 |
| Titel |
Deformation and failure of single- and multi-phase silicate liquids: seismic precursors and mechanical work |
| VerfasserIn |
Jeremie Vasseur, Yan Lavallée, Kai-Uwe Hess, Joachim Wassermann, Donald B. Dingwell |
| Konferenz |
EGU General Assembly 2013
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 15 (2013) |
| Datensatznummer |
250081481
|
|
|
|
|
|
| Zusammenfassung |
| Along with many others, volcanic unrest is regarded as a catastrophic material failure
phenomenon and is often preceded by diverse precursory signals. Although a volcanic system
intrinsically behave in a non-linear and stochastic way, these precursors display
systematic evolutionary trends to upcoming eruptions. Seismic signals in particular are
in general dramatically increasing prior to an eruption and have been extensively
reported to show accelerating rates through time, as well as in the laboratory before
failure of rock samples. At the lab-scale, acoustic emissions (AE) are high frequency
transient stress waves used to track fracture initiation and propagation inside a rock
sample. Synthesized glass samples featuring a range of porosities (0Â –Â 30%) and
natural rock samples from volcán de Colima, Mexico, have been failed under high
temperature uniaxial compression experiments at constant stresses and strain rates. Using
the monitored AEs and the generated mechanical work during deformation, we
investigated the evolutionary trends of energy patterns associated to different degrees of
heterogeneity. We observed that the failure of dense, poorly porous glasses is achieved by
exceeding elevated strength and thus requires a significant accumulation of strain,
meaning only pervasive small-scale cracking is occurring. More porous glasses as
well as volcanic samples need much lower applied stress and deformation to fail,
as fractures are nucleating, propagating and coalescing into localized large-scale
cracks, taking the advantage of the existence of numerous defects (voids for glasses,
voids and crystals for volcanic rocks). These observations demonstrate that the
mechanical work generated through cracking is efficiently distributed inside denser and
more homogeneous samples, as underlined by the overall lower AE energy released
during experiments. In contrast, the quicker and larger AE energy released during
the loading of heterogeneous samples shows that the mechanical work tends to
concentrate in specific weak regions facilitating dynamical failure of the material
through dissipation of the accumulated strain energy. Applying a statistical Global
Linearization Method (GLM) in multi-phase silicate liquids samples leads to a
maximum likelihood power-law fit of the accelerating rates of released AEs. The
calculated α exponent of the famous empirical Failure Forecast Method (FFM)
tends to decrease from 2 towards 1 with increasing porosity, suggesting a shift
towards an idealized exponential-like acceleration. Single-phase silicate liquids
behave more elastically during deformation without much cracking and suddenly
releasing their accumulated strain energy at failure, implying less clear trends in
monitored AEs. In a predictive prospective, these results support the fact that failure
forecasting power is enhanced by the presence of heterogeneities inside a material. |
|
|
|
|
|
|