Skip to main content

Mass-waste modelling across scales

Periodic Reporting for period 1 - MAWAMOSCA (Mass-waste modelling across scales)

Reporting period: 2017-10-01 to 2019-09-30

Mass-wasting is the general term for the transfer of Earth material down hillslopes. Some examples are rockfalls, debris flows, ice avalanches or soil creep. Not only do such events sculpt the local topography but in many cases they pose as natural hazards, having significant economic impact or by seriously endangering human lives.
Mass-waste events usually consist of 3 phases: a solid phase (a large number of solid particles), a liquid phase (usually water) and a gaseous phase. Capturing the exact properties of the solid phase is a very important task in a number of cases, as some specific properties of the dynamics of the mass movement can only then be understood, like segregation (when for example different sized particles wander to different parts of the flow, changing the movement dynamics). These properties are usually modeled by so called discreet element models (DiEM), that follow each particles trajectory individually. However these codes are usually written for spherical particles, while in real life most particle shapes are non-spherical. This has led to following project objectives:
1. Develop an efficient, open-source 3D discrete element model based software called: MWDiEM, that can model realistic particle shapes found in dry or mostly dry mass-waste events and make this code ready for application to real-world natural hazard events (within efficiency and model limitations).
2. Validate the developed code by comparing its results to laboratory experiments and real-world case studies. Connect MWDiEM with the two-phase continuum model based software r.avaflow in order to create a framework that makes modeling across scales possible.
3. Apply this framework along with experiments to investigate the importance of segregation on dry mass-flow dynamics and runout zone geometry.
4. Determine possible future fundamental research and development directions and engineering applications for both the framework and MWDiEM.
The simulation framework MWDiEM was fully developed during the Project, and is freely available via GitHub under the General Public License. In order for this, during the Project implementation a number of algorithms found in the literature were compared, modified if necessary, coded and then tested. The code itself is ported to the GPU making sure that the computational performance of the simulations allow for reasonable running times even for bigger systems.
Validation of the code was first done by testing each and every module of the code on simple geometries to see if the results are physically meaningful as expected. The code is currently being further validated on a number of experimental cases and real-world examples. Furthermore simulations in order to investigate the exact effect of shape segregation have already began. Future research and development directions and engineering applications for code were also identified. The results were disseminated at a number of conferences, workshops and meetings, and papers detailing the work are under preparation.
MWDiEM is a free, open source GPU based discreet element model simulation package that is able to efficiently model a large number of cohesionless polyhedral particles, with a special emphasis on real world geomorphic systems. This is achieved by adding a number of features for the simulation of such systems, like GIS file support, particle breakup and erosion, possibility of entrainment, etc. As to our knowledge this makes the code a unique modeling tool for both geoscientists and granular scientists alike.
With the help of the simulation package way we will be able to answer a whole set of very interesting questions bringing both geoscientists and granular scientists closer to a deeper understanding on granular mass-waste events:
(a) What kind of particle shape and/or size distribution causes the largest change in the runout zone geometry?
(b) Is there any distinct point during the flow when segregation is not important any more, and both the r.avaflow and the MWDiEM predictions reveal the same results?
(c) Which (and under what conditions) is the stronger effect: shape or size segregation?
(d) How does the length and shape of the flowpath influence the runout geometry?
(e) Can we see any fingering or self- channeling of the flow due to shape segregation?
(f) What is the mechanism behind any observed segregation? Is it similar to kinetic sieving or something very different?
(g) Is there any crystallization associated with the polyhedral shape of the particles?
(h) Is there any kind of “lubrication” effect associated with the shape-segregation, so is there less force needed to shear the material once the segregation has started (leads to less dissipation in the system)?
Internal working structure of MWDiEM