Rapid injection of particles and gas into non-fluidized granular material, and some volcanological implications

In diatremes and other volcanic vents, steep bodies of volcaniclastic material having differing properties (particle size distribution, proportion of lithic fragments, etc.) from those of the surrounding vent-filling volcaniclastic material are often found. It has been proposed that cylindrical or cone-shaped bodies result from the passage of “debris jets” generated after phreatomagmatic explosions or other discrete subterranean bursts. To learn more about such phenomena, we model experimentally the injection of gas-particulate dispersions through other particles. Analogue materials (glass beads or sand) and a finite amount of compressed air are used in the laboratory. The gas is made available by rapidly opening a valve—therefore the injection of gas and coloured particles into a granular host is a brief (<1 s), discrete event, comparable to what occurs in nature following subterranean explosions. The injection assumes a bubble shape while expanding and propagating upwards. In reaction, the upper part of the clastic host moves upward and outward above the ‘bubble’, forming a ‘dome’. The doming effect is much more pronounced for shallow injection depths (thin hosts), with dome angles reaching more than 45°. Significant surface doming is also observed for some full-scale subterranean blasts (e.g. buried nuclear explosions), so it is not an artefact of our setup. What happens next in the experiments depends on the depth of injection and the nature of the host material. With shallow injection into a permeable host (glass beads), the compressed air in the “bubble’ is able to diffuse rapidly through the roof. Meanwhile the coloured beads sediment into the transient cavity, which is also closing laterally because of inward-directed granular flow of the host. Depending on the initial gas pressure in the reservoir, the two-phase flow can “erupt” or not; non-erupting injections produce cylindrical bodies of coloured beads whereas erupting runs produce flaring upward or conical deposits. Changing the particle size of the host glass beads does not have a large effect under the size range investigated (100–200 to 300–400 μm). Doubling the host thickness (injection depth) requires a doubling of the initial gas pressure to produce similar phenomena. Such injections—whether erupting or wholly subterranean—provide a compelling explanation for the origin and characteristics of multiple cross-cutting bodies that have been documented for diatreme and other vent deposits.     

Platform loading from explosions in saturated sand using avisco-plastic model

A set of experiments were conducted at the Aberdeen Proving Grounds 1 (Taylor, L.C., Skaggs, R.R. and Gault, W., 2005. Vertical impulse measurements of mines buried in saturated sand. Proceedings of the 31st Annual Conference on Explosives and Blasting Technique, Orlando, FL, 6 - 9 February, 2.) in which explosive charges were buried in saturated sand beneath a suspended rigid platform. The goal of these experiments was to measure the dependence of the impulse transmitted to the platform on the standoff distance and the charge burial depth. Simulations of these experiments were performed using the BUB2D axi-symmetric code using a frictional-cohesive visco-plastic model to describe the response of the saturated sand. This code solves a constrained set of conservation laws in which the liquid region (in this case saturated sand) is assumed incompressible. The explosion is initialized as a high pressure gas bubble (void) within the fluid. Comparisons of the simulations to the experiments are presented together with a study of the physical phenomenology associated with the loading process. In particular, it is shown that the force imparted to the platform is a combination of the impact of the sand on top of the explosion gas bubble and the pressure of the bubble as it expands before venting into the atmosphere. Under certain conditions, when the platform standoff is sufficiently small and the platform is sufficiently large, the bubble can over-expand before venting and pull the platform downward. This phenomenon was studied further through carefully measured and photographed small-scale experiments performed at the University of Maryland 2 (Fourney, W., Leiste, U., Bonenberger, R. and Goodings, D., 2005, Predicting explosive impulse by means of small scale tests. Proceedings of the 31st Annual Conference on Explosives and Blasting Technique, Orlando, FL, 6 - 9 February, 2.). The small-scale experiments provide additional important validation benchmarks for our model.     

Simulation of strong earthquake motion by explosions — experiments at the Lyaur testing range in Tajikistan

Strong motion data of 10 controlled explosion experiments conducted in 1977 at the Lyaur testing range in the Republic of Tajikistan are revisited. The explosions were detonated in arrays, with time delay between detonation of array lines. Ground accelerations, as large as 1.6g, were recorded at 4 sites by SMA-1 accelerographs. The records were recently digitized and processed with modern accelerogram data processing software. The amplitude and spectral characteristics of these data are here compared with those of strong earthquake shaking data and other published explosion data. The comparison of the Fourier amplitude spectra with estimates by recent empirical scaling laws for strong ground motion, in the near-field of earthquakes, suggests that such explosions can offer powerful possibilities (at present forgotten and neglected) for testing of almost full-scale structures (1/2 to 1/3 scaled models). It is suggested that by going into rather than avoiding the nonlinear zone surrounding the explosions, new testing methods can be developed to simulate near-field nonlinear strong motion of soft soils, found in most metropolitan areas in the seismically active regions.     

Neural network methods in hydrodynamic yield estimation

Hydrodynamic theory allows us to use the speed of a shock wave front to determine the yield of an explosion. On the basis of this theory we developed a neural network to estimate a yield of underground explosions from the shock wave radius versus time (RVT) data, as measured by continuous reflectometry for radius versus time experiments (CORRTEX). The proposed method not only replaces the subjective elements of conventional algorithms, but produces significantly improved yield estimates. The network was trained with real hydrodynamic data and its performance on unseen test events was studied. A backpropagation network was employed; the architecture consisted of ten input units, a hidden layer with eleven hidden units, and one output unit. The network was trained with thousands of input-output measurement vectors, the feasible input set, derived from the RVT data from only four other training or standard events (selected on the basis of the given RVT data from the unknown event). The feasible input vectors were propagated through the trained network and the network output was used to select the optimum yield estimate. Elements of the input vector were: center of energy (COE) offsets, shock front radii, and time onset and interval of analysis for both the standard and unknown events. We studied the performance of the proposed system using 24 Nevada Test Site (NTS) events that were located in the geologic medium tuff. Sensitivity analysis of the proposed method to the assumed nominal COE offset is discussed. Variations of the proposed system that might lead to further improvements in performance are suggested.