scg-xiii international symposium on landslides. …

SCG-XIII INTERNATIONAL SYMPOSIUM ON LANDSLIDES. CARTAGENA, COLOMBIA- JUNE 15th-19th-2020

Laboratory test to study the effect of comminution in rockfalls

Matas G, Parras E, Lantada N, Gili J, Ruiz-Carulla R, Corominas J, Moya J, Prades A,

Buil F, Nuñez-Andres MA, Puig C

Civil and Environmental Engineering Department. Universitat Politècnica de Catalunya

(UPC.BarcelonaTech)

[email protected]

Abstract

Fragmentation in rockfalls reduces the size of the blocks and increases the number of fragments. It affects the runout

and the impact energies, having implications in the associated hazard and therefore in the subsequent risk. In some

rockfall events, a young debris cover is formed during the first impacts, with a substantial reduction of the particle

size. The large number of small particles generated in the debris cover suggests that beside the breakage of the

particles, comminution by crushing and grinding may also occur during the impact. This contribution presents the

results of the comminution tests carried out in the laboratory. A set of one, three and five stacked bricks of different

colors were released. The tests aim at verifying and quantifying the following hypotheses for the brick located in the

lower position of the stack: i) the fragments generated should be smaller, ii) the number of fragments generated

should increase, and iii) both effects should intensify as the number of the stacked bricks increases. A total of 78

reduced scale tests were carried out using a set of stacked bricks that were dropped from a height of 4.26 m. The

resulting fragments were separated by colors according to their initial position in the stacking and sieved obtaining

the grain size distribution. A high-speed camera was used to record and interpret each test. A 3D photogrammetric

reconstruction after each test was generated to register the final deposit. The obtained grain size distributions confirm

the proposed hypothesis. Confinement reduces the maximum fragment size while increases substantially the total

number of generated fragments and those effects increase with the number of stacked bricks.


1 INTRODUCTION

Fragmentation is a mechanism frequently observed in rockfalls. It consists of the generation of small pieces by breakage and/or disaggregation of the blocks composing the initial rock mass. Rockfalls are often modeled by computing the trajectory of a single block. Fragmentation, or any reduction of particle sized during motion, is largely neglected (Haug et al 2016). However, fragmentation affects the propagation of rockfalls and rock avalanches (Davies and McSaveney, 2009). The trajectories of the generated fragments, the impact energies, and the runout are modified by fragmentation with significant but contrasting effects on risk assessment (Corominas et al. 2019).

Pollet and Schneider (2004) separates the fragmentation into two types: A primary or static fragmentation, where the rock mass separates by breaking rock bridges connecting fragments of more competent rock together (Eberhardt et al., 2004), and a dynamic fragmentation where these particles are continuously reduced in size by comminution (Pollet and Schneider, 2004; Imre et al., 2010). Comminution is the breakage mechanism that occurs when multiple stacked blocks impact together with the ground and the upper ones produce enough overburden stress to break the lower ones. The analysis of inventoried rockfall events suggests that comminution tends to increase with the size of the falling mass (Ruiz-Carulla et al. 2016). In rockfalls of middle to large size, from hundreds to thousands of cubic meters, a Young Debris Cover (YDC) is formed due to the comminution of the detached blocks stacking itself and breaking. Some fragments continue their propagation individually and reach longer runouts (Figure 1).

Real scale rockfall tests are typically performed to characterize rockfall motion parameters (Richtie, 1963; Labiouse & Heidenreich, 2009: Spadari et al. 2012) and very few of these tests study the effect of primary fragmentation (Giacomini et al. 2009; Gili et al. 2016). In some rockfall events, a young debris cover is formed during the first impacts, with a substantial reduction of the particle size (figure 1). The large amount of small particles generated in the debris cover suggests that beside the breakage of the particles, comminution by crushing and grinding may also occur during the impact. However, as far as the authors are aware, the effect of the comminution in rockfall events has not been studied yet.

The experiment presented here is aimed at studying the effect of block confinement on the dynamic fragmentation and on the resultant grain size distribution. The main hypotheses to check are that the lower is the position of the brick in the stack: i) the smaller the fragments generated ii) the greater the number of fragments created, and iii) both effects should intensify with the number of stacked bricks.

Figure 1. Young debris cover generated at the first impact location of the rockfall while several individual blocks traveled larger distances (Esterri d’Àneu, Eastern Pyrenees).


To this purpose, stacks of 1, 3 and 5 bricks of baked clay are released from a height of 4.26 m and generated fragments after breakage are measured to obtain grain size distributions. All tests were recorded using a high-speed camera and several additional cameras to properly interpret the phenomena. Finally, a 3D model of the scattered fragments was captured to properly document each test for future analyses.

2 TEST DESIGN

2.1 Materials

The material to be tested were selected based on two main criteria. First, the material must be weak enough for breaking under impacting energy conditions of the test and secondly, the commercial availability of enough quantity from different colors to be able to distinguish which fragments correspond to each initial piece. The material selected for the test were baked clay bricks. Low strength concrete was discarded since the maximum fall height of our testing site was not enough to reach a high fragmentation degree. Bricks have a standard size of 5 x 10 x 20 cm and five different colors. Figure 2 shows a sample of the bricks types with their relative position on the stacks. Their density ranges between 1.81 and 2.24 g/cm³ depending on the color.

Three settings were tested to evaluate the effect of the added dynamic load due to piece stacking: single piece, three stacked pieces and five stacked pieces (Figure 3).

Figure 2. Numbering of the blocks which corresponds to their relative position on the stacks. Number 1 goes on the bottom of the stacks and 5 on the top.

Figure 3. Distribution of the blocks in sets to release.

2.2 Test set-up

The tests were carried out in the Laboratory of Technology of Structures and Materials (UPC). A device was specially designed to place and release the bricks. It allows the bricks to fall vertically without rotational velocity. The total fall height was 4.26 m, which determines the impacting energy of the bricks stack. The concrete slab of the floor was protected with a 10mm thick steel plate placed on the impacting area. To stop the fragments ejected after breakage a wood frame was built around the steel plate. This protective frame had wood boards around the perimeter with a height of 1.2 m and a plastic rack above the boards to allow light to enter the scene (Figure 4).

Each one of the releases were recorded using a

high-speed camera recording at 400 fps in HD and

a GoPro camera. This high frame rate avoided

blurring effect of the blocks at each frame. Since

the tests were performed indoor, to match light

requirements of the high-speed camera 4 spotlights

were used with a total power of 4000 W. The

camera and the spotlights pointed the scene through

holes in the wood boards protected with

polycarbonate sheets. Targets were placed both in

the floor and the wood boards to allow cinematic

estimations using the video records.


Figure 4. Experimental test in laboratory with bricks release device and wood frame around the impact area.

3 TEST PERFORMANCE

A strict security protocol was designed to minimize risks during the execution of the tests. For each release, the procedure followed consisted of:

1. Placement of tracing paper and graphical targets in the impacting point to record the contact area during impact.

2. Turning on the spotlights and the cameras. 3. Security check: everybody is placed at a safe

distance. 4. Release of the bricks’ stack (see an example in

Figure 5). 5. Turning off all cameras. 6. Photogrammetry: place a control volume in the

scene for calibration and take a photographic coverage to build a 3D model of the fragment deposit.

7. After each release the tracing paper and all fragments were carefully removed, stored and marked with the release reference number to later proceed to the fragment classification and measurement.

Figure 5. Release number 33 corresponding to a 5 bricks stack before (above) and after the impact (below).

4 RESULTS

The grain size distribution of each brick must be obtained to characterize different fragmentation degree for each position within the stack. For each stack first all fragments where sieved using a 4.76 mm sieve to discard very fine particles or dust. Then, all fragments retained in the sieve were visually classified by color, which gave the position on the stack. Once classified by color, the weight of each fragment was measured using a high precision weighing scale. The data acquisition of these weights was semi-automatic. Finally, knowing the density of each of the bricks the weights could be transformed to volumes thus obtaining a granulometric curve of each tested brick. This whole process lasted a couple of weeks.

After this classification and measurement, we know the maximum volume and number of fragments of each tested brick classified by color depending on the position in the stack and by size depending on the number of bricks in the stack (3 or 5).

Figure 6 and Figure 7 show the grain size distributions obtained on a test of 3 and 5 stacked bricks respectively. The relative position of the bricks within the stack is indicated from 1 to 3 or 5, being the brick "1" the located at the bottom. Both plots show the different fragmentation behavior as a function of the position in the stack.


Figure 6. Cumulative number of fragments against fragment volume of the test number 62 (3 bricks stack).

Figure 6 and Figure 7 show how the biggest produced fragment after each brick breaks decreases as lower is the position in the stack. They also show how the number of generated fragments increases as lower is the position. The slopes of the distributions look pretty similar except by brick in position 5 on test 33 (Figure 7) which broke in three main fragments as shown in figure 5.

Figure 7. Cumulative number of fragments against fragment volume of the test number 33 (5 bricks stack).

The number of generated fragments on each position of the stacks for 5 bricks tests are shown on figure 8. It clearly shows how the number of fragments increases as lower is the position on the stack. The maximum number of fragments was 378 for the brick at the bottom position of test #84. The maximum, minimum and average number of fragments generated at each position are shown on Table 1. In a few cases, there was a brick that did not break at all. Analyzing the videos, we observed that the unbroken bricks landed on an edge or on top of a fragments’ pile of the bottom block which had already broke.

Figure 8. Number of generated fragments in 5 brick stacked

tests for each position in the stack.

5E-08 5E-07 5E-06 5E-05 0,00051

10

100

Position 3

Position 2

Poisition 1

Fragment volume [m³]

Cu

mu

lati

ve

nu

mb

er o

f fr

ag

men

ts [

-]

1E-08 1E-07 1E-06 1E-05 0,0001 0,0011

10

100

Position 5

Position 4

Position 3

Position 2

Position 1

Fragment volume [m³]

Cu

mu

lati

ve

nu

mb

er o

f fr

ag

men

ts [

-]


Table 1. Statistics on generated fragments depending on the position of the bricks in the 5 bricks stacks.

Position Max #frag Min #frag Average #frag

1 378 39 227.5

2 307 3 143

3 301 1 114

4 223 1 63.4

5 58 1 20.2

In many tests most of the fragments generated by the brick placed at the bottom of the stack remained in place at the same point they contacted the steel plate (Figure 9). In Figure 10 a sequence of test #60 (5 brick stack) shows how fragments are quickly ejected but that the bottom brick remains in place. At the first contact with the steel plate, the fragments generated in the outer edge are ejected but the ones at the center cannot escape and tend to stay at the impact point. This phenomenon is observed in the deposit when the generated fragments remain on the site as if they were a puzzle (Figure 9).

Figure 9. Detail of the bottom block of the release #60 (fig. 8) The cracking pattern is observed but the fragments in the middle remain in in place like a puzzle.

Figure 10. Sequence of test #60 where a stack of five bricks was dropped. Note how the brick on the bottom stays in place.


5 CONCLUDING REMARKS

The grain size distribution of fragments obtained from this experimental test confirms that the blocks undergoing greater confinement (which increase towards the lowest position on the stack) generate a greater number of fragments while decreasing their maximum size.

This phenomenon, combined with the reduction

of mobility of the confined fragments, which

remain in place at the impact point, may explain the

formation of Young Debris Cover in real cases

where the initial released rock masses are big

enough to produce both the confinement and

comminution effect.

The fragments generated at the edges of the

bricks were quickly ejected. In this case, the

release velocity increased with the confinement.

As a final remark, many researchers

have focused on fragmentation recently since

new technologies allow better data acquisition in

field and laboratory tests and new

methodologies for numerical simulations.

However, there are still few publications in

which the kinematics of the fragments is

analyzed experimentally measured.

The distribution after

and/or

energy

fragmentation and subsequent momentum increase

of some of the fragments remains as

unexplored topic.

ACKNOWLEDGMENT

This study was supported by the research projects Rockmodels (ref. BIA2016-75668-P, AEI/FEDER,UE) funded by the Spanish Ministry of Economy and Competitiveness, co-funded by the Agencia Estatal de Investigación (AEI) and The European Regional Development Fund (ERDF or FEDER in Spanish), and GeoRISK (ref. PID2019-103974RB-I00/AEI/10.13039/501100011033) funded by Ministry of Science and Innovation and AEI.

REFERENCES

Davies, T.R. and McSaveney, M.J. (2009). “The role of

rock fragmentation in the motion of large landslides”. Engineering Geology, 109: 67-79.

Eberhardt, E., Stead, D., and Coggan J. (2004), “Numerical analysis of initiation and progressive failure in natural rock slopes the 1991 Randa rockslide”. Int. J. Rock Mech. Min. Sci., 41(1): 69-87, doi:10.1016/S1365-1609(03)00076-5.

Giacomini, A., Buzzi, O., Renard, B., and Giani, G. (2009). “Experimental studies on fragmentation of

rock falls on impact with rock surfaces”. Int. J. Rock Mech. Min., 46: 708-715.

Gili, JA, Ruiz-Carulla, R., Matas, G, Corominas, J, Lantada, N, Núñez, MA, Mavrouli, O, Buill, F, Moya, J, Prades, A & Moreno, S (2016). “Experimental study on rockfall fragmentation: in situ test design and firsts results”. International Symposium Landslides.

Haug, Ø.T., Rosenau, M., Leever, K. and Oncken O. (2016). “On the energy budgets of fragmenting rockfalls and rockslides: Insights from experiments”. J. Geophys. Res. Earth Surface, 121: 1310-1327.doi:10.1002/2014JF003406.

Imre, B., Laue, J., and Springman S. M. (2010). “Fractal fragmentation of rocks within sturzstroms: Insight derived from physical experiments within the ETH geotechnical drum centrifuge”. Granular Matter, 12(3): 267–285, doi:10.1007/s10035-009-0163-1.

Labiouse, V. and Heidenreich, B. (2009). “Half-scale experimental study of rockfall impacts on sandy slopes”. Nat Hazards Earth Syst Sci; 9: 1981-93.

Parras, E. (2019) “Diseño y realización de ensayos de lanzamiento de bloques de roca a escala reducida para el estudio de la fragmentación por conminución”. Master thesis. Universitat Politècnica de Catalunya (UPC.BarcelonaTech) (in Spanish)

Pollet, N., and Schneider, J.L. (2004). “Dynamic disintegration processes accompanying transport of the Holocene Flims sturzstrom (Swiss Alps)”. Earth Planet Sc. Lett., 221(1-4): 433-448, doi:10.1016/S0012-821X(04)00071-8.

Ritchie, A. (1963). “Evaluation of rockfall and its control”. Highway Res Rec; 17:13–28.

Ruiz-Carulla, R., Corominas, J., Mavrouli, O. (2016). “Comparison of block size Distribution in rockfalls”. In: Landslides and Engineered Slopes. Experience, Theory and Practice – Aversa et al. (Eds). Associazione Geotecnica Italiana, Rome, Italy, Vol 3: 1767-1774.

Spadari, M., Giacomini, A., Buzzi, O., Fityus, F., Giani, P. (2012). “In situ rockfall testing in New South Wales,Australia”. International Journal of Rock Mechanicsand Mining Sciences, 49: 84-93.

scg-xiii international symposium on landslides. …

Documents