Interaction of regular non-spherical particles with rigid barrier in gravity-driven dry granular flow

Structural countermeasures such as rigid and flexible barriers are commonly installed in mountainous regions to intercept mass-wasting processes. Without sufficient and reliable comparable physical data, the study of impact mechanisms remains difficult and not well understood. In this study, a newly developed flexible model barrier together with a rigid barrier are used to simulate either dry granular or viscous liquid impacts on these model barriers in a geotechnical centrifuge. The novel flexible barrier is made of four instrumented cables controlled by spring mechanisms to replicate a bilinear prototype loading response. Tests revealed that regardless of barrier type, both dry granular and viscous flows could have similar frontal dynamic impact coefficients around unity. Compared with the kinetic energy of flow mass (∼10 MJ), only 249 kJ of flexible barrier energy capacity was mobilized. This implies that debris-resisting barriers may only be required to intercept the dynamic flow front as the subsequent flow energy may mainly be dissipated through internal shearing. Attributing to the large deformation of the flexible barrier, the granular static load acting on the flexible barrier could be 39% lower than that on the rigid barrier, resulting in an active failure mode and a lower earth pressure.

Debris flows pose threats to sustainable development in many countries worldwide, including China, Japan, Switzerland and USA. To mitigate these flows, rigid and flexible barriers are commonly installed along the predicted flow paths. To arrest large volumes of debris flow, several barriers may be installed in series to create a cascading effect to progressively decelerate and retain the debris. Barriers may even be designed with a basal clearance to allow small discharges to pass underneath the barrier to reduce the peak impact force. Despite the importance of barriers as life-saving assets, their design remains essentially empirical because of the highly heterogeneous and scale-dependent nature of debris flow. These features of debris flow have hindered an understanding of their fundamental impact mechanisms, thereby hampering the development of scientific design guidelines to enable robust and cost-effective barriers. This forum paper presents a collection of physical experiments modelling the impact mechanisms of the two extreme cases of water and dry granular flows, and two-phase debris flows against single and dual rigid barriers, and a single flexible barrier. Furthermore, the effects of a basal clearance on the impact dynamics of dry granular flow against a single rigid barrier are examined. Experiments were conducted at two different scales, including 5 m-long and 28 m-long flumes. Based on the observed impact mechanisms and measured data, a newly developed analytical framework for designing multiple rigid barriers was evaluated. Recommendations and procedures are provided for the design of single and multiple rigid barriers with and without a basal clearance.KeywordsDebris flowImpactRigid barrierFlexible barrierMultiple barriersPhysical modellingBasal clearance

Effective design of rigid debris-resisting barriers remains a challenging problem for practitioners. At present, a pseudo-static approach is commonly used for the estimation of debris impact loads acting on barriers. In this approach, the dynamic impact pressure is estimated using the hydrodynamic equation and an assumed value of the dynamic pressure coefficient. The bulk density and travel velocity of the landslide debris are also required for the calculations. To this end, the debris velocity under free-field conditions can be obtained from a conventional debris mobility assessment. A limitation of this approach is that it ignores the attenuation of the debris’ velocity during the deposition process behind the barrier. To provide a scientific basis for assessing the velocity of granular debris flows during impact on rigid barriers, this paper presents results of debris velocity from a laboratory study of debris-barrier interaction using physical flume modelling. High-speed cameras were used to capture the dynamic motion of the flowing debris. The images obtained were analysed using the particle image velocimetry (PIV) technique to estimate the velocity of the debris. The results show that, following the first stage of impact, the accumulated debris behind the model barrier formed a stationary zone and caused the remaining debris to slow down in a run-up process. As a result, the peak debris momentum was 30% lower compared to that observed under free-field conditions. The findings of this study will allow practitioners to optimise the design of debris-resisting barriers in the future.

SummaryThis study uses an incompressible smoothed‐particle hydrodynamics model to investigate the interaction between dry granular material flows and rigid barriers. The primary aim is to summarise some practical guidelines for the design of debris‐resisting barriers. The granular materials are modelled as a rigid‐perfectly plastic material where the plastic flow corresponds to the critical state. The coupled continuity equation and momentum equation are solved by a semi‐implicit algorithm. Compared with flows in controlled flume experiments, the model adequately reproduces both the kinetic of the flows and the impact force under various conditions. Then the numerical simulations are used to study the detailed interaction process. It is illustrated quantitatively that the interaction force consists of two parts, ie, the earth pressure force caused by the weight of the soil and a dynamic force caused by the internal deformation (flowing mass on top of a dead zone). For the estimation of impact load, this study suggests that an increased earth pressure coefficient depending on the Froude number should be incorporated into the hydrostatic model.

Steep creek hazards are characterised according to their solid fraction, νs, which governs the flow dynamics. To arrest these hazards, rigid multiple barriers are installed along a channel. However, existing design guidelines are deficient because the fundamental overflow and landing mechanisms are not well understood. In this study, experiments were conducted using a 5 m long flume to model the impact of flows with different νs values on a rigid barrier. By varying the νs of sand–water mixtures, the dynamics of different steep creek hazards were modelled. The barrier height was varied to obtain barrier Froude numbers Frb (ratio of flow inertia to potential energy related to barrier height) from 0·7 to 3·4. Results show that barriers should be designed so that Frb < 1, which leads to downward overflow with reduced landing distances. In addition, the landing distance for the watery flows (νs = 0 to 0·1) is up to 87% longer than that of dry granular flows. This implies that the design spacing between barriers should cater for different types of steep creek hazards and Frb. An existing framework for multiple barriers is modified to provide guidance on the selection of the barrier height using the Frb.

Constructing multiple barriers along the predicted flow paths has been recognised as an effective engineering measure to intercept large volumes of debris flows. In the field, barriers usually have a basal clearance, which is an opening between the channel bed and the base of the barrier, to allow small discharges pass through the barrier and prevent the accumulation of stream loads. However, the current design of basal clearance is still based on an empirical criterion. The objectives of this research are to investigate the effects of basal clearance on the interaction between debris flow and multiple rigid barriers and to propose and verify a new analytical framework for flow impact on multiple rigid barriers with basal clearance. Findings from this study can serve to improve the current design guidelines on multiple-barrier system. The newly proposed analytical framework consists of an impact equation for barrier with a basal clearance, discharge equations to estimate flow properties of the basal discharge and a landing equation to estimate the flow properties after flow lands on the basal discharge. To verify the proposed framework, physical and numerical modelling were adopted. A series of dry granular flow tests has been carried out with a 5-m flume model to investigate the fundamental impact behaviour of idealised uniform particles on a rigid barrier with basal clearance. A discrete element model was calibrated by the physical tests and a series of numerical parametric studies has been conducted with varying barrier heights and barrier spacing. Furthermore, due to scale dependent nature of debris flow, a 28-m flume facility has been newly developed to investigate the impact dynamics of debris flow on dual rigid barriers with basal clearance. It is found that basal clearance can be designed to optimise barrier size and minimise maintenance work for a multiple-barrier system. Basal clearance with a ratio between clearance height and flow depth, H<sub>c</sub>/h<sub>i</sub>, ranging from 0.5 to 1.0 with Fr = 7 can reduce the impact force and runup height of two-phase debris flow by up to 55% and 43%, respectively. The effectiveness of basal clearance on mitigating impact can be characterised by Friction number, N<sub>fric</sub>, of the flow. A larger N<sub>fric</sub> represents a higher shear resistance among particles that would resist the flow to be discharged and lead the flow tend to be redirected to parallel to the barrier. N<sub>fric</sub> of dry granular flow that is by up to 55 times higher than the two-phase flow leads the basal clearance about 30% less effective on reducing the impact force and runup height of dry granular flow. Characteristic particle size, d<sub>max</sub> that represents the largest particle of the debris material, is found capable of determining the occurrence of jamming at the basal clearance. Basal clearance height H<sub>c</sub>/d<sub>max</sub> ≤ 2.0 and H<sub>c</sub>/d<sub>max</sub> ≥ 3.0 may lead and not lead jamming at the clearance, respectively, and could be adopted for a barrier to regulate debris flow without losing barrier retention and without retaining any debris, respectively. The proposed impact equation can well estimate both dry granular flow and debris flow impact on a rigid barrier with a basal clearance. The proposed landing equation can provide an upper bound for flow depth and velocity near the landing location. However, with an increasing flow distance, the decrease of downstream flow depth may lead that only the discharge equations are needed to estimate the flow properties.

Some types of barriers are designed with a clearance between the bottom of the barrier and the channel bed. This feature allows small discharges to pass, thereby reducing the maintenance required over the service life of the barrier. Aside from the practical function of a clearance, it influences the impact force, jump height, and discharge. In this paper, a series of physical experiments was conducted using a 6 m long flume to model the interaction between dry granular flow and rigid barrier with a basal clearance. The ratio between the clearance and particle diameter Hc/D was varied from 0 to 10. The channel inclination was varied from 15° to 35° to achieve different Froude numbers before impact. A new impact model for predicting impact force exerted on the barrier with a basal clearance is presented and evaluated. Results reveal that Hc ≥ 3D is capable of reducing the impact force and overflow. Findings from this study highlight the importance of considering the effects of basal clearance on the design of multiple-barrier systems.

Impact and erosion dynamics of inclusion-enriched dry granular flows on rigid barriers with basal clearance: Numerical insights from hybrid MP-DEM simulations

Both the solid and fluid phases in a debris flow vitally influence the impact dynamics against a rigid barrier. However, previous numerical and analytical models commonly adopt an equivalent fluid approach, which does not explicitly consider solid–fluid interaction when impacting a barrier. This study investigates the role of solid–fluid interaction on the impact mechanism and load exerted on a rigid barrier. A coupled computational fluid dynamics and discrete-element method (CFD–DEM) is used to study the effect of solid–fluid interaction on impact by varying the solid fraction from 0 to 0·5. The mesoscopic insight from CFD–DEM modelling reveals that the particle–fluid interaction plays two roles in increasing the kinematic energy of particles: (a) imposing driving force to the particles; (b) reducing inter-particle contact forces and therefore energy dissipation from shearing among grains by applying buoyancy to the particles. Consequently, the run-up height and impact pressure of the water–particle mixture flows at the barrier are higher than those of dry granular flows with the same Froude number (Fr). While impacting the rigid barrier, most of the energy of water–particle mixtures (exhibiting a run-up mechanism) is dissipated as the fluid phase segregates from the mixture and rolls back towards subsequent flow. This differs from the conventional impact mechanism observed for dry granular flows, where the energy is mainly dissipated by way of shearing between layers of sand piling up behind a barrier.

The maximum impact force of granular flow and its action point on a rigid barrier are the key indices of the anti-slip and the anti-overturning calculation, respectively. On the basis of the present impact force models and the observation in model experiments, this paper proposes a new semi-empirical impact force model focusing on the normal impact force and its point of action. By comparing the analytical solution and the experimental results, the new semi-empirical analytical model can estimate the normal impact force composed of both dynamic and static components with an error margin within ± 20% compared with the experimental results and so is for the corresponding point of action and tangential force. To calculate the residual force generated only by the static dead zone, the static friction angle between the dead zone and the chute base should be less than what was measured under kinetic conditions. Based on a large number of tests, it was found that assuming that the reaction force generated by the chute base operates at 2/3 of the deposition length makes the estimated normal impact force best suited to the experimental results.

Granular debris flow baffles are commonly installed in front of rigid barriers to dissipate flow energy and reduce the required barrier impact capacity. Despite the engineering value of baffles, their influence on rigid barrier impact is still not well understood. A previously calibrated discrete element method (DEM) model using a series of flume experiments was adopted to study the effectiveness of installing baffles in front of a rigid barrier. Froude scaling was used to characterize the flow front. Different baffle configurations were examined, namely number of rows, spacing between successive rows (L), and baffle height. Results reveal an optimum row spacing of L/D = 3 (D is the slit size). Row spacing less than L/D = 3 leads to increased peak dynamic force from overflow impacting the barrier, whereas row spacing greater than L/D = 3 results in increased peak dynamic force from the granular debris flow front. Increasing spacing greater than L/D = 3 allows the dispersion of debris between rows and decreases the effectiveness of the second row. Adopting baffle heights greater than 1.5 times the approach flow depth (h) reveals little influence on the peak impact force induced on the barrier.

From the understanding of dynamics and processes of rapid granular flows and the granular-segregation mechanism in gravity-driven flow, we can clarify the particle-composition structure in the downstream areas of avalanches in geophysical contexts, such as landslides, rock falls, and snow-slab avalanches. Such dynamics also provide a basis for geophysical studies. This study experimentally investigates the dynamic behavior and segregation phenomena of a density-bidisperse, rapid, granular flow down a quasi-2D, rough, inclined rectangular chute. Particles with two density ratios are used to investigate the mechanism of density-induced segregation, and four chute-inclination angles are tested to examine the influence of driving forces. The dynamics of the mixture flow—which includes the flow-depth evolution, stream-wise and depth-wise velocity profiles, shear rate, and granular temperature in the upper high-shear band of the flow—are obtained from particle image velocimetry (PIV) measurements. The two-dimensional concentration distributions of the particles in the stream-wise direction are also obtained using 2D image processing to determine the segregation state. In the upstream region, the variation in the concentration of heavier particles is defined as the strength of the density-induced segregation state, Sd. Our results indicate that the mixture-flow parameter—particularly the shear rate and the granular temperature in the upper high-shear band—crucially influence the strength of particle segregation in granular avalanches. In the upstream region, a higher shear rate and a higher granular temperature in the upper high-velocity band result in a smaller drag force in the mixture flow, causing stronger density-induced particle segregation. These results well describe the entire processes of dense granular flows, from upstream initiation to the downstream steady state. Therefore, they reveal the structure of the mixed flow in the depth direction and are expected to explain various gravity-driven mixture granular flows.

Granular flows comprise a wide range of particle sizes. The particle size governs the degree of grain contact and inertial stresses in the flow, thus influencing the mechanism of impact against a rigid barrier. The current commonly adopted design approaches estimating the run-up height are based on the energy principle and the momentum approach. However, both neglect the discrete nature of flows, and do not consider the effects of particle size on the flow regime. In this study, physical experiments using different sizes of monodispersed sand and glass spheres were carried out to investigate the run-up mechanisms on a rigid barrier. Results have shown that the run-up height is not only dependent on the Froude number of the flow before impact, but also on the particle size which principally governs the mechanism of run-up. Inertial flows comprising large particles (the Savage number (NSav) > 0·1) profoundly transfer momentum vertically upon impact, resulting in significant grain saltation and high run-up heights. In contrast, frictional flows comprising fine particles (NSav < 0·1) tend to pile up without significant run-up due to a high degree of contact stresses. This implies that for flows with coarse particles entrained at the front of the flow by way of particle-size segregation, the run-up height is principally influenced by large particles that accumulate at the flow front.

b1450697 HKUST Electronic Theses Computational study of granular debris flow impact on rigid barriers and baffles by Law Pak Hei thesis 2015 xviii, 19-251 pages : illustrations (some color) ; 30 cm Debris flow is one type of natural terrain…Read more ›

Two concepts in modeling the effects of the evolution of porosity in dry granular flows are investigated to illuminate their performance and limitations. To this end, the thermodynamic analysis, based on the Muller-Liu entropy principle, and the quasi-linear theory, are employed to deduce the ultimate constitutive models and the restrictions on their thermodynamic consistencies. The models are employed to study an isothermal dry granular slow flow down an inclined moving plane, of which the results are compared with the experimental outcomes. Results show that, while the two models deliver appropriate equilibrium expressions of the Cauchy stress tensor for compressible grains, the model in which the evolution of porosity is treated kinematically yields a spherical stress tensor for incompressible grains. Only the model with a dynamic evolution of porosity can give rise to a non-spherical stress tensor at equilibrium. Moreover, whilst the former model can better capture the characteristics of flows with slow to moderate speeds, the latter model is more able to describe the features of very rapid flows like avalanches. The present study illustrates the essential difference between the two concepts in modeling the effects of the evolution of porosity, and can be extended for further studies on other microstructural effects in granular flows.