The Evolution of Grain Size Distribution in Explosive Rock Fragmentation – Sequential Fragmentation Theory Revisited

Abstract:

Fragmentation is a ubiquitous phenomenon in many natural and engineering systems. It is the process by which an initially competent medium, solid or liquid, is broken up into a population of constituents. Examples occur in collisions and impacts of asteroids/meteorites, explosion driven fragmentation of munitions on a battlefield, as well as of magma in a volcanic conduit causing explosive volcanic eruptions and break-up of liquid drops. Besides the mechanism of fragmentation the resulting frequency-size distribution of the generated constituents is of central interest. Initially their distributions were fitted empirically using lognormal, Rosin-Rammler and Weibull distributions (e.g. Brown & Wohletz 1995). The sequential fragmentation theory (Brown 1989, Wohletz at al. 1989, Wohletz & Brown 1995) and the application of fractal theory to fragmentation products (Turcotte 1986, Perfect 1997, Perugini & Kueppers 2012) attempt to overcome this shortcoming by providing a more physical basis for the applied distribution. Both rely on an at least partially scale-invariant and thus self-similar random fragmentation process.

Here we provide a stochastic model for the evolution of grain size distribution during the explosion process. Our model is based on laboratory experiments in which volcanic rock samples explode naturally when rapidly depressurized from initial pressures of several MPa to ambient conditions. The physics governing this fragmentation process has been successfully modelled and the observed fragmentation pattern could be numerically reproduced (Fowler et al. 2010). The fragmentation of these natural rocks leads to grain size distributions which vary depending on the experimental starting conditions. Our model provides a theoretical description of these different grain size distributions. Our model combines a sequential model of the type outlined by Turcotte (1986), but generalized to cater for the explosive process appropriate here, in particular by including in the description of the fracturing events in which the rock fragments, with a recipe for the production of fines, as observed in the experiments. To our knowledge, this implementation of a deterministic fracturing process into a stochastic (sequential) model is unique, further it provides the model with some forecasting power.