Cigala, Valeria (2017): The dynamics of starting gas-particle jets: a volcanic scenario. Dissertation, LMU München: Faculty of Geosciences |
Preview |
PDF
Cigala_Valeria.pdf 6MB |
Abstract
Explosive volcanic eruptions are a threat for a large part of global population and infrastructures. Explosive eruptions are the results of energetic magma fragmentation, where only gas exsolved in the magma drive the eruption, or of the interaction with external water. The mechanisms of fragmentation are complex and various, but despite that at explosive eruption onset the potential energy stored in gas bubbles in the magma always transforms into kinetic energy via gas expansion and produce the ejection of pyroclasts and/or non-juvenile material in the atmosphere. Particle ejection rate, velocity and trajectory differ depending on source conditions, e.g. magma composition, gas overpressure, conduit length, vent geometry, etc. Field observations, when possible, can help to characterize an ejection from which then the source conditions are indirectly retrieved. High-speed and infrared videos of volcanic ejections, seismic and acoustic measurements, as well as petrographycal and geochemical analysis on the pyroclasts ejected offer insight on the eruptive event. Nevertheless, to link observations and source parameters is not trivial and it still requires a certain number of assumptions. Therefore, the knowledge of source conditions stays uncertain. On the other hand, empirical studies can help linking observations and input parameters, since the latter are chosen experimental conditions. In general, laboratory experiments are far less complex than natural eruptions. However, the simplifications imposed benefit the investigation of single processes as well as the understanding of the effects of boundary conditions on such observed dynamics. The goal, at the end, is to learn the patterns of certain dynamics and possibly, to recognize certain characteristics of volcanic eruptions and be able to associate them to source conditions. Additionally, empirical results provide input parameters for numerical modelling and thus hazard assessment. I perform rapid decompression experiments of gas-particle mixtures generating starting jets. I use two different experimental apparatus, the first is the “fragmentation bomb” at the LMU facility and the second the “jet buster” at INGV Rome. With the two setups, it is possible to characterize the effect of boundary conditions such as: 1) vent geometry, 2) tube length, 3) particle load and size, 4) temperature, and 5) overpressure in the reservoir on the dynamics of the ejection of natural particles of different initial size distribution (from 0.125 to 4 mm). In particular, I focus the analysis on particle velocity and trajectory. Observations on particle fragmentation, mass ejection rate and lightning generation are also possible on experiments from the “fragmentation bomb”. The experiments are recorded with a high-speed camera, which provides visual observation of the dynamics. On the “jet buster” experiments, the video recordings are coupled with piezoelectric sensors providing microseismic signals of the related propagation dynamics. The two apparatus are different and complementary. The “fragmentation bomb”, a shock-tube made of metal, is 24 cm long, allows high overpressures (here 150 bar) and temperatures (here 500°C), gas and particles are pressurized in the same chamber and the observations are made at vent exit. The “jet buster”, on the other hand, with its 3 m of transparent PMMA tube allows the observation of the whole propagation and dynamics inside the pipe as well at vent exit. The overpressure threshold is in the order of few bar (here 2 bar), and the gas reservoir is separated and below the sample chamber. In the “fragmentation bomb” experiments, maximum particle velocity shows, in order of importance, 1) negative correlation with tube length; 2) positive correlation with particle load; 3) positive correlation with flaring vent walls, with peaks for funnel 15; 4) positive correlation with temperature, and 5) negative correlation with particle size. The evolution of particle velocity with time in non-linear and is mostly affected by particle load and tube length. Gas maximum initial spreading angle shows, in order of importance: 1) negative correlation with flaring vent walls; 2) negative correlation with experimental temperature; 3) positive correlation with tube length; 4) positive correlation with particle size, and 5) negative correlation with particle load. The gas spreading angle evolution with time shows a bell shape pattern and it is especially appreciable in setup 1 experiments, due to the particles later arrival. This is the main affecting parameter. The particle initial spreading angle shows: 1) positive correlation with particle load, 2) negative correlation with particle size; 3) negative correlation with vent geometry; 4) positive correlation with tube length, and 5) negative correlation with temperature. The particle spreading angle evolution with time shows patterns varying in particular with particle load and tube length. Estimations of the mass ejection rate (MER) and instantaneous mass or particle concentration show peaks of 26kg/s for setup 2 experiments, 7 kg/s for setup 3 and 4.6 kg/s for setup 1. The evolution of the MER with time reflects the evolution of particle velocity with time. Finally, mm to cm electrical discharges, i.e. lightning, are observed. Their appearance is positively correlated with particle load, and negatively correlated with tube length, temperature, particle size, and flaring of vent walls. In the “jet buster”, I perform both gas only and gas-particle mixture experiments. This to compare the elastic response of the system and jets’ dynamics. The gas only experiments includes a pinch of kaolin powder in order to make the flow front propagation visible in the camera. The gas flow front shows an initial fast propagation (up to 500m/s) in the pipe accompanied by an abrupt deceleration (to 150 m/s) at vent exit were it generates a vortex ring. On the other hand, particles show maximum velocities between 40 to 100 cm in the pipe in respect to initial sample position. In addition, in this case, maximum particle velocity shows negative correlation with particle size and the evolution of particle velocity displays a non-linear trend. Good correlation between microseismic signals and process occurring in the pipe is observed. The comparison of the experimental results with natural data collected on Stromboli volcano, Italy, is far from trivial. As mentioned above, volcanic eruptions are characterized by the interaction of several processes, thus making them far more complex. Nevertheless, I think the data set present here provides a promising link for both field volcanology (visual observations and quantitative monitoring) as well as numerical modelling in order to advance our understanding of explosive volcanic eruptions and assess the related hazard.
Item Type: | Theses (Dissertation, LMU Munich) |
---|---|
Subjects: | 500 Natural sciences and mathematics 500 Natural sciences and mathematics > 550 Earth sciences |
Faculties: | Faculty of Geosciences |
Language: | English |
Date of oral examination: | 20. December 2017 |
1. Referee: | Dingwell, Donald Bruce |
MD5 Checksum of the PDF-file: | fa5d1ac83667aaee19313dc836a0b361 |
Signature of the printed copy: | 0001/UMC 25330 |
ID Code: | 21893 |
Deposited On: | 09. Mar 2018 14:58 |
Last Modified: | 23. Oct 2020 17:49 |