6.1 Modeling repacking in a granular medium from a continuum perspective
The loading curves for Hoyos et al. (2022), in some cases, display sharp discontinuities. These discontinuities were termed “load drops” by Hoyos et al. (2022) and were attributed to the disruption or failure of force chains. Force chains - collections of jammed particles prevented from freely rotating and translating (Bergantz et al. , 2017, Philpotts et al. , 1998, Qin & Suckale, 2020, Schleicher & Bergantz, 2017) - build up over the duration of the experiment. The formation of force chains acts to jam the granular medium and is a highly stochastic process, as illustrated from the variable evolution of \(\Sigma\) for repeated experiments using the same particle shape, size, and distribution. Hoyos et al.(2022) demonstrated that the buildup of force chains may be attributed to the fact that the ability of particles to freely rotate is impinged by the apparatus walls, with larger particles experiencing locking more frequently.
The development of coupled modeling-experimental studies that validate models for compaction is rare as of now but necessary. The results of the optimizations and the load curves of the experiments of Hoyoset al. (2022) and the porosity profiles of the olivine centrifuge experiments are in good agreement (Fig. 5 , Fig. 6 , andFig. 9 ). It is an important result because compaction models have been applied to model phase separation in magma reservoirs with limited experimental validation. We find that across all particle types tested by Hoyos et al. (2022) – both in the case of pure and mixture endmembers – the model fits the data well, even for the experiments with load curves that exhibit sharp discontinuities. The fact that the continuum-scale model fits the data well and within reasonable bounds for the inverted parameters is promising.