris flow head was
In this paper, a power model was introduced. In order to en-
greater than that of the debris flow body. During the 1 s impact
hance the robustness and extrapolation capability of the regres-
of the debris flow head, the grain impacts at the three depths
sion model, the impact data from previous studies were used.
were 490 ± 69, 493 ± 52 and 336 ± 114 times. These reveal
Some data were obtained from miniaturized flume tests by
that most particles were concentrated in the debris flow head,
Hübl and Holzinger (2003), Tiberghien et al. (2007), Scheidl
and the number may have increased with height.
et al. (2013) and others were estimated from field observed
events by Zhang and Yuan (1985) and Costa (1984) (Figure 11).
The empirical coefficient was estimated at 0.12–18.6, and the
Froude number at 0.5–10.8.
The values of a and b were determined from a power model
Discussion
based on 155 sets of data (Figure 11). The regression hydrody- Miniaturized flume experiments were carried out to measure
namic model is shown in Equation (12): the impact force of viscous debris flow. The detected impact
signals were denoised and filtered using wavelet analysis; this
P ¼ 5:3Fr 1:5 ρm v 2 (12) method performs well for preserving the real impact signals.
The signals for the dynamic pressure of the slurry and for the
Grain impact characteristic impact load of grains were separated successfully. Based on
the separated signals, the impact processes of debris flow and
The grain impact pressure was obtained by subtracting the dy- characteristics of grain impact pressure were studied in detail.
namic pressure of the slurry from the denoised signal (Figure 6). From the practical application of this and its potential contribu-
The grain impact pressure is random and does not increase tion to validating the fluid–solid phase model of our study, the
with fluid pressure. The impulsive load of the grains was re- relationships of the impact characteristics and the engineering
corded at different depths. The load at the top sensor was design and the particle motion mechanism are discussed.
greater than that at the bottom and middle sensors in some ex-
periments. For example, during the fifth experiment, when the
debris flow depth reached 10.2 cm, the impact loads on the
top sensor, the middle sensor and the bottom sensor were The flume test results are consistent with those
recorded in field observations
The impact signals at different depths in the experiment show a
typical impact process of debris flow, and is in accordance with
that recorded in Jiangjia Ravine by Hu et al. (2011), which con-
firms the attempt to approach real debris flow behavior in our
study. The consensus also can be established on separating de-
bris flow impact signals into slurry and big grains for both flume
tests and field observations. The dynamic pressure of slurry was
18–60 kPa in Jiangjia Ravine, when debris flow velocity was
5–10 m/s (Hu et al., 2011), which would be same as the records
in our experiment if the velocities were the same. The grain im-
pact loading is random, and the impact frequency of big grains
was likely to increase with the depth, which indicates that large
grains are prone to concentrate in the surface and middle parts
of the flow, and this phenomenon has also been found in natu-
ral debris flows in Jiangjia Ravine. All these reveal that our tests
Figure 11. Relationships between Froude number and empirical coef- are reliable and applicable for understanding debris flow im-
ficients. This figure is available in colour online at wileyonlinelibrary. pacts. However, flume size still limits simulating the effects of
com/journal/espl real, big boulders.
The hydrodynamic model was improved by Therefore, if we only consider the rapid impact of the debris
introducing Froude number flow front for engineering design, but ignore the continuous im-
pact processes, the risk to structures from the impact of a debris
It was found that the flow regime influenced the way that the flow will tend to be underestimated.
debris flow impacted on the structure. In previous studies, the
empirical coefficients varied from place to place (Table II),
and there was no formula for calculating the dynamic pressure
of slurry with different flow regimes. The dimensionless number
Fr is defined as the ratio of a characteristic velocity to a gravita-
tional wave velocity: it is used to determine the flow pattern
and it is frequently used in the construction of dynamically sim-
ilar models. Therefore, the Fr numbers were introduced to de-
scribe the differences in flow patterns, and to achieve scale
invariance for miniaturized flume tests.
The Froude number estimated from miniaturized flume tests
was in the range 0.6 to 10.8 (Figure 11), which is similar to that
estimated from observed events (which was 0.5–7.6). The
power model performed well in fitting 155 sets of the data from
miniaturized and field experiments (coefficient of correlation
0.91). This indicates that the model proposed in this paper
can be widely applied to the calculation of debris flow
dynamic pressure. However, there are still other types of debris
flow to consider, with Froude numbers lower than 0.5 or larger Figure 13. Building hit by debris flow in Longchi, Du jiangyan,
than 10.8. China. A column was destroyed at the column bottom as debris flow
reached the middle of the column. This figure is available in colour on-
line at wileyonlinelibrary.com/journal/espl
The risk of structure damage may be reduced by
considering the impact processes
The behavior of debris flow towards structures can be summa-
rized as the powerful impact by the debris flow head leading to
structural damage followed by buried damage with the tail. The
processes described below are shown in Figure 12.
1. The structure is hit by the debris flow head with a turbulent

flow regime. The dynamic pressure can be simplified to a
uniform load on the structure, with the maximum stress
point located at 1/3 of the depth. The structure’s bottom is
more vulnerable to shearing stress and bending.
2. Continuous dynamic pressure on the structure leads to
sustainable deformation. The peak pressure appears at
the top of flow and the load distribution is approximately
trapezoidal. The maximum stress point is reached at 2/3
of the depth. In this condition, the top of the structure
would be damaged first due to maximum shearing and
bending. Figure 14. The column was destroyed from its top point as debris flow
3. The discharge and velocity of the debris flow decreases. was higher than the column in Qipan ravine, Wenchuan County, 2013.
The stress point shifts down along the structure and there This figure is available in colour online at wileyonlinelibrary.com/jour-
will be deposits or buried damage. nal/espl
Figure 12. Simplification of the impact processes of debris flow on structures.
1654 P. CUI ET AL.
There is some evidence that the location of the stress point Breakers. In Proceedings of the International Conference on Hydrau-
affects structural damage (Zeng et al., 2014). For example, dur- lics of Dams and River Structures, Yazdandoost F, Attari J (eds). Taylor
ing a building impact by debris flow in Longchi, Du jiangyan & Francis Group: London.
city, China, the bottom of a column was destroyed as the debris Armanini A, Larcher M, Odorizzi M. 2011. Dynamic Impact of a Debris
Flow Front Against a Vertical Wall. In 5th International Conference on
flow front reached the middle of the column (Figure 13).
Debris-flow Hazard Mitigation, Genevois R, Douglas L (eds). Casa
Another building in Qipan ravine, Wenchuan county, China, Editrice Università La Sapienza: Roma; 1041–1049.
was damaged at the top of a column when debris flow depth Bugnion L, McArdel B, Bartelt P, Wendeler C. 2012. Measurements of
is higher than the column (Figure 14). hillslope debris flow impact pressure on obstacles. Landslides 9(2):
179–187.
Chen HK, Tang HM. 2006. Method to calculate impact force and im-
pact time of two-phase debris flow. China Journal of Highway and
Conclusions Transport 19(3): 19–23(in Chinese).
Costa JE. 1984. Physical geomorphology of debris flows. In Develop-
The impact processes of debris flow on structures were divided
ments and Applications of Geomorphology. Springer: Berlin
into three stages, according to changes in the flow regime and
Heidelberg; 268–317.
the impact pressure series of the debris flow. The first stage is Cui P. 1999. Impact of debris flow on river chanel in the upper reaches
the sudden and powerful impact of the debris flow head, which of the Yangtze River [J]. International Journal of Sediment Research
was sustained for 1 s with a peak impulse shape. The Bingham 14(2): 201–203.
Reynolds number of the head was 0.003 × 105–4.49 × 105 Fei XJ, Shu AP. 2004. Movement Mechanism and Disaster Control for
where it appeared stable or slightly turbulent. The second stage Debris Flow. Tsinghua University Press: Beijing (in Chinese).
is the continuous and steady impact by the debris flow body Hu KH, Wei FQ, Hong Y, Li XY. 2006. Field measurement of impact
sustained for 3–4 s, where the debris flow approximated steady force of debris flow. Chinese Journal of Rock Mechanics and
laminar flow and the peak impact pressure was 0.68 ± 0.14 Engineering 25(s1): 2813–2819 (in Chinese).
Hu KH, Wei FQ, Li Y. 2011. Real-time measurement and preliminary
times that of the debris flow head. The last stage was the slip
analysis of debris-flow impact force at Jiangjia Ravine, China. Earth
flow of the debris flow tail sustained for 3–4 s. The velocity be- Surface and Processes Land 36(9): 1268–1278.
came much lower, and the sensors were subjected to static Huang HP, Yang KC, Lai SW. 2007. Impact force of debris flow on
pressure. filter dam. Geophysical Research Abstracts. European Geosciences
The wavelet analysis method performed well in separating Union-General Assembly. Vienna, Austria, 03218.
the impact pressures of the slurry and the coarse grains. An em- Hungr O, Morgan GC, Kellerhals R. 1984. Quantitative analysis of
pirical model for calculating the dynamic pressure of the slurry debris torrent hazards for design of remedial measures. Canadian
was proposed. In order to describe the differences of flow pat- Geotechnical Journal 21(4): 663–677.
terns and to achieve scale invariance for miniaturized flume HüBl J, Holzinger G. 2003. Development of design basis for crest open
structures for debris flow management in torrents: miniaturized tests
tests, Froude numbers were introduced into the model.
for the efficiency estimation of debris flow breakers. WLS Report 50
The grain impact pressure was random and did not increase
Band 3.2003. Im Auftrag des BMLFUW VC 7a, Vienna: University
with fluid pressure. The grain impact frequency of the debris of Natural Resources and Applied Life Sciences. (in German).
flow head was greater than that of the debris flow body, which Hübl J, Suda J, Proske D, Kaitna R, Scheidl C. 2009. Debris Flow Impact
revealed that most particles were concentrated in the debris Estimation. Ohrid: Macedonia.
flow head, and the number increased with height. The re- Ishikawa N, Inoue R, Hayashi K, Hasegawa Y, Mizuyama T. 2008.
sponses to the grain impact pressure at the three depths were Experimental approach on measurement of impulsive fluid force
not synchronous. The impact frequencies from bottom to top using debris flow model, INTERPRAEVENT 2008- Conference
were 210 ± 50, 287 ± 29 and 226 ± 35 times per second, Proceedings: 343–354. Dornbirn, Austria.
and the frequencies for the debris flow head were 490 ± 69, Kang ZC, Li ZF, Ma AN, Luo JT. 2004. Debris Flow Research in China.
Science Press: Beijing (in Chinese).
493 ± 52 and 336 ± 114 times per second. The peak impact
Misiti M, Misiti Y, Oppenheim G, Poggi J. 1996. Wavelet Toolbox. The
pressure of particles at different flow depths was 40–160 kPa
Math Works Inc.: Natick, MA.
for the 24 measurements of the surge. It was 3.2 ± 1.5 times Mizuyama T. 1979. Computational method and some considerations
the dynamic pressure of slurry at the bottom, 3.1 ± 0.9 times on impulsive force of debris flow acting on sabo dams. Journal
for the middle sensor and 3.3 ± 0.9 times for the surface sensor. of the Japan Society of Erosion Control Engineering 11(2): 40–43
Some valuable results were obtained in the miniaturized (in Japanese).
flume experiment, but the dynamic and kinematic similarities Proske D, Suda J, Hübl J. 2011. Debris flow impact estimation for brea-
with the prototype were hard to achieve; in particular, the im- kers. Georisk: Assessment and Management of Risk for Engineered
pact load of big boulders was difficult to simulate. Therefore, Systems and Geohazards 5(2): 143–155.
Qian N, Wang ZY. 1984. A preliminary study on the mechanism of
there are still some limits to the study of the nature of the impact
debris flow. Acta Geographica Sinica 39(1): 33–43 (in Chinese).
force of big grains in our experiment.
Scheidl C, Chiari M, Kaitna R, Müllegger M, Krawtschuk A,
Zimmermann T, Proske D. 2013. Analysing Debris-Flow Impact
Acknowledgements—This work has been supported by the National Models, Based on a Small Scale Modelling Approach. Survey of
Natural Science Foundation of China (41030742, 41190084) and Key Geophysics 34(1): 121–140.
Project of the Chinese Academy of Sciences (Grant No.KZZD-EW-05-01). Scotten P, Trivellato F. 1995. Dynamic pressure on check-dam due to
debris flow collision. In Twelfth Australian Fluid Mechanics Confer-
ence, The University of Sydney, Australia, 187–190.
References Suda J, Strauss A, Rudolf-Miklau F, Hübl J. 2009. Safety assessment of
barrier structures. Structure and Infrastructure Engineering 5(4):
Arattano M, Franzi L. 2003. On the evaluation of debris flows dynamics 311–324.
by means of mathematical models. Natural Hazards and Earth System Suwa H, Okud S, Yokoya K. 1973. Observation system on rocky mud-
Sciences 3(6): 539–544. flow. Bulletin of the Disaster Prevention Research Institute 23(3–4):
Armanini A. 1997. On the dynamic impact of debris flows: Lecture 59–73.
Notes in Earth Sciences.In Recent Development on Debris Flows, Tang JB, Hu KH, Zhou GD, Chen HY, Zhu XH, Ma C. 2013. Debris
Armanini A, Michiue M (eds). Springer: Berlin; 64: 208–226. flow impact pressure signal processing by Wavelet Analysis.
Armanini A, Dalrì C, Putta FD, Larcher M, Rampanelli L, Righetti M. Journal of Sichuan University (Engineering Science Edition) 45(1):
2004. Experimental Analysis on the Hydraulic Efficiency of Mudflow 8–13 (in Chinese).
Takahashi T. 2007. Debris Flow: Mechanics, Prediction and Counter- Zanchetta G, Sulpizio R, Pareschi MT, Leoni FM, Santacroce R. 2004.
measures. Taylor & Francis/Balkema: Leiden, Netherlands; 19–24. Characteristics of May 5-6, 1998 volcaniclastic debris flows in the Sarno
Tiberghien D, Laigle D, Naaim M, Thibert E, Ousset F. 2007. Experi- area (Campania, southern Italy) relationships to structural damage and
mental investigations of interaction between mudflow and an obsta- hazard zonation. Journal of Volcanology and Geothermal Research
cle. In Debris Flow Hazard s Mitigation: Mechanics, Prediction, and 133(1): 377–393.
Assessment, Chen CC, Major JJ (eds). Chengdu; 681–687. Zanuttigh B, Lamberti A. 2004. Analysis of debris wave development
Wang YY, Jan CD, Chen XQ, Han W. 2003. Self-organization criticality with one-dimensional shallow-water equations. Journal of Hydraulic
of debris flow rheology. Chinese Science Bulletin 48(17): 1857–1861. Engineering 130: 293–304.
Watanabe M, Ikeya H. 1981. Investigation and analysis of volcanic Zanuttigh B, Lamberti A. 2006. Experimental analysis of the impact of
mud flows on Mount Sakurajima Japan. In Erosion Sediment Trans- dry avalanches on structures and implication for debris flows. Journal
port Measurement (Proceedings of the Florence Symposium, June of Hydraulic Research 44(4): 522–534.
1981); 245–256. Zeng C, Cui P, Su ZM, Lei Y, Chen R. 2014. Failure modes of reinforced
Wei H. 1996. Experimental study on impact force of debris flow heads. concrete columns of buildings under debris flow impact. Landslides
China Railway Science 17(3): 50–62 (in Chinese). 1–11. DOI: 10.1007/s10346-014-0490-0.
Wendeler C, Volkwein A, Roth A, Denk M, Wartmann S. 2007. Field Zhang SC. 1993. A comprehensive approach to the observation and
measurements used for numerical modelling of flexible debris flow prevention of debris flows in China. National Hazards 7(1): 1–23.
barriers. In Debris-Flow Hazards Mitigation: Mechanics, Prediction, Zhang SC, Yuan JM. 1985. Impact force of debris flow and its detection.
and Assessment, Chen CC, Major JJ (eds). Chengdu; 681–687. Memoirs of Lanzhou Institute of Glaciology and Cryopedology
Yang H, Wei F, Hu K, Chernomorets S, Hong Y, Li X, Xie T. 2011. Measuring Chinese Academy of Sciences, 269–274 (in Chinese).
the internal velocity of debris flows using impact pressure detecting in the Zhang YP. 2009. Signal Identification Method to Debris Flow Impac-
flume experiment. Journal of Mountain Science 8(2): 109–116. tion. Chongqing Jiaotong University: Chongqing, China (in Chinese).

Earth Surf Processes Landf - 2015 - Cui - Experimental Analysis On The Impact Force of Viscous Debris

Uploaded by

Copyright:

Available Formats

Earth Surf Processes Landf - 2015 - Cui - Experimental Analysis On The Impact Force of Viscous Debris

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Earth Surf Processes Landf - 2015 - Cui - Experimental Analysis On The Impact Force of Viscous Debris

Uploaded by

Copyright:

Available Formats

EARTH SURFACE PROCESSES AND LANDFORMS

Earth Surf. Process. Landforms 40, 1644–1655 (2015)

Experimental analysis on the impact force of

Received 26 June 2014; Revised 19 March 2015; Accepted 23 March 2015

Introduction phase vitally influences the motion and impact force of a

Figure 1. The experimental equipment.

Figure 2. Cumulative grain size distribution of the debris flow.

Table I. Tests and fluid properties

Density Water Volumetric Total Flume Dynamic Yield Froude

1 1000 1 0.00 0.23 13 0.001 — 5.9

20 which is 0.4 m from the sensor support. The distance of 0.4

12 estimated with the models proposed by Fei and Shu (2004)

Results and Analysis

D8f–D1f represents the grain impulse signal, which implies a 35

2. Threshold detail coefficients 20

Pressure of slurry (kPa) Pressure of grain (kPa)

1 1000 4.86 7.00 11.15 12.54 10.82 — — —

(c) The static pressure of the debris flow tail on sensors

Figure 7. Impact processes of debris flow on sensors.

16 8 Empirical model to calculate dynamic pressure

Figure 9. Comparison of debris flow impact pressure (measured by

1. The structure is hit by the debris flow head with a turbulent

Figure 12. Simplification of the impact processes of debris flow on structures.

You might also like